This commit is contained in:
2023-07-03 20:39:45 +02:00
parent 25445b2a98
commit 5152f44924
149 changed files with 17896 additions and 5 deletions

View File

@ -0,0 +1,161 @@
# Generated by the protocol buffer compiler. DO NOT EDIT!
from google.protobuf import descriptor
from google.protobuf import message
from google.protobuf import reflection
from google.protobuf import descriptor_pb2
# @@protoc_insertion_point(imports)
import google.protobuf.descriptor_pb2
DESCRIPTOR = descriptor.FileDescriptor(
name='google/protobuf/compiler/plugin.proto',
package='google.protobuf.compiler',
serialized_pb='\n%google/protobuf/compiler/plugin.proto\x12\x18google.protobuf.compiler\x1a google/protobuf/descriptor.proto\"}\n\x14\x43odeGeneratorRequest\x12\x18\n\x10\x66ile_to_generate\x18\x01 \x03(\t\x12\x11\n\tparameter\x18\x02 \x01(\t\x12\x38\n\nproto_file\x18\x0f \x03(\x0b\x32$.google.protobuf.FileDescriptorProto\"\xaa\x01\n\x15\x43odeGeneratorResponse\x12\r\n\x05\x65rror\x18\x01 \x01(\t\x12\x42\n\x04\x66ile\x18\x0f \x03(\x0b\x32\x34.google.protobuf.compiler.CodeGeneratorResponse.File\x1a>\n\x04\x46ile\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x17\n\x0finsertion_point\x18\x02 \x01(\t\x12\x0f\n\x07\x63ontent\x18\x0f \x01(\t')
_CODEGENERATORREQUEST = descriptor.Descriptor(
name='CodeGeneratorRequest',
full_name='google.protobuf.compiler.CodeGeneratorRequest',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
descriptor.FieldDescriptor(
name='file_to_generate', full_name='google.protobuf.compiler.CodeGeneratorRequest.file_to_generate', index=0,
number=1, type=9, cpp_type=9, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
descriptor.FieldDescriptor(
name='parameter', full_name='google.protobuf.compiler.CodeGeneratorRequest.parameter', index=1,
number=2, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=unicode("", "utf-8"),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
descriptor.FieldDescriptor(
name='proto_file', full_name='google.protobuf.compiler.CodeGeneratorRequest.proto_file', index=2,
number=15, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
extension_ranges=[],
serialized_start=101,
serialized_end=226,
)
_CODEGENERATORRESPONSE_FILE = descriptor.Descriptor(
name='File',
full_name='google.protobuf.compiler.CodeGeneratorResponse.File',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
descriptor.FieldDescriptor(
name='name', full_name='google.protobuf.compiler.CodeGeneratorResponse.File.name', index=0,
number=1, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=unicode("", "utf-8"),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
descriptor.FieldDescriptor(
name='insertion_point', full_name='google.protobuf.compiler.CodeGeneratorResponse.File.insertion_point', index=1,
number=2, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=unicode("", "utf-8"),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
descriptor.FieldDescriptor(
name='content', full_name='google.protobuf.compiler.CodeGeneratorResponse.File.content', index=2,
number=15, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=unicode("", "utf-8"),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
extension_ranges=[],
serialized_start=337,
serialized_end=399,
)
_CODEGENERATORRESPONSE = descriptor.Descriptor(
name='CodeGeneratorResponse',
full_name='google.protobuf.compiler.CodeGeneratorResponse',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
descriptor.FieldDescriptor(
name='error', full_name='google.protobuf.compiler.CodeGeneratorResponse.error', index=0,
number=1, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=unicode("", "utf-8"),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
descriptor.FieldDescriptor(
name='file', full_name='google.protobuf.compiler.CodeGeneratorResponse.file', index=1,
number=15, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[_CODEGENERATORRESPONSE_FILE, ],
enum_types=[
],
options=None,
is_extendable=False,
extension_ranges=[],
serialized_start=229,
serialized_end=399,
)
_CODEGENERATORREQUEST.fields_by_name['proto_file'].message_type = google.protobuf.descriptor_pb2._FILEDESCRIPTORPROTO
_CODEGENERATORRESPONSE_FILE.containing_type = _CODEGENERATORRESPONSE;
_CODEGENERATORRESPONSE.fields_by_name['file'].message_type = _CODEGENERATORRESPONSE_FILE
DESCRIPTOR.message_types_by_name['CodeGeneratorRequest'] = _CODEGENERATORREQUEST
DESCRIPTOR.message_types_by_name['CodeGeneratorResponse'] = _CODEGENERATORRESPONSE
class CodeGeneratorRequest(message.Message):
__metaclass__ = reflection.GeneratedProtocolMessageType
DESCRIPTOR = _CODEGENERATORREQUEST
# @@protoc_insertion_point(class_scope:google.protobuf.compiler.CodeGeneratorRequest)
class CodeGeneratorResponse(message.Message):
__metaclass__ = reflection.GeneratedProtocolMessageType
class File(message.Message):
__metaclass__ = reflection.GeneratedProtocolMessageType
DESCRIPTOR = _CODEGENERATORRESPONSE_FILE
# @@protoc_insertion_point(class_scope:google.protobuf.compiler.CodeGeneratorResponse.File)
DESCRIPTOR = _CODEGENERATORRESPONSE
# @@protoc_insertion_point(class_scope:google.protobuf.compiler.CodeGeneratorResponse)
# @@protoc_insertion_point(module_scope)

View File

@ -0,0 +1,598 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Descriptors essentially contain exactly the information found in a .proto
file, in types that make this information accessible in Python.
"""
__author__ = 'robinson@google.com (Will Robinson)'
from google.protobuf.internal import api_implementation
if api_implementation.Type() == 'cpp':
from google.protobuf.internal import cpp_message
class Error(Exception):
"""Base error for this module."""
class DescriptorBase(object):
"""Descriptors base class.
This class is the base of all descriptor classes. It provides common options
related functionaility.
Attributes:
has_options: True if the descriptor has non-default options. Usually it
is not necessary to read this -- just call GetOptions() which will
happily return the default instance. However, it's sometimes useful
for efficiency, and also useful inside the protobuf implementation to
avoid some bootstrapping issues.
"""
def __init__(self, options, options_class_name):
"""Initialize the descriptor given its options message and the name of the
class of the options message. The name of the class is required in case
the options message is None and has to be created.
"""
self._options = options
self._options_class_name = options_class_name
# Does this descriptor have non-default options?
self.has_options = options is not None
def GetOptions(self):
"""Retrieves descriptor options.
This method returns the options set or creates the default options for the
descriptor.
"""
if self._options:
return self._options
from google.protobuf import descriptor_pb2
try:
options_class = getattr(descriptor_pb2, self._options_class_name)
except AttributeError:
raise RuntimeError('Unknown options class name %s!' %
(self._options_class_name))
self._options = options_class()
return self._options
class _NestedDescriptorBase(DescriptorBase):
"""Common class for descriptors that can be nested."""
def __init__(self, options, options_class_name, name, full_name,
file, containing_type, serialized_start=None,
serialized_end=None):
"""Constructor.
Args:
options: Protocol message options or None
to use default message options.
options_class_name: (str) The class name of the above options.
name: (str) Name of this protocol message type.
full_name: (str) Fully-qualified name of this protocol message type,
which will include protocol "package" name and the name of any
enclosing types.
file: (FileDescriptor) Reference to file info.
containing_type: if provided, this is a nested descriptor, with this
descriptor as parent, otherwise None.
serialized_start: The start index (inclusive) in block in the
file.serialized_pb that describes this descriptor.
serialized_end: The end index (exclusive) in block in the
file.serialized_pb that describes this descriptor.
"""
super(_NestedDescriptorBase, self).__init__(
options, options_class_name)
self.name = name
# TODO(falk): Add function to calculate full_name instead of having it in
# memory?
self.full_name = full_name
self.file = file
self.containing_type = containing_type
self._serialized_start = serialized_start
self._serialized_end = serialized_end
def GetTopLevelContainingType(self):
"""Returns the root if this is a nested type, or itself if its the root."""
desc = self
while desc.containing_type is not None:
desc = desc.containing_type
return desc
def CopyToProto(self, proto):
"""Copies this to the matching proto in descriptor_pb2.
Args:
proto: An empty proto instance from descriptor_pb2.
Raises:
Error: If self couldnt be serialized, due to to few constructor arguments.
"""
if (self.file is not None and
self._serialized_start is not None and
self._serialized_end is not None):
proto.ParseFromString(self.file.serialized_pb[
self._serialized_start:self._serialized_end])
else:
raise Error('Descriptor does not contain serialization.')
class Descriptor(_NestedDescriptorBase):
"""Descriptor for a protocol message type.
A Descriptor instance has the following attributes:
name: (str) Name of this protocol message type.
full_name: (str) Fully-qualified name of this protocol message type,
which will include protocol "package" name and the name of any
enclosing types.
containing_type: (Descriptor) Reference to the descriptor of the
type containing us, or None if this is top-level.
fields: (list of FieldDescriptors) Field descriptors for all
fields in this type.
fields_by_number: (dict int -> FieldDescriptor) Same FieldDescriptor
objects as in |fields|, but indexed by "number" attribute in each
FieldDescriptor.
fields_by_name: (dict str -> FieldDescriptor) Same FieldDescriptor
objects as in |fields|, but indexed by "name" attribute in each
FieldDescriptor.
nested_types: (list of Descriptors) Descriptor references
for all protocol message types nested within this one.
nested_types_by_name: (dict str -> Descriptor) Same Descriptor
objects as in |nested_types|, but indexed by "name" attribute
in each Descriptor.
enum_types: (list of EnumDescriptors) EnumDescriptor references
for all enums contained within this type.
enum_types_by_name: (dict str ->EnumDescriptor) Same EnumDescriptor
objects as in |enum_types|, but indexed by "name" attribute
in each EnumDescriptor.
enum_values_by_name: (dict str -> EnumValueDescriptor) Dict mapping
from enum value name to EnumValueDescriptor for that value.
extensions: (list of FieldDescriptor) All extensions defined directly
within this message type (NOT within a nested type).
extensions_by_name: (dict, string -> FieldDescriptor) Same FieldDescriptor
objects as |extensions|, but indexed by "name" attribute of each
FieldDescriptor.
is_extendable: Does this type define any extension ranges?
options: (descriptor_pb2.MessageOptions) Protocol message options or None
to use default message options.
file: (FileDescriptor) Reference to file descriptor.
"""
def __init__(self, name, full_name, filename, containing_type, fields,
nested_types, enum_types, extensions, options=None,
is_extendable=True, extension_ranges=None, file=None,
serialized_start=None, serialized_end=None):
"""Arguments to __init__() are as described in the description
of Descriptor fields above.
Note that filename is an obsolete argument, that is not used anymore.
Please use file.name to access this as an attribute.
"""
super(Descriptor, self).__init__(
options, 'MessageOptions', name, full_name, file,
containing_type, serialized_start=serialized_start,
serialized_end=serialized_start)
# We have fields in addition to fields_by_name and fields_by_number,
# so that:
# 1. Clients can index fields by "order in which they're listed."
# 2. Clients can easily iterate over all fields with the terse
# syntax: for f in descriptor.fields: ...
self.fields = fields
for field in self.fields:
field.containing_type = self
self.fields_by_number = dict((f.number, f) for f in fields)
self.fields_by_name = dict((f.name, f) for f in fields)
self.nested_types = nested_types
self.nested_types_by_name = dict((t.name, t) for t in nested_types)
self.enum_types = enum_types
for enum_type in self.enum_types:
enum_type.containing_type = self
self.enum_types_by_name = dict((t.name, t) for t in enum_types)
self.enum_values_by_name = dict(
(v.name, v) for t in enum_types for v in t.values)
self.extensions = extensions
for extension in self.extensions:
extension.extension_scope = self
self.extensions_by_name = dict((f.name, f) for f in extensions)
self.is_extendable = is_extendable
self.extension_ranges = extension_ranges
self._serialized_start = serialized_start
self._serialized_end = serialized_end
def CopyToProto(self, proto):
"""Copies this to a descriptor_pb2.DescriptorProto.
Args:
proto: An empty descriptor_pb2.DescriptorProto.
"""
# This function is overriden to give a better doc comment.
super(Descriptor, self).CopyToProto(proto)
# TODO(robinson): We should have aggressive checking here,
# for example:
# * If you specify a repeated field, you should not be allowed
# to specify a default value.
# * [Other examples here as needed].
#
# TODO(robinson): for this and other *Descriptor classes, we
# might also want to lock things down aggressively (e.g.,
# prevent clients from setting the attributes). Having
# stronger invariants here in general will reduce the number
# of runtime checks we must do in reflection.py...
class FieldDescriptor(DescriptorBase):
"""Descriptor for a single field in a .proto file.
A FieldDescriptor instance has the following attriubtes:
name: (str) Name of this field, exactly as it appears in .proto.
full_name: (str) Name of this field, including containing scope. This is
particularly relevant for extensions.
index: (int) Dense, 0-indexed index giving the order that this
field textually appears within its message in the .proto file.
number: (int) Tag number declared for this field in the .proto file.
type: (One of the TYPE_* constants below) Declared type.
cpp_type: (One of the CPPTYPE_* constants below) C++ type used to
represent this field.
label: (One of the LABEL_* constants below) Tells whether this
field is optional, required, or repeated.
has_default_value: (bool) True if this field has a default value defined,
otherwise false.
default_value: (Varies) Default value of this field. Only
meaningful for non-repeated scalar fields. Repeated fields
should always set this to [], and non-repeated composite
fields should always set this to None.
containing_type: (Descriptor) Descriptor of the protocol message
type that contains this field. Set by the Descriptor constructor
if we're passed into one.
Somewhat confusingly, for extension fields, this is the
descriptor of the EXTENDED message, not the descriptor
of the message containing this field. (See is_extension and
extension_scope below).
message_type: (Descriptor) If a composite field, a descriptor
of the message type contained in this field. Otherwise, this is None.
enum_type: (EnumDescriptor) If this field contains an enum, a
descriptor of that enum. Otherwise, this is None.
is_extension: True iff this describes an extension field.
extension_scope: (Descriptor) Only meaningful if is_extension is True.
Gives the message that immediately contains this extension field.
Will be None iff we're a top-level (file-level) extension field.
options: (descriptor_pb2.FieldOptions) Protocol message field options or
None to use default field options.
"""
# Must be consistent with C++ FieldDescriptor::Type enum in
# descriptor.h.
#
# TODO(robinson): Find a way to eliminate this repetition.
TYPE_DOUBLE = 1
TYPE_FLOAT = 2
TYPE_INT64 = 3
TYPE_UINT64 = 4
TYPE_INT32 = 5
TYPE_FIXED64 = 6
TYPE_FIXED32 = 7
TYPE_BOOL = 8
TYPE_STRING = 9
TYPE_GROUP = 10
TYPE_MESSAGE = 11
TYPE_BYTES = 12
TYPE_UINT32 = 13
TYPE_ENUM = 14
TYPE_SFIXED32 = 15
TYPE_SFIXED64 = 16
TYPE_SINT32 = 17
TYPE_SINT64 = 18
MAX_TYPE = 18
# Must be consistent with C++ FieldDescriptor::CppType enum in
# descriptor.h.
#
# TODO(robinson): Find a way to eliminate this repetition.
CPPTYPE_INT32 = 1
CPPTYPE_INT64 = 2
CPPTYPE_UINT32 = 3
CPPTYPE_UINT64 = 4
CPPTYPE_DOUBLE = 5
CPPTYPE_FLOAT = 6
CPPTYPE_BOOL = 7
CPPTYPE_ENUM = 8
CPPTYPE_STRING = 9
CPPTYPE_MESSAGE = 10
MAX_CPPTYPE = 10
# Must be consistent with C++ FieldDescriptor::Label enum in
# descriptor.h.
#
# TODO(robinson): Find a way to eliminate this repetition.
LABEL_OPTIONAL = 1
LABEL_REQUIRED = 2
LABEL_REPEATED = 3
MAX_LABEL = 3
def __init__(self, name, full_name, index, number, type, cpp_type, label,
default_value, message_type, enum_type, containing_type,
is_extension, extension_scope, options=None,
has_default_value=True):
"""The arguments are as described in the description of FieldDescriptor
attributes above.
Note that containing_type may be None, and may be set later if necessary
(to deal with circular references between message types, for example).
Likewise for extension_scope.
"""
super(FieldDescriptor, self).__init__(options, 'FieldOptions')
self.name = name
self.full_name = full_name
self.index = index
self.number = number
self.type = type
self.cpp_type = cpp_type
self.label = label
self.has_default_value = has_default_value
self.default_value = default_value
self.containing_type = containing_type
self.message_type = message_type
self.enum_type = enum_type
self.is_extension = is_extension
self.extension_scope = extension_scope
if api_implementation.Type() == 'cpp':
if is_extension:
self._cdescriptor = cpp_message.GetExtensionDescriptor(full_name)
else:
self._cdescriptor = cpp_message.GetFieldDescriptor(full_name)
else:
self._cdescriptor = None
class EnumDescriptor(_NestedDescriptorBase):
"""Descriptor for an enum defined in a .proto file.
An EnumDescriptor instance has the following attributes:
name: (str) Name of the enum type.
full_name: (str) Full name of the type, including package name
and any enclosing type(s).
values: (list of EnumValueDescriptors) List of the values
in this enum.
values_by_name: (dict str -> EnumValueDescriptor) Same as |values|,
but indexed by the "name" field of each EnumValueDescriptor.
values_by_number: (dict int -> EnumValueDescriptor) Same as |values|,
but indexed by the "number" field of each EnumValueDescriptor.
containing_type: (Descriptor) Descriptor of the immediate containing
type of this enum, or None if this is an enum defined at the
top level in a .proto file. Set by Descriptor's constructor
if we're passed into one.
file: (FileDescriptor) Reference to file descriptor.
options: (descriptor_pb2.EnumOptions) Enum options message or
None to use default enum options.
"""
def __init__(self, name, full_name, filename, values,
containing_type=None, options=None, file=None,
serialized_start=None, serialized_end=None):
"""Arguments are as described in the attribute description above.
Note that filename is an obsolete argument, that is not used anymore.
Please use file.name to access this as an attribute.
"""
super(EnumDescriptor, self).__init__(
options, 'EnumOptions', name, full_name, file,
containing_type, serialized_start=serialized_start,
serialized_end=serialized_start)
self.values = values
for value in self.values:
value.type = self
self.values_by_name = dict((v.name, v) for v in values)
self.values_by_number = dict((v.number, v) for v in values)
self._serialized_start = serialized_start
self._serialized_end = serialized_end
def CopyToProto(self, proto):
"""Copies this to a descriptor_pb2.EnumDescriptorProto.
Args:
proto: An empty descriptor_pb2.EnumDescriptorProto.
"""
# This function is overriden to give a better doc comment.
super(EnumDescriptor, self).CopyToProto(proto)
class EnumValueDescriptor(DescriptorBase):
"""Descriptor for a single value within an enum.
name: (str) Name of this value.
index: (int) Dense, 0-indexed index giving the order that this
value appears textually within its enum in the .proto file.
number: (int) Actual number assigned to this enum value.
type: (EnumDescriptor) EnumDescriptor to which this value
belongs. Set by EnumDescriptor's constructor if we're
passed into one.
options: (descriptor_pb2.EnumValueOptions) Enum value options message or
None to use default enum value options options.
"""
def __init__(self, name, index, number, type=None, options=None):
"""Arguments are as described in the attribute description above."""
super(EnumValueDescriptor, self).__init__(options, 'EnumValueOptions')
self.name = name
self.index = index
self.number = number
self.type = type
class ServiceDescriptor(_NestedDescriptorBase):
"""Descriptor for a service.
name: (str) Name of the service.
full_name: (str) Full name of the service, including package name.
index: (int) 0-indexed index giving the order that this services
definition appears withing the .proto file.
methods: (list of MethodDescriptor) List of methods provided by this
service.
options: (descriptor_pb2.ServiceOptions) Service options message or
None to use default service options.
file: (FileDescriptor) Reference to file info.
"""
def __init__(self, name, full_name, index, methods, options=None, file=None,
serialized_start=None, serialized_end=None):
super(ServiceDescriptor, self).__init__(
options, 'ServiceOptions', name, full_name, file,
None, serialized_start=serialized_start,
serialized_end=serialized_end)
self.index = index
self.methods = methods
# Set the containing service for each method in this service.
for method in self.methods:
method.containing_service = self
def FindMethodByName(self, name):
"""Searches for the specified method, and returns its descriptor."""
for method in self.methods:
if name == method.name:
return method
return None
def CopyToProto(self, proto):
"""Copies this to a descriptor_pb2.ServiceDescriptorProto.
Args:
proto: An empty descriptor_pb2.ServiceDescriptorProto.
"""
# This function is overriden to give a better doc comment.
super(ServiceDescriptor, self).CopyToProto(proto)
class MethodDescriptor(DescriptorBase):
"""Descriptor for a method in a service.
name: (str) Name of the method within the service.
full_name: (str) Full name of method.
index: (int) 0-indexed index of the method inside the service.
containing_service: (ServiceDescriptor) The service that contains this
method.
input_type: The descriptor of the message that this method accepts.
output_type: The descriptor of the message that this method returns.
options: (descriptor_pb2.MethodOptions) Method options message or
None to use default method options.
"""
def __init__(self, name, full_name, index, containing_service,
input_type, output_type, options=None):
"""The arguments are as described in the description of MethodDescriptor
attributes above.
Note that containing_service may be None, and may be set later if necessary.
"""
super(MethodDescriptor, self).__init__(options, 'MethodOptions')
self.name = name
self.full_name = full_name
self.index = index
self.containing_service = containing_service
self.input_type = input_type
self.output_type = output_type
class FileDescriptor(DescriptorBase):
"""Descriptor for a file. Mimics the descriptor_pb2.FileDescriptorProto.
name: name of file, relative to root of source tree.
package: name of the package
serialized_pb: (str) Byte string of serialized
descriptor_pb2.FileDescriptorProto.
"""
def __init__(self, name, package, options=None, serialized_pb=None):
"""Constructor."""
super(FileDescriptor, self).__init__(options, 'FileOptions')
self.message_types_by_name = {}
self.name = name
self.package = package
self.serialized_pb = serialized_pb
if (api_implementation.Type() == 'cpp' and
self.serialized_pb is not None):
cpp_message.BuildFile(self.serialized_pb)
def CopyToProto(self, proto):
"""Copies this to a descriptor_pb2.FileDescriptorProto.
Args:
proto: An empty descriptor_pb2.FileDescriptorProto.
"""
proto.ParseFromString(self.serialized_pb)
def _ParseOptions(message, string):
"""Parses serialized options.
This helper function is used to parse serialized options in generated
proto2 files. It must not be used outside proto2.
"""
message.ParseFromString(string)
return message

File diff suppressed because one or more lines are too long

View File

@ -0,0 +1,64 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""
This module is the central entity that determines which implementation of the
API is used.
"""
__author__ = 'petar@google.com (Petar Petrov)'
import os
# This environment variable can be used to switch to a certain implementation
# of the Python API. Right now only 'python' and 'cpp' are valid values. Any
# other value will be ignored.
_implementation_type = os.getenv('PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION',
'python')
if _implementation_type != 'python':
# For now, by default use the pure-Python implementation.
# The code below checks if the C extension is available and
# uses it if it is available.
_implementation_type = 'cpp'
## Determine automatically which implementation to use.
#try:
# from google.protobuf.internal import cpp_message
# _implementation_type = 'cpp'
#except ImportError, e:
# _implementation_type = 'python'
# Usage of this function is discouraged. Clients shouldn't care which
# implementation of the API is in use. Note that there is no guarantee
# that differences between APIs will be maintained.
# Please don't use this function if possible.
def Type():
return _implementation_type

View File

@ -0,0 +1,259 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Contains container classes to represent different protocol buffer types.
This file defines container classes which represent categories of protocol
buffer field types which need extra maintenance. Currently these categories
are:
- Repeated scalar fields - These are all repeated fields which aren't
composite (e.g. they are of simple types like int32, string, etc).
- Repeated composite fields - Repeated fields which are composite. This
includes groups and nested messages.
"""
__author__ = 'petar@google.com (Petar Petrov)'
class BaseContainer(object):
"""Base container class."""
# Minimizes memory usage and disallows assignment to other attributes.
__slots__ = ['_message_listener', '_values']
def __init__(self, message_listener):
"""
Args:
message_listener: A MessageListener implementation.
The RepeatedScalarFieldContainer will call this object's
Modified() method when it is modified.
"""
self._message_listener = message_listener
self._values = []
def __getitem__(self, key):
"""Retrieves item by the specified key."""
return self._values[key]
def __len__(self):
"""Returns the number of elements in the container."""
return len(self._values)
def __ne__(self, other):
"""Checks if another instance isn't equal to this one."""
# The concrete classes should define __eq__.
return not self == other
def __hash__(self):
raise TypeError('unhashable object')
def __repr__(self):
return repr(self._values)
def sort(self, sort_function=cmp):
self._values.sort(sort_function)
class RepeatedScalarFieldContainer(BaseContainer):
"""Simple, type-checked, list-like container for holding repeated scalars."""
# Disallows assignment to other attributes.
__slots__ = ['_type_checker']
def __init__(self, message_listener, type_checker):
"""
Args:
message_listener: A MessageListener implementation.
The RepeatedScalarFieldContainer will call this object's
Modified() method when it is modified.
type_checker: A type_checkers.ValueChecker instance to run on elements
inserted into this container.
"""
super(RepeatedScalarFieldContainer, self).__init__(message_listener)
self._type_checker = type_checker
def append(self, value):
"""Appends an item to the list. Similar to list.append()."""
self._type_checker.CheckValue(value)
self._values.append(value)
if not self._message_listener.dirty:
self._message_listener.Modified()
def insert(self, key, value):
"""Inserts the item at the specified position. Similar to list.insert()."""
self._type_checker.CheckValue(value)
self._values.insert(key, value)
if not self._message_listener.dirty:
self._message_listener.Modified()
def extend(self, elem_seq):
"""Extends by appending the given sequence. Similar to list.extend()."""
if not elem_seq:
return
new_values = []
for elem in elem_seq:
self._type_checker.CheckValue(elem)
new_values.append(elem)
self._values.extend(new_values)
self._message_listener.Modified()
def MergeFrom(self, other):
"""Appends the contents of another repeated field of the same type to this
one. We do not check the types of the individual fields.
"""
self._values.extend(other._values)
self._message_listener.Modified()
def remove(self, elem):
"""Removes an item from the list. Similar to list.remove()."""
self._values.remove(elem)
self._message_listener.Modified()
def __setitem__(self, key, value):
"""Sets the item on the specified position."""
self._type_checker.CheckValue(value)
self._values[key] = value
self._message_listener.Modified()
def __getslice__(self, start, stop):
"""Retrieves the subset of items from between the specified indices."""
return self._values[start:stop]
def __setslice__(self, start, stop, values):
"""Sets the subset of items from between the specified indices."""
new_values = []
for value in values:
self._type_checker.CheckValue(value)
new_values.append(value)
self._values[start:stop] = new_values
self._message_listener.Modified()
def __delitem__(self, key):
"""Deletes the item at the specified position."""
del self._values[key]
self._message_listener.Modified()
def __delslice__(self, start, stop):
"""Deletes the subset of items from between the specified indices."""
del self._values[start:stop]
self._message_listener.Modified()
def __eq__(self, other):
"""Compares the current instance with another one."""
if self is other:
return True
# Special case for the same type which should be common and fast.
if isinstance(other, self.__class__):
return other._values == self._values
# We are presumably comparing against some other sequence type.
return other == self._values
class RepeatedCompositeFieldContainer(BaseContainer):
"""Simple, list-like container for holding repeated composite fields."""
# Disallows assignment to other attributes.
__slots__ = ['_message_descriptor']
def __init__(self, message_listener, message_descriptor):
"""
Note that we pass in a descriptor instead of the generated directly,
since at the time we construct a _RepeatedCompositeFieldContainer we
haven't yet necessarily initialized the type that will be contained in the
container.
Args:
message_listener: A MessageListener implementation.
The RepeatedCompositeFieldContainer will call this object's
Modified() method when it is modified.
message_descriptor: A Descriptor instance describing the protocol type
that should be present in this container. We'll use the
_concrete_class field of this descriptor when the client calls add().
"""
super(RepeatedCompositeFieldContainer, self).__init__(message_listener)
self._message_descriptor = message_descriptor
def add(self, **kwargs):
"""Adds a new element at the end of the list and returns it. Keyword
arguments may be used to initialize the element.
"""
new_element = self._message_descriptor._concrete_class(**kwargs)
new_element._SetListener(self._message_listener)
self._values.append(new_element)
if not self._message_listener.dirty:
self._message_listener.Modified()
return new_element
def extend(self, elem_seq):
"""Extends by appending the given sequence of elements of the same type
as this one, copying each individual message.
"""
message_class = self._message_descriptor._concrete_class
listener = self._message_listener
values = self._values
for message in elem_seq:
new_element = message_class()
new_element._SetListener(listener)
new_element.MergeFrom(message)
values.append(new_element)
listener.Modified()
def MergeFrom(self, other):
"""Appends the contents of another repeated field of the same type to this
one, copying each individual message.
"""
self.extend(other._values)
def __getslice__(self, start, stop):
"""Retrieves the subset of items from between the specified indices."""
return self._values[start:stop]
def __delitem__(self, key):
"""Deletes the item at the specified position."""
del self._values[key]
self._message_listener.Modified()
def __delslice__(self, start, stop):
"""Deletes the subset of items from between the specified indices."""
del self._values[start:stop]
self._message_listener.Modified()
def __eq__(self, other):
"""Compares the current instance with another one."""
if self is other:
return True
if not isinstance(other, self.__class__):
raise TypeError('Can only compare repeated composite fields against '
'other repeated composite fields.')
return self._values == other._values

View File

@ -0,0 +1,616 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Contains helper functions used to create protocol message classes from
Descriptor objects at runtime backed by the protocol buffer C++ API.
"""
__author__ = 'petar@google.com (Petar Petrov)'
import operator
from google.protobuf.internal import _net_proto2___python
from google.protobuf import message
_LABEL_REPEATED = _net_proto2___python.LABEL_REPEATED
_LABEL_OPTIONAL = _net_proto2___python.LABEL_OPTIONAL
_CPPTYPE_MESSAGE = _net_proto2___python.CPPTYPE_MESSAGE
_TYPE_MESSAGE = _net_proto2___python.TYPE_MESSAGE
def GetDescriptorPool():
"""Creates a new DescriptorPool C++ object."""
return _net_proto2___python.NewCDescriptorPool()
_pool = GetDescriptorPool()
def GetFieldDescriptor(full_field_name):
"""Searches for a field descriptor given a full field name."""
return _pool.FindFieldByName(full_field_name)
def BuildFile(content):
"""Registers a new proto file in the underlying C++ descriptor pool."""
_net_proto2___python.BuildFile(content)
def GetExtensionDescriptor(full_extension_name):
"""Searches for extension descriptor given a full field name."""
return _pool.FindExtensionByName(full_extension_name)
def NewCMessage(full_message_name):
"""Creates a new C++ protocol message by its name."""
return _net_proto2___python.NewCMessage(full_message_name)
def ScalarProperty(cdescriptor):
"""Returns a scalar property for the given descriptor."""
def Getter(self):
return self._cmsg.GetScalar(cdescriptor)
def Setter(self, value):
self._cmsg.SetScalar(cdescriptor, value)
return property(Getter, Setter)
def CompositeProperty(cdescriptor, message_type):
"""Returns a Python property the given composite field."""
def Getter(self):
sub_message = self._composite_fields.get(cdescriptor.name, None)
if sub_message is None:
cmessage = self._cmsg.NewSubMessage(cdescriptor)
sub_message = message_type._concrete_class(__cmessage=cmessage)
self._composite_fields[cdescriptor.name] = sub_message
return sub_message
return property(Getter)
class RepeatedScalarContainer(object):
"""Container for repeated scalar fields."""
__slots__ = ['_message', '_cfield_descriptor', '_cmsg']
def __init__(self, msg, cfield_descriptor):
self._message = msg
self._cmsg = msg._cmsg
self._cfield_descriptor = cfield_descriptor
def append(self, value):
self._cmsg.AddRepeatedScalar(
self._cfield_descriptor, value)
def extend(self, sequence):
for element in sequence:
self.append(element)
def insert(self, key, value):
values = self[slice(None, None, None)]
values.insert(key, value)
self._cmsg.AssignRepeatedScalar(self._cfield_descriptor, values)
def remove(self, value):
values = self[slice(None, None, None)]
values.remove(value)
self._cmsg.AssignRepeatedScalar(self._cfield_descriptor, values)
def __setitem__(self, key, value):
values = self[slice(None, None, None)]
values[key] = value
self._cmsg.AssignRepeatedScalar(self._cfield_descriptor, values)
def __getitem__(self, key):
return self._cmsg.GetRepeatedScalar(self._cfield_descriptor, key)
def __delitem__(self, key):
self._cmsg.DeleteRepeatedField(self._cfield_descriptor, key)
def __len__(self):
return len(self[slice(None, None, None)])
def __eq__(self, other):
if self is other:
return True
if not operator.isSequenceType(other):
raise TypeError(
'Can only compare repeated scalar fields against sequences.')
# We are presumably comparing against some other sequence type.
return other == self[slice(None, None, None)]
def __ne__(self, other):
return not self == other
def __hash__(self):
raise TypeError('unhashable object')
def sort(self, sort_function=cmp):
values = self[slice(None, None, None)]
values.sort(sort_function)
self._cmsg.AssignRepeatedScalar(self._cfield_descriptor, values)
def RepeatedScalarProperty(cdescriptor):
"""Returns a Python property the given repeated scalar field."""
def Getter(self):
container = self._composite_fields.get(cdescriptor.name, None)
if container is None:
container = RepeatedScalarContainer(self, cdescriptor)
self._composite_fields[cdescriptor.name] = container
return container
def Setter(self, new_value):
raise AttributeError('Assignment not allowed to repeated field '
'"%s" in protocol message object.' % cdescriptor.name)
doc = 'Magic attribute generated for "%s" proto field.' % cdescriptor.name
return property(Getter, Setter, doc=doc)
class RepeatedCompositeContainer(object):
"""Container for repeated composite fields."""
__slots__ = ['_message', '_subclass', '_cfield_descriptor', '_cmsg']
def __init__(self, msg, cfield_descriptor, subclass):
self._message = msg
self._cmsg = msg._cmsg
self._subclass = subclass
self._cfield_descriptor = cfield_descriptor
def add(self, **kwargs):
cmessage = self._cmsg.AddMessage(self._cfield_descriptor)
return self._subclass(__cmessage=cmessage, __owner=self._message, **kwargs)
def extend(self, elem_seq):
"""Extends by appending the given sequence of elements of the same type
as this one, copying each individual message.
"""
for message in elem_seq:
self.add().MergeFrom(message)
def MergeFrom(self, other):
for message in other[:]:
self.add().MergeFrom(message)
def __getitem__(self, key):
cmessages = self._cmsg.GetRepeatedMessage(
self._cfield_descriptor, key)
subclass = self._subclass
if not isinstance(cmessages, list):
return subclass(__cmessage=cmessages, __owner=self._message)
return [subclass(__cmessage=m, __owner=self._message) for m in cmessages]
def __delitem__(self, key):
self._cmsg.DeleteRepeatedField(
self._cfield_descriptor, key)
def __len__(self):
return self._cmsg.FieldLength(self._cfield_descriptor)
def __eq__(self, other):
"""Compares the current instance with another one."""
if self is other:
return True
if not isinstance(other, self.__class__):
raise TypeError('Can only compare repeated composite fields against '
'other repeated composite fields.')
messages = self[slice(None, None, None)]
other_messages = other[slice(None, None, None)]
return messages == other_messages
def __hash__(self):
raise TypeError('unhashable object')
def sort(self, sort_function=cmp):
messages = []
for index in range(len(self)):
# messages[i][0] is where the i-th element of the new array has to come
# from.
# messages[i][1] is where the i-th element of the old array has to go.
messages.append([index, 0, self[index]])
messages.sort(lambda x,y: sort_function(x[2], y[2]))
# Remember which position each elements has to move to.
for i in range(len(messages)):
messages[messages[i][0]][1] = i
# Apply the transposition.
for i in range(len(messages)):
from_position = messages[i][0]
if i == from_position:
continue
self._cmsg.SwapRepeatedFieldElements(
self._cfield_descriptor, i, from_position)
messages[messages[i][1]][0] = from_position
def RepeatedCompositeProperty(cdescriptor, message_type):
"""Returns a Python property for the given repeated composite field."""
def Getter(self):
container = self._composite_fields.get(cdescriptor.name, None)
if container is None:
container = RepeatedCompositeContainer(
self, cdescriptor, message_type._concrete_class)
self._composite_fields[cdescriptor.name] = container
return container
def Setter(self, new_value):
raise AttributeError('Assignment not allowed to repeated field '
'"%s" in protocol message object.' % cdescriptor.name)
doc = 'Magic attribute generated for "%s" proto field.' % cdescriptor.name
return property(Getter, Setter, doc=doc)
class ExtensionDict(object):
"""Extension dictionary added to each protocol message."""
def __init__(self, msg):
self._message = msg
self._cmsg = msg._cmsg
self._values = {}
def __setitem__(self, extension, value):
from google.protobuf import descriptor
if not isinstance(extension, descriptor.FieldDescriptor):
raise KeyError('Bad extension %r.' % (extension,))
cdescriptor = extension._cdescriptor
if (cdescriptor.label != _LABEL_OPTIONAL or
cdescriptor.cpp_type == _CPPTYPE_MESSAGE):
raise TypeError('Extension %r is repeated and/or a composite type.' % (
extension.full_name,))
self._cmsg.SetScalar(cdescriptor, value)
self._values[extension] = value
def __getitem__(self, extension):
from google.protobuf import descriptor
if not isinstance(extension, descriptor.FieldDescriptor):
raise KeyError('Bad extension %r.' % (extension,))
cdescriptor = extension._cdescriptor
if (cdescriptor.label != _LABEL_REPEATED and
cdescriptor.cpp_type != _CPPTYPE_MESSAGE):
return self._cmsg.GetScalar(cdescriptor)
ext = self._values.get(extension, None)
if ext is not None:
return ext
ext = self._CreateNewHandle(extension)
self._values[extension] = ext
return ext
def ClearExtension(self, extension):
from google.protobuf import descriptor
if not isinstance(extension, descriptor.FieldDescriptor):
raise KeyError('Bad extension %r.' % (extension,))
self._cmsg.ClearFieldByDescriptor(extension._cdescriptor)
if extension in self._values:
del self._values[extension]
def HasExtension(self, extension):
from google.protobuf import descriptor
if not isinstance(extension, descriptor.FieldDescriptor):
raise KeyError('Bad extension %r.' % (extension,))
return self._cmsg.HasFieldByDescriptor(extension._cdescriptor)
def _FindExtensionByName(self, name):
"""Tries to find a known extension with the specified name.
Args:
name: Extension full name.
Returns:
Extension field descriptor.
"""
return self._message._extensions_by_name.get(name, None)
def _CreateNewHandle(self, extension):
cdescriptor = extension._cdescriptor
if (cdescriptor.label != _LABEL_REPEATED and
cdescriptor.cpp_type == _CPPTYPE_MESSAGE):
cmessage = self._cmsg.NewSubMessage(cdescriptor)
return extension.message_type._concrete_class(__cmessage=cmessage)
if cdescriptor.label == _LABEL_REPEATED:
if cdescriptor.cpp_type == _CPPTYPE_MESSAGE:
return RepeatedCompositeContainer(
self._message, cdescriptor, extension.message_type._concrete_class)
else:
return RepeatedScalarContainer(self._message, cdescriptor)
# This shouldn't happen!
assert False
return None
def NewMessage(message_descriptor, dictionary):
"""Creates a new protocol message *class*."""
_AddClassAttributesForNestedExtensions(message_descriptor, dictionary)
_AddEnumValues(message_descriptor, dictionary)
_AddDescriptors(message_descriptor, dictionary)
def InitMessage(message_descriptor, cls):
"""Constructs a new message instance (called before instance's __init__)."""
cls._extensions_by_name = {}
_AddInitMethod(message_descriptor, cls)
_AddMessageMethods(message_descriptor, cls)
_AddPropertiesForExtensions(message_descriptor, cls)
def _AddDescriptors(message_descriptor, dictionary):
"""Sets up a new protocol message class dictionary.
Args:
message_descriptor: A Descriptor instance describing this message type.
dictionary: Class dictionary to which we'll add a '__slots__' entry.
"""
dictionary['__descriptors'] = {}
for field in message_descriptor.fields:
dictionary['__descriptors'][field.name] = GetFieldDescriptor(
field.full_name)
dictionary['__slots__'] = list(dictionary['__descriptors'].iterkeys()) + [
'_cmsg', '_owner', '_composite_fields', 'Extensions']
def _AddEnumValues(message_descriptor, dictionary):
"""Sets class-level attributes for all enum fields defined in this message.
Args:
message_descriptor: Descriptor object for this message type.
dictionary: Class dictionary that should be populated.
"""
for enum_type in message_descriptor.enum_types:
for enum_value in enum_type.values:
dictionary[enum_value.name] = enum_value.number
def _AddClassAttributesForNestedExtensions(message_descriptor, dictionary):
"""Adds class attributes for the nested extensions."""
extension_dict = message_descriptor.extensions_by_name
for extension_name, extension_field in extension_dict.iteritems():
assert extension_name not in dictionary
dictionary[extension_name] = extension_field
def _AddInitMethod(message_descriptor, cls):
"""Adds an __init__ method to cls."""
# Create and attach message field properties to the message class.
# This can be done just once per message class, since property setters and
# getters are passed the message instance.
# This makes message instantiation extremely fast, and at the same time it
# doesn't require the creation of property objects for each message instance,
# which saves a lot of memory.
for field in message_descriptor.fields:
field_cdescriptor = cls.__descriptors[field.name]
if field.label == _LABEL_REPEATED:
if field.cpp_type == _CPPTYPE_MESSAGE:
value = RepeatedCompositeProperty(field_cdescriptor, field.message_type)
else:
value = RepeatedScalarProperty(field_cdescriptor)
elif field.cpp_type == _CPPTYPE_MESSAGE:
value = CompositeProperty(field_cdescriptor, field.message_type)
else:
value = ScalarProperty(field_cdescriptor)
setattr(cls, field.name, value)
# Attach a constant with the field number.
constant_name = field.name.upper() + '_FIELD_NUMBER'
setattr(cls, constant_name, field.number)
def Init(self, **kwargs):
"""Message constructor."""
cmessage = kwargs.pop('__cmessage', None)
if cmessage is None:
self._cmsg = NewCMessage(message_descriptor.full_name)
else:
self._cmsg = cmessage
# Keep a reference to the owner, as the owner keeps a reference to the
# underlying protocol buffer message.
owner = kwargs.pop('__owner', None)
if owner is not None:
self._owner = owner
self.Extensions = ExtensionDict(self)
self._composite_fields = {}
for field_name, field_value in kwargs.iteritems():
field_cdescriptor = self.__descriptors.get(field_name, None)
if field_cdescriptor is None:
raise ValueError('Protocol message has no "%s" field.' % field_name)
if field_cdescriptor.label == _LABEL_REPEATED:
if field_cdescriptor.cpp_type == _CPPTYPE_MESSAGE:
for val in field_value:
getattr(self, field_name).add().MergeFrom(val)
else:
getattr(self, field_name).extend(field_value)
elif field_cdescriptor.cpp_type == _CPPTYPE_MESSAGE:
getattr(self, field_name).MergeFrom(field_value)
else:
setattr(self, field_name, field_value)
Init.__module__ = None
Init.__doc__ = None
cls.__init__ = Init
def _IsMessageSetExtension(field):
"""Checks if a field is a message set extension."""
return (field.is_extension and
field.containing_type.has_options and
field.containing_type.GetOptions().message_set_wire_format and
field.type == _TYPE_MESSAGE and
field.message_type == field.extension_scope and
field.label == _LABEL_OPTIONAL)
def _AddMessageMethods(message_descriptor, cls):
"""Adds the methods to a protocol message class."""
if message_descriptor.is_extendable:
def ClearExtension(self, extension):
self.Extensions.ClearExtension(extension)
def HasExtension(self, extension):
return self.Extensions.HasExtension(extension)
def HasField(self, field_name):
return self._cmsg.HasField(field_name)
def ClearField(self, field_name):
if field_name in self._composite_fields:
del self._composite_fields[field_name]
self._cmsg.ClearField(field_name)
def Clear(self):
return self._cmsg.Clear()
def IsInitialized(self, errors=None):
if self._cmsg.IsInitialized():
return True
if errors is not None:
errors.extend(self.FindInitializationErrors());
return False
def SerializeToString(self):
if not self.IsInitialized():
raise message.EncodeError(
'Message is missing required fields: ' +
','.join(self.FindInitializationErrors()))
return self._cmsg.SerializeToString()
def SerializePartialToString(self):
return self._cmsg.SerializePartialToString()
def ParseFromString(self, serialized):
self.Clear()
self.MergeFromString(serialized)
def MergeFromString(self, serialized):
byte_size = self._cmsg.MergeFromString(serialized)
if byte_size < 0:
raise message.DecodeError('Unable to merge from string.')
return byte_size
def MergeFrom(self, msg):
if not isinstance(msg, cls):
raise TypeError(
"Parameter to MergeFrom() must be instance of same class.")
self._cmsg.MergeFrom(msg._cmsg)
def CopyFrom(self, msg):
self._cmsg.CopyFrom(msg._cmsg)
def ByteSize(self):
return self._cmsg.ByteSize()
def SetInParent(self):
return self._cmsg.SetInParent()
def ListFields(self):
all_fields = []
field_list = self._cmsg.ListFields()
fields_by_name = cls.DESCRIPTOR.fields_by_name
for is_extension, field_name in field_list:
if is_extension:
extension = cls._extensions_by_name[field_name]
all_fields.append((extension, self.Extensions[extension]))
else:
field_descriptor = fields_by_name[field_name]
all_fields.append(
(field_descriptor, getattr(self, field_name)))
all_fields.sort(key=lambda item: item[0].number)
return all_fields
def FindInitializationErrors(self):
return self._cmsg.FindInitializationErrors()
def __str__(self):
return self._cmsg.DebugString()
def __eq__(self, other):
if self is other:
return True
if not isinstance(other, self.__class__):
return False
return self.ListFields() == other.ListFields()
def __ne__(self, other):
return not self == other
def __hash__(self):
raise TypeError('unhashable object')
def __unicode__(self):
return text_format.MessageToString(self, as_utf8=True).decode('utf-8')
# Attach the local methods to the message class.
for key, value in locals().copy().iteritems():
if key not in ('key', 'value', '__builtins__', '__name__', '__doc__'):
setattr(cls, key, value)
# Static methods:
def RegisterExtension(extension_handle):
extension_handle.containing_type = cls.DESCRIPTOR
cls._extensions_by_name[extension_handle.full_name] = extension_handle
if _IsMessageSetExtension(extension_handle):
# MessageSet extension. Also register under type name.
cls._extensions_by_name[
extension_handle.message_type.full_name] = extension_handle
cls.RegisterExtension = staticmethod(RegisterExtension)
def FromString(string):
msg = cls()
msg.MergeFromString(string)
return msg
cls.FromString = staticmethod(FromString)
def _AddPropertiesForExtensions(message_descriptor, cls):
"""Adds properties for all fields in this protocol message type."""
extension_dict = message_descriptor.extensions_by_name
for extension_name, extension_field in extension_dict.iteritems():
constant_name = extension_name.upper() + '_FIELD_NUMBER'
setattr(cls, constant_name, extension_field.number)

View File

@ -0,0 +1,714 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Code for decoding protocol buffer primitives.
This code is very similar to encoder.py -- read the docs for that module first.
A "decoder" is a function with the signature:
Decode(buffer, pos, end, message, field_dict)
The arguments are:
buffer: The string containing the encoded message.
pos: The current position in the string.
end: The position in the string where the current message ends. May be
less than len(buffer) if we're reading a sub-message.
message: The message object into which we're parsing.
field_dict: message._fields (avoids a hashtable lookup).
The decoder reads the field and stores it into field_dict, returning the new
buffer position. A decoder for a repeated field may proactively decode all of
the elements of that field, if they appear consecutively.
Note that decoders may throw any of the following:
IndexError: Indicates a truncated message.
struct.error: Unpacking of a fixed-width field failed.
message.DecodeError: Other errors.
Decoders are expected to raise an exception if they are called with pos > end.
This allows callers to be lax about bounds checking: it's fineto read past
"end" as long as you are sure that someone else will notice and throw an
exception later on.
Something up the call stack is expected to catch IndexError and struct.error
and convert them to message.DecodeError.
Decoders are constructed using decoder constructors with the signature:
MakeDecoder(field_number, is_repeated, is_packed, key, new_default)
The arguments are:
field_number: The field number of the field we want to decode.
is_repeated: Is the field a repeated field? (bool)
is_packed: Is the field a packed field? (bool)
key: The key to use when looking up the field within field_dict.
(This is actually the FieldDescriptor but nothing in this
file should depend on that.)
new_default: A function which takes a message object as a parameter and
returns a new instance of the default value for this field.
(This is called for repeated fields and sub-messages, when an
instance does not already exist.)
As with encoders, we define a decoder constructor for every type of field.
Then, for every field of every message class we construct an actual decoder.
That decoder goes into a dict indexed by tag, so when we decode a message
we repeatedly read a tag, look up the corresponding decoder, and invoke it.
"""
__author__ = 'kenton@google.com (Kenton Varda)'
import struct
from google.protobuf.internal import encoder
from google.protobuf.internal import wire_format
from google.protobuf import message
# This will overflow and thus become IEEE-754 "infinity". We would use
# "float('inf')" but it doesn't work on Windows pre-Python-2.6.
_POS_INF = 1e10000
_NEG_INF = -_POS_INF
_NAN = _POS_INF * 0
# This is not for optimization, but rather to avoid conflicts with local
# variables named "message".
_DecodeError = message.DecodeError
def _VarintDecoder(mask):
"""Return an encoder for a basic varint value (does not include tag).
Decoded values will be bitwise-anded with the given mask before being
returned, e.g. to limit them to 32 bits. The returned decoder does not
take the usual "end" parameter -- the caller is expected to do bounds checking
after the fact (often the caller can defer such checking until later). The
decoder returns a (value, new_pos) pair.
"""
local_ord = ord
def DecodeVarint(buffer, pos):
result = 0
shift = 0
while 1:
b = local_ord(buffer[pos])
result |= ((b & 0x7f) << shift)
pos += 1
if not (b & 0x80):
result &= mask
return (result, pos)
shift += 7
if shift >= 64:
raise _DecodeError('Too many bytes when decoding varint.')
return DecodeVarint
def _SignedVarintDecoder(mask):
"""Like _VarintDecoder() but decodes signed values."""
local_ord = ord
def DecodeVarint(buffer, pos):
result = 0
shift = 0
while 1:
b = local_ord(buffer[pos])
result |= ((b & 0x7f) << shift)
pos += 1
if not (b & 0x80):
if result > 0x7fffffffffffffff:
result -= (1 << 64)
result |= ~mask
else:
result &= mask
return (result, pos)
shift += 7
if shift >= 64:
raise _DecodeError('Too many bytes when decoding varint.')
return DecodeVarint
_DecodeVarint = _VarintDecoder((1 << 64) - 1)
_DecodeSignedVarint = _SignedVarintDecoder((1 << 64) - 1)
# Use these versions for values which must be limited to 32 bits.
_DecodeVarint32 = _VarintDecoder((1 << 32) - 1)
_DecodeSignedVarint32 = _SignedVarintDecoder((1 << 32) - 1)
def ReadTag(buffer, pos):
"""Read a tag from the buffer, and return a (tag_bytes, new_pos) tuple.
We return the raw bytes of the tag rather than decoding them. The raw
bytes can then be used to look up the proper decoder. This effectively allows
us to trade some work that would be done in pure-python (decoding a varint)
for work that is done in C (searching for a byte string in a hash table).
In a low-level language it would be much cheaper to decode the varint and
use that, but not in Python.
"""
start = pos
while ord(buffer[pos]) & 0x80:
pos += 1
pos += 1
return (buffer[start:pos], pos)
# --------------------------------------------------------------------
def _SimpleDecoder(wire_type, decode_value):
"""Return a constructor for a decoder for fields of a particular type.
Args:
wire_type: The field's wire type.
decode_value: A function which decodes an individual value, e.g.
_DecodeVarint()
"""
def SpecificDecoder(field_number, is_repeated, is_packed, key, new_default):
if is_packed:
local_DecodeVarint = _DecodeVarint
def DecodePackedField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
(endpoint, pos) = local_DecodeVarint(buffer, pos)
endpoint += pos
if endpoint > end:
raise _DecodeError('Truncated message.')
while pos < endpoint:
(element, pos) = decode_value(buffer, pos)
value.append(element)
if pos > endpoint:
del value[-1] # Discard corrupt value.
raise _DecodeError('Packed element was truncated.')
return pos
return DecodePackedField
elif is_repeated:
tag_bytes = encoder.TagBytes(field_number, wire_type)
tag_len = len(tag_bytes)
def DecodeRepeatedField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
while 1:
(element, new_pos) = decode_value(buffer, pos)
value.append(element)
# Predict that the next tag is another copy of the same repeated
# field.
pos = new_pos + tag_len
if buffer[new_pos:pos] != tag_bytes or new_pos >= end:
# Prediction failed. Return.
if new_pos > end:
raise _DecodeError('Truncated message.')
return new_pos
return DecodeRepeatedField
else:
def DecodeField(buffer, pos, end, message, field_dict):
(field_dict[key], pos) = decode_value(buffer, pos)
if pos > end:
del field_dict[key] # Discard corrupt value.
raise _DecodeError('Truncated message.')
return pos
return DecodeField
return SpecificDecoder
def _ModifiedDecoder(wire_type, decode_value, modify_value):
"""Like SimpleDecoder but additionally invokes modify_value on every value
before storing it. Usually modify_value is ZigZagDecode.
"""
# Reusing _SimpleDecoder is slightly slower than copying a bunch of code, but
# not enough to make a significant difference.
def InnerDecode(buffer, pos):
(result, new_pos) = decode_value(buffer, pos)
return (modify_value(result), new_pos)
return _SimpleDecoder(wire_type, InnerDecode)
def _StructPackDecoder(wire_type, format):
"""Return a constructor for a decoder for a fixed-width field.
Args:
wire_type: The field's wire type.
format: The format string to pass to struct.unpack().
"""
value_size = struct.calcsize(format)
local_unpack = struct.unpack
# Reusing _SimpleDecoder is slightly slower than copying a bunch of code, but
# not enough to make a significant difference.
# Note that we expect someone up-stack to catch struct.error and convert
# it to _DecodeError -- this way we don't have to set up exception-
# handling blocks every time we parse one value.
def InnerDecode(buffer, pos):
new_pos = pos + value_size
result = local_unpack(format, buffer[pos:new_pos])[0]
return (result, new_pos)
return _SimpleDecoder(wire_type, InnerDecode)
def _FloatDecoder():
"""Returns a decoder for a float field.
This code works around a bug in struct.unpack for non-finite 32-bit
floating-point values.
"""
local_unpack = struct.unpack
def InnerDecode(buffer, pos):
# We expect a 32-bit value in little-endian byte order. Bit 1 is the sign
# bit, bits 2-9 represent the exponent, and bits 10-32 are the significand.
new_pos = pos + 4
float_bytes = buffer[pos:new_pos]
# If this value has all its exponent bits set, then it's non-finite.
# In Python 2.4, struct.unpack will convert it to a finite 64-bit value.
# To avoid that, we parse it specially.
if ((float_bytes[3] in '\x7F\xFF')
and (float_bytes[2] >= '\x80')):
# If at least one significand bit is set...
if float_bytes[0:3] != '\x00\x00\x80':
return (_NAN, new_pos)
# If sign bit is set...
if float_bytes[3] == '\xFF':
return (_NEG_INF, new_pos)
return (_POS_INF, new_pos)
# Note that we expect someone up-stack to catch struct.error and convert
# it to _DecodeError -- this way we don't have to set up exception-
# handling blocks every time we parse one value.
result = local_unpack('<f', float_bytes)[0]
return (result, new_pos)
return _SimpleDecoder(wire_format.WIRETYPE_FIXED32, InnerDecode)
def _DoubleDecoder():
"""Returns a decoder for a double field.
This code works around a bug in struct.unpack for not-a-number.
"""
local_unpack = struct.unpack
def InnerDecode(buffer, pos):
# We expect a 64-bit value in little-endian byte order. Bit 1 is the sign
# bit, bits 2-12 represent the exponent, and bits 13-64 are the significand.
new_pos = pos + 8
double_bytes = buffer[pos:new_pos]
# If this value has all its exponent bits set and at least one significand
# bit set, it's not a number. In Python 2.4, struct.unpack will treat it
# as inf or -inf. To avoid that, we treat it specially.
if ((double_bytes[7] in '\x7F\xFF')
and (double_bytes[6] >= '\xF0')
and (double_bytes[0:7] != '\x00\x00\x00\x00\x00\x00\xF0')):
return (_NAN, new_pos)
# Note that we expect someone up-stack to catch struct.error and convert
# it to _DecodeError -- this way we don't have to set up exception-
# handling blocks every time we parse one value.
result = local_unpack('<d', double_bytes)[0]
return (result, new_pos)
return _SimpleDecoder(wire_format.WIRETYPE_FIXED64, InnerDecode)
# --------------------------------------------------------------------
Int32Decoder = EnumDecoder = _SimpleDecoder(
wire_format.WIRETYPE_VARINT, _DecodeSignedVarint32)
Int64Decoder = _SimpleDecoder(
wire_format.WIRETYPE_VARINT, _DecodeSignedVarint)
UInt32Decoder = _SimpleDecoder(wire_format.WIRETYPE_VARINT, _DecodeVarint32)
UInt64Decoder = _SimpleDecoder(wire_format.WIRETYPE_VARINT, _DecodeVarint)
SInt32Decoder = _ModifiedDecoder(
wire_format.WIRETYPE_VARINT, _DecodeVarint32, wire_format.ZigZagDecode)
SInt64Decoder = _ModifiedDecoder(
wire_format.WIRETYPE_VARINT, _DecodeVarint, wire_format.ZigZagDecode)
# Note that Python conveniently guarantees that when using the '<' prefix on
# formats, they will also have the same size across all platforms (as opposed
# to without the prefix, where their sizes depend on the C compiler's basic
# type sizes).
Fixed32Decoder = _StructPackDecoder(wire_format.WIRETYPE_FIXED32, '<I')
Fixed64Decoder = _StructPackDecoder(wire_format.WIRETYPE_FIXED64, '<Q')
SFixed32Decoder = _StructPackDecoder(wire_format.WIRETYPE_FIXED32, '<i')
SFixed64Decoder = _StructPackDecoder(wire_format.WIRETYPE_FIXED64, '<q')
FloatDecoder = _FloatDecoder()
DoubleDecoder = _DoubleDecoder()
BoolDecoder = _ModifiedDecoder(
wire_format.WIRETYPE_VARINT, _DecodeVarint, bool)
def StringDecoder(field_number, is_repeated, is_packed, key, new_default):
"""Returns a decoder for a string field."""
local_DecodeVarint = _DecodeVarint
local_unicode = unicode
assert not is_packed
if is_repeated:
tag_bytes = encoder.TagBytes(field_number,
wire_format.WIRETYPE_LENGTH_DELIMITED)
tag_len = len(tag_bytes)
def DecodeRepeatedField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
while 1:
(size, pos) = local_DecodeVarint(buffer, pos)
new_pos = pos + size
if new_pos > end:
raise _DecodeError('Truncated string.')
value.append(local_unicode(buffer[pos:new_pos], 'utf-8'))
# Predict that the next tag is another copy of the same repeated field.
pos = new_pos + tag_len
if buffer[new_pos:pos] != tag_bytes or new_pos == end:
# Prediction failed. Return.
return new_pos
return DecodeRepeatedField
else:
def DecodeField(buffer, pos, end, message, field_dict):
(size, pos) = local_DecodeVarint(buffer, pos)
new_pos = pos + size
if new_pos > end:
raise _DecodeError('Truncated string.')
field_dict[key] = local_unicode(buffer[pos:new_pos], 'utf-8')
return new_pos
return DecodeField
def BytesDecoder(field_number, is_repeated, is_packed, key, new_default):
"""Returns a decoder for a bytes field."""
local_DecodeVarint = _DecodeVarint
assert not is_packed
if is_repeated:
tag_bytes = encoder.TagBytes(field_number,
wire_format.WIRETYPE_LENGTH_DELIMITED)
tag_len = len(tag_bytes)
def DecodeRepeatedField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
while 1:
(size, pos) = local_DecodeVarint(buffer, pos)
new_pos = pos + size
if new_pos > end:
raise _DecodeError('Truncated string.')
value.append(buffer[pos:new_pos])
# Predict that the next tag is another copy of the same repeated field.
pos = new_pos + tag_len
if buffer[new_pos:pos] != tag_bytes or new_pos == end:
# Prediction failed. Return.
return new_pos
return DecodeRepeatedField
else:
def DecodeField(buffer, pos, end, message, field_dict):
(size, pos) = local_DecodeVarint(buffer, pos)
new_pos = pos + size
if new_pos > end:
raise _DecodeError('Truncated string.')
field_dict[key] = buffer[pos:new_pos]
return new_pos
return DecodeField
def GroupDecoder(field_number, is_repeated, is_packed, key, new_default):
"""Returns a decoder for a group field."""
end_tag_bytes = encoder.TagBytes(field_number,
wire_format.WIRETYPE_END_GROUP)
end_tag_len = len(end_tag_bytes)
assert not is_packed
if is_repeated:
tag_bytes = encoder.TagBytes(field_number,
wire_format.WIRETYPE_START_GROUP)
tag_len = len(tag_bytes)
def DecodeRepeatedField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
while 1:
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
# Read sub-message.
pos = value.add()._InternalParse(buffer, pos, end)
# Read end tag.
new_pos = pos+end_tag_len
if buffer[pos:new_pos] != end_tag_bytes or new_pos > end:
raise _DecodeError('Missing group end tag.')
# Predict that the next tag is another copy of the same repeated field.
pos = new_pos + tag_len
if buffer[new_pos:pos] != tag_bytes or new_pos == end:
# Prediction failed. Return.
return new_pos
return DecodeRepeatedField
else:
def DecodeField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
# Read sub-message.
pos = value._InternalParse(buffer, pos, end)
# Read end tag.
new_pos = pos+end_tag_len
if buffer[pos:new_pos] != end_tag_bytes or new_pos > end:
raise _DecodeError('Missing group end tag.')
return new_pos
return DecodeField
def MessageDecoder(field_number, is_repeated, is_packed, key, new_default):
"""Returns a decoder for a message field."""
local_DecodeVarint = _DecodeVarint
assert not is_packed
if is_repeated:
tag_bytes = encoder.TagBytes(field_number,
wire_format.WIRETYPE_LENGTH_DELIMITED)
tag_len = len(tag_bytes)
def DecodeRepeatedField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
while 1:
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
# Read length.
(size, pos) = local_DecodeVarint(buffer, pos)
new_pos = pos + size
if new_pos > end:
raise _DecodeError('Truncated message.')
# Read sub-message.
if value.add()._InternalParse(buffer, pos, new_pos) != new_pos:
# The only reason _InternalParse would return early is if it
# encountered an end-group tag.
raise _DecodeError('Unexpected end-group tag.')
# Predict that the next tag is another copy of the same repeated field.
pos = new_pos + tag_len
if buffer[new_pos:pos] != tag_bytes or new_pos == end:
# Prediction failed. Return.
return new_pos
return DecodeRepeatedField
else:
def DecodeField(buffer, pos, end, message, field_dict):
value = field_dict.get(key)
if value is None:
value = field_dict.setdefault(key, new_default(message))
# Read length.
(size, pos) = local_DecodeVarint(buffer, pos)
new_pos = pos + size
if new_pos > end:
raise _DecodeError('Truncated message.')
# Read sub-message.
if value._InternalParse(buffer, pos, new_pos) != new_pos:
# The only reason _InternalParse would return early is if it encountered
# an end-group tag.
raise _DecodeError('Unexpected end-group tag.')
return new_pos
return DecodeField
# --------------------------------------------------------------------
MESSAGE_SET_ITEM_TAG = encoder.TagBytes(1, wire_format.WIRETYPE_START_GROUP)
def MessageSetItemDecoder(extensions_by_number):
"""Returns a decoder for a MessageSet item.
The parameter is the _extensions_by_number map for the message class.
The message set message looks like this:
message MessageSet {
repeated group Item = 1 {
required int32 type_id = 2;
required string message = 3;
}
}
"""
type_id_tag_bytes = encoder.TagBytes(2, wire_format.WIRETYPE_VARINT)
message_tag_bytes = encoder.TagBytes(3, wire_format.WIRETYPE_LENGTH_DELIMITED)
item_end_tag_bytes = encoder.TagBytes(1, wire_format.WIRETYPE_END_GROUP)
local_ReadTag = ReadTag
local_DecodeVarint = _DecodeVarint
local_SkipField = SkipField
def DecodeItem(buffer, pos, end, message, field_dict):
type_id = -1
message_start = -1
message_end = -1
# Technically, type_id and message can appear in any order, so we need
# a little loop here.
while 1:
(tag_bytes, pos) = local_ReadTag(buffer, pos)
if tag_bytes == type_id_tag_bytes:
(type_id, pos) = local_DecodeVarint(buffer, pos)
elif tag_bytes == message_tag_bytes:
(size, message_start) = local_DecodeVarint(buffer, pos)
pos = message_end = message_start + size
elif tag_bytes == item_end_tag_bytes:
break
else:
pos = SkipField(buffer, pos, end, tag_bytes)
if pos == -1:
raise _DecodeError('Missing group end tag.')
if pos > end:
raise _DecodeError('Truncated message.')
if type_id == -1:
raise _DecodeError('MessageSet item missing type_id.')
if message_start == -1:
raise _DecodeError('MessageSet item missing message.')
extension = extensions_by_number.get(type_id)
if extension is not None:
value = field_dict.get(extension)
if value is None:
value = field_dict.setdefault(
extension, extension.message_type._concrete_class())
if value._InternalParse(buffer, message_start,message_end) != message_end:
# The only reason _InternalParse would return early is if it encountered
# an end-group tag.
raise _DecodeError('Unexpected end-group tag.')
return pos
return DecodeItem
# --------------------------------------------------------------------
# Optimization is not as heavy here because calls to SkipField() are rare,
# except for handling end-group tags.
def _SkipVarint(buffer, pos, end):
"""Skip a varint value. Returns the new position."""
while ord(buffer[pos]) & 0x80:
pos += 1
pos += 1
if pos > end:
raise _DecodeError('Truncated message.')
return pos
def _SkipFixed64(buffer, pos, end):
"""Skip a fixed64 value. Returns the new position."""
pos += 8
if pos > end:
raise _DecodeError('Truncated message.')
return pos
def _SkipLengthDelimited(buffer, pos, end):
"""Skip a length-delimited value. Returns the new position."""
(size, pos) = _DecodeVarint(buffer, pos)
pos += size
if pos > end:
raise _DecodeError('Truncated message.')
return pos
def _SkipGroup(buffer, pos, end):
"""Skip sub-group. Returns the new position."""
while 1:
(tag_bytes, pos) = ReadTag(buffer, pos)
new_pos = SkipField(buffer, pos, end, tag_bytes)
if new_pos == -1:
return pos
pos = new_pos
def _EndGroup(buffer, pos, end):
"""Skipping an END_GROUP tag returns -1 to tell the parent loop to break."""
return -1
def _SkipFixed32(buffer, pos, end):
"""Skip a fixed32 value. Returns the new position."""
pos += 4
if pos > end:
raise _DecodeError('Truncated message.')
return pos
def _RaiseInvalidWireType(buffer, pos, end):
"""Skip function for unknown wire types. Raises an exception."""
raise _DecodeError('Tag had invalid wire type.')
def _FieldSkipper():
"""Constructs the SkipField function."""
WIRETYPE_TO_SKIPPER = [
_SkipVarint,
_SkipFixed64,
_SkipLengthDelimited,
_SkipGroup,
_EndGroup,
_SkipFixed32,
_RaiseInvalidWireType,
_RaiseInvalidWireType,
]
wiretype_mask = wire_format.TAG_TYPE_MASK
local_ord = ord
def SkipField(buffer, pos, end, tag_bytes):
"""Skips a field with the specified tag.
|pos| should point to the byte immediately after the tag.
Returns:
The new position (after the tag value), or -1 if the tag is an end-group
tag (in which case the calling loop should break).
"""
# The wire type is always in the first byte since varints are little-endian.
wire_type = local_ord(tag_bytes[0]) & wiretype_mask
return WIRETYPE_TO_SKIPPER[wire_type](buffer, pos, end)
return SkipField
SkipField = _FieldSkipper()

View File

@ -0,0 +1,769 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Code for encoding protocol message primitives.
Contains the logic for encoding every logical protocol field type
into one of the 5 physical wire types.
This code is designed to push the Python interpreter's performance to the
limits.
The basic idea is that at startup time, for every field (i.e. every
FieldDescriptor) we construct two functions: a "sizer" and an "encoder". The
sizer takes a value of this field's type and computes its byte size. The
encoder takes a writer function and a value. It encodes the value into byte
strings and invokes the writer function to write those strings. Typically the
writer function is the write() method of a cStringIO.
We try to do as much work as possible when constructing the writer and the
sizer rather than when calling them. In particular:
* We copy any needed global functions to local variables, so that we do not need
to do costly global table lookups at runtime.
* Similarly, we try to do any attribute lookups at startup time if possible.
* Every field's tag is encoded to bytes at startup, since it can't change at
runtime.
* Whatever component of the field size we can compute at startup, we do.
* We *avoid* sharing code if doing so would make the code slower and not sharing
does not burden us too much. For example, encoders for repeated fields do
not just call the encoders for singular fields in a loop because this would
add an extra function call overhead for every loop iteration; instead, we
manually inline the single-value encoder into the loop.
* If a Python function lacks a return statement, Python actually generates
instructions to pop the result of the last statement off the stack, push
None onto the stack, and then return that. If we really don't care what
value is returned, then we can save two instructions by returning the
result of the last statement. It looks funny but it helps.
* We assume that type and bounds checking has happened at a higher level.
"""
__author__ = 'kenton@google.com (Kenton Varda)'
import struct
from google.protobuf.internal import wire_format
# This will overflow and thus become IEEE-754 "infinity". We would use
# "float('inf')" but it doesn't work on Windows pre-Python-2.6.
_POS_INF = 1e10000
_NEG_INF = -_POS_INF
def _VarintSize(value):
"""Compute the size of a varint value."""
if value <= 0x7f: return 1
if value <= 0x3fff: return 2
if value <= 0x1fffff: return 3
if value <= 0xfffffff: return 4
if value <= 0x7ffffffff: return 5
if value <= 0x3ffffffffff: return 6
if value <= 0x1ffffffffffff: return 7
if value <= 0xffffffffffffff: return 8
if value <= 0x7fffffffffffffff: return 9
return 10
def _SignedVarintSize(value):
"""Compute the size of a signed varint value."""
if value < 0: return 10
if value <= 0x7f: return 1
if value <= 0x3fff: return 2
if value <= 0x1fffff: return 3
if value <= 0xfffffff: return 4
if value <= 0x7ffffffff: return 5
if value <= 0x3ffffffffff: return 6
if value <= 0x1ffffffffffff: return 7
if value <= 0xffffffffffffff: return 8
if value <= 0x7fffffffffffffff: return 9
return 10
def _TagSize(field_number):
"""Returns the number of bytes required to serialize a tag with this field
number."""
# Just pass in type 0, since the type won't affect the tag+type size.
return _VarintSize(wire_format.PackTag(field_number, 0))
# --------------------------------------------------------------------
# In this section we define some generic sizers. Each of these functions
# takes parameters specific to a particular field type, e.g. int32 or fixed64.
# It returns another function which in turn takes parameters specific to a
# particular field, e.g. the field number and whether it is repeated or packed.
# Look at the next section to see how these are used.
def _SimpleSizer(compute_value_size):
"""A sizer which uses the function compute_value_size to compute the size of
each value. Typically compute_value_size is _VarintSize."""
def SpecificSizer(field_number, is_repeated, is_packed):
tag_size = _TagSize(field_number)
if is_packed:
local_VarintSize = _VarintSize
def PackedFieldSize(value):
result = 0
for element in value:
result += compute_value_size(element)
return result + local_VarintSize(result) + tag_size
return PackedFieldSize
elif is_repeated:
def RepeatedFieldSize(value):
result = tag_size * len(value)
for element in value:
result += compute_value_size(element)
return result
return RepeatedFieldSize
else:
def FieldSize(value):
return tag_size + compute_value_size(value)
return FieldSize
return SpecificSizer
def _ModifiedSizer(compute_value_size, modify_value):
"""Like SimpleSizer, but modify_value is invoked on each value before it is
passed to compute_value_size. modify_value is typically ZigZagEncode."""
def SpecificSizer(field_number, is_repeated, is_packed):
tag_size = _TagSize(field_number)
if is_packed:
local_VarintSize = _VarintSize
def PackedFieldSize(value):
result = 0
for element in value:
result += compute_value_size(modify_value(element))
return result + local_VarintSize(result) + tag_size
return PackedFieldSize
elif is_repeated:
def RepeatedFieldSize(value):
result = tag_size * len(value)
for element in value:
result += compute_value_size(modify_value(element))
return result
return RepeatedFieldSize
else:
def FieldSize(value):
return tag_size + compute_value_size(modify_value(value))
return FieldSize
return SpecificSizer
def _FixedSizer(value_size):
"""Like _SimpleSizer except for a fixed-size field. The input is the size
of one value."""
def SpecificSizer(field_number, is_repeated, is_packed):
tag_size = _TagSize(field_number)
if is_packed:
local_VarintSize = _VarintSize
def PackedFieldSize(value):
result = len(value) * value_size
return result + local_VarintSize(result) + tag_size
return PackedFieldSize
elif is_repeated:
element_size = value_size + tag_size
def RepeatedFieldSize(value):
return len(value) * element_size
return RepeatedFieldSize
else:
field_size = value_size + tag_size
def FieldSize(value):
return field_size
return FieldSize
return SpecificSizer
# ====================================================================
# Here we declare a sizer constructor for each field type. Each "sizer
# constructor" is a function that takes (field_number, is_repeated, is_packed)
# as parameters and returns a sizer, which in turn takes a field value as
# a parameter and returns its encoded size.
Int32Sizer = Int64Sizer = EnumSizer = _SimpleSizer(_SignedVarintSize)
UInt32Sizer = UInt64Sizer = _SimpleSizer(_VarintSize)
SInt32Sizer = SInt64Sizer = _ModifiedSizer(
_SignedVarintSize, wire_format.ZigZagEncode)
Fixed32Sizer = SFixed32Sizer = FloatSizer = _FixedSizer(4)
Fixed64Sizer = SFixed64Sizer = DoubleSizer = _FixedSizer(8)
BoolSizer = _FixedSizer(1)
def StringSizer(field_number, is_repeated, is_packed):
"""Returns a sizer for a string field."""
tag_size = _TagSize(field_number)
local_VarintSize = _VarintSize
local_len = len
assert not is_packed
if is_repeated:
def RepeatedFieldSize(value):
result = tag_size * len(value)
for element in value:
l = local_len(element.encode('utf-8'))
result += local_VarintSize(l) + l
return result
return RepeatedFieldSize
else:
def FieldSize(value):
l = local_len(value.encode('utf-8'))
return tag_size + local_VarintSize(l) + l
return FieldSize
def BytesSizer(field_number, is_repeated, is_packed):
"""Returns a sizer for a bytes field."""
tag_size = _TagSize(field_number)
local_VarintSize = _VarintSize
local_len = len
assert not is_packed
if is_repeated:
def RepeatedFieldSize(value):
result = tag_size * len(value)
for element in value:
l = local_len(element)
result += local_VarintSize(l) + l
return result
return RepeatedFieldSize
else:
def FieldSize(value):
l = local_len(value)
return tag_size + local_VarintSize(l) + l
return FieldSize
def GroupSizer(field_number, is_repeated, is_packed):
"""Returns a sizer for a group field."""
tag_size = _TagSize(field_number) * 2
assert not is_packed
if is_repeated:
def RepeatedFieldSize(value):
result = tag_size * len(value)
for element in value:
result += element.ByteSize()
return result
return RepeatedFieldSize
else:
def FieldSize(value):
return tag_size + value.ByteSize()
return FieldSize
def MessageSizer(field_number, is_repeated, is_packed):
"""Returns a sizer for a message field."""
tag_size = _TagSize(field_number)
local_VarintSize = _VarintSize
assert not is_packed
if is_repeated:
def RepeatedFieldSize(value):
result = tag_size * len(value)
for element in value:
l = element.ByteSize()
result += local_VarintSize(l) + l
return result
return RepeatedFieldSize
else:
def FieldSize(value):
l = value.ByteSize()
return tag_size + local_VarintSize(l) + l
return FieldSize
# --------------------------------------------------------------------
# MessageSet is special.
def MessageSetItemSizer(field_number):
"""Returns a sizer for extensions of MessageSet.
The message set message looks like this:
message MessageSet {
repeated group Item = 1 {
required int32 type_id = 2;
required string message = 3;
}
}
"""
static_size = (_TagSize(1) * 2 + _TagSize(2) + _VarintSize(field_number) +
_TagSize(3))
local_VarintSize = _VarintSize
def FieldSize(value):
l = value.ByteSize()
return static_size + local_VarintSize(l) + l
return FieldSize
# ====================================================================
# Encoders!
def _VarintEncoder():
"""Return an encoder for a basic varint value (does not include tag)."""
local_chr = chr
def EncodeVarint(write, value):
bits = value & 0x7f
value >>= 7
while value:
write(local_chr(0x80|bits))
bits = value & 0x7f
value >>= 7
return write(local_chr(bits))
return EncodeVarint
def _SignedVarintEncoder():
"""Return an encoder for a basic signed varint value (does not include
tag)."""
local_chr = chr
def EncodeSignedVarint(write, value):
if value < 0:
value += (1 << 64)
bits = value & 0x7f
value >>= 7
while value:
write(local_chr(0x80|bits))
bits = value & 0x7f
value >>= 7
return write(local_chr(bits))
return EncodeSignedVarint
_EncodeVarint = _VarintEncoder()
_EncodeSignedVarint = _SignedVarintEncoder()
def _VarintBytes(value):
"""Encode the given integer as a varint and return the bytes. This is only
called at startup time so it doesn't need to be fast."""
pieces = []
_EncodeVarint(pieces.append, value)
return "".join(pieces)
def TagBytes(field_number, wire_type):
"""Encode the given tag and return the bytes. Only called at startup."""
return _VarintBytes(wire_format.PackTag(field_number, wire_type))
# --------------------------------------------------------------------
# As with sizers (see above), we have a number of common encoder
# implementations.
def _SimpleEncoder(wire_type, encode_value, compute_value_size):
"""Return a constructor for an encoder for fields of a particular type.
Args:
wire_type: The field's wire type, for encoding tags.
encode_value: A function which encodes an individual value, e.g.
_EncodeVarint().
compute_value_size: A function which computes the size of an individual
value, e.g. _VarintSize().
"""
def SpecificEncoder(field_number, is_repeated, is_packed):
if is_packed:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
def EncodePackedField(write, value):
write(tag_bytes)
size = 0
for element in value:
size += compute_value_size(element)
local_EncodeVarint(write, size)
for element in value:
encode_value(write, element)
return EncodePackedField
elif is_repeated:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeRepeatedField(write, value):
for element in value:
write(tag_bytes)
encode_value(write, element)
return EncodeRepeatedField
else:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeField(write, value):
write(tag_bytes)
return encode_value(write, value)
return EncodeField
return SpecificEncoder
def _ModifiedEncoder(wire_type, encode_value, compute_value_size, modify_value):
"""Like SimpleEncoder but additionally invokes modify_value on every value
before passing it to encode_value. Usually modify_value is ZigZagEncode."""
def SpecificEncoder(field_number, is_repeated, is_packed):
if is_packed:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
def EncodePackedField(write, value):
write(tag_bytes)
size = 0
for element in value:
size += compute_value_size(modify_value(element))
local_EncodeVarint(write, size)
for element in value:
encode_value(write, modify_value(element))
return EncodePackedField
elif is_repeated:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeRepeatedField(write, value):
for element in value:
write(tag_bytes)
encode_value(write, modify_value(element))
return EncodeRepeatedField
else:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeField(write, value):
write(tag_bytes)
return encode_value(write, modify_value(value))
return EncodeField
return SpecificEncoder
def _StructPackEncoder(wire_type, format):
"""Return a constructor for an encoder for a fixed-width field.
Args:
wire_type: The field's wire type, for encoding tags.
format: The format string to pass to struct.pack().
"""
value_size = struct.calcsize(format)
def SpecificEncoder(field_number, is_repeated, is_packed):
local_struct_pack = struct.pack
if is_packed:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
def EncodePackedField(write, value):
write(tag_bytes)
local_EncodeVarint(write, len(value) * value_size)
for element in value:
write(local_struct_pack(format, element))
return EncodePackedField
elif is_repeated:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeRepeatedField(write, value):
for element in value:
write(tag_bytes)
write(local_struct_pack(format, element))
return EncodeRepeatedField
else:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeField(write, value):
write(tag_bytes)
return write(local_struct_pack(format, value))
return EncodeField
return SpecificEncoder
def _FloatingPointEncoder(wire_type, format):
"""Return a constructor for an encoder for float fields.
This is like StructPackEncoder, but catches errors that may be due to
passing non-finite floating-point values to struct.pack, and makes a
second attempt to encode those values.
Args:
wire_type: The field's wire type, for encoding tags.
format: The format string to pass to struct.pack().
"""
value_size = struct.calcsize(format)
if value_size == 4:
def EncodeNonFiniteOrRaise(write, value):
# Remember that the serialized form uses little-endian byte order.
if value == _POS_INF:
write('\x00\x00\x80\x7F')
elif value == _NEG_INF:
write('\x00\x00\x80\xFF')
elif value != value: # NaN
write('\x00\x00\xC0\x7F')
else:
raise
elif value_size == 8:
def EncodeNonFiniteOrRaise(write, value):
if value == _POS_INF:
write('\x00\x00\x00\x00\x00\x00\xF0\x7F')
elif value == _NEG_INF:
write('\x00\x00\x00\x00\x00\x00\xF0\xFF')
elif value != value: # NaN
write('\x00\x00\x00\x00\x00\x00\xF8\x7F')
else:
raise
else:
raise ValueError('Can\'t encode floating-point values that are '
'%d bytes long (only 4 or 8)' % value_size)
def SpecificEncoder(field_number, is_repeated, is_packed):
local_struct_pack = struct.pack
if is_packed:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
def EncodePackedField(write, value):
write(tag_bytes)
local_EncodeVarint(write, len(value) * value_size)
for element in value:
# This try/except block is going to be faster than any code that
# we could write to check whether element is finite.
try:
write(local_struct_pack(format, element))
except SystemError:
EncodeNonFiniteOrRaise(write, element)
return EncodePackedField
elif is_repeated:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeRepeatedField(write, value):
for element in value:
write(tag_bytes)
try:
write(local_struct_pack(format, element))
except SystemError:
EncodeNonFiniteOrRaise(write, element)
return EncodeRepeatedField
else:
tag_bytes = TagBytes(field_number, wire_type)
def EncodeField(write, value):
write(tag_bytes)
try:
write(local_struct_pack(format, value))
except SystemError:
EncodeNonFiniteOrRaise(write, value)
return EncodeField
return SpecificEncoder
# ====================================================================
# Here we declare an encoder constructor for each field type. These work
# very similarly to sizer constructors, described earlier.
Int32Encoder = Int64Encoder = EnumEncoder = _SimpleEncoder(
wire_format.WIRETYPE_VARINT, _EncodeSignedVarint, _SignedVarintSize)
UInt32Encoder = UInt64Encoder = _SimpleEncoder(
wire_format.WIRETYPE_VARINT, _EncodeVarint, _VarintSize)
SInt32Encoder = SInt64Encoder = _ModifiedEncoder(
wire_format.WIRETYPE_VARINT, _EncodeVarint, _VarintSize,
wire_format.ZigZagEncode)
# Note that Python conveniently guarantees that when using the '<' prefix on
# formats, they will also have the same size across all platforms (as opposed
# to without the prefix, where their sizes depend on the C compiler's basic
# type sizes).
Fixed32Encoder = _StructPackEncoder(wire_format.WIRETYPE_FIXED32, '<I')
Fixed64Encoder = _StructPackEncoder(wire_format.WIRETYPE_FIXED64, '<Q')
SFixed32Encoder = _StructPackEncoder(wire_format.WIRETYPE_FIXED32, '<i')
SFixed64Encoder = _StructPackEncoder(wire_format.WIRETYPE_FIXED64, '<q')
FloatEncoder = _FloatingPointEncoder(wire_format.WIRETYPE_FIXED32, '<f')
DoubleEncoder = _FloatingPointEncoder(wire_format.WIRETYPE_FIXED64, '<d')
def BoolEncoder(field_number, is_repeated, is_packed):
"""Returns an encoder for a boolean field."""
false_byte = chr(0)
true_byte = chr(1)
if is_packed:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
def EncodePackedField(write, value):
write(tag_bytes)
local_EncodeVarint(write, len(value))
for element in value:
if element:
write(true_byte)
else:
write(false_byte)
return EncodePackedField
elif is_repeated:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_VARINT)
def EncodeRepeatedField(write, value):
for element in value:
write(tag_bytes)
if element:
write(true_byte)
else:
write(false_byte)
return EncodeRepeatedField
else:
tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_VARINT)
def EncodeField(write, value):
write(tag_bytes)
if value:
return write(true_byte)
return write(false_byte)
return EncodeField
def StringEncoder(field_number, is_repeated, is_packed):
"""Returns an encoder for a string field."""
tag = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
local_len = len
assert not is_packed
if is_repeated:
def EncodeRepeatedField(write, value):
for element in value:
encoded = element.encode('utf-8')
write(tag)
local_EncodeVarint(write, local_len(encoded))
write(encoded)
return EncodeRepeatedField
else:
def EncodeField(write, value):
encoded = value.encode('utf-8')
write(tag)
local_EncodeVarint(write, local_len(encoded))
return write(encoded)
return EncodeField
def BytesEncoder(field_number, is_repeated, is_packed):
"""Returns an encoder for a bytes field."""
tag = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
local_len = len
assert not is_packed
if is_repeated:
def EncodeRepeatedField(write, value):
for element in value:
write(tag)
local_EncodeVarint(write, local_len(element))
write(element)
return EncodeRepeatedField
else:
def EncodeField(write, value):
write(tag)
local_EncodeVarint(write, local_len(value))
return write(value)
return EncodeField
def GroupEncoder(field_number, is_repeated, is_packed):
"""Returns an encoder for a group field."""
start_tag = TagBytes(field_number, wire_format.WIRETYPE_START_GROUP)
end_tag = TagBytes(field_number, wire_format.WIRETYPE_END_GROUP)
assert not is_packed
if is_repeated:
def EncodeRepeatedField(write, value):
for element in value:
write(start_tag)
element._InternalSerialize(write)
write(end_tag)
return EncodeRepeatedField
else:
def EncodeField(write, value):
write(start_tag)
value._InternalSerialize(write)
return write(end_tag)
return EncodeField
def MessageEncoder(field_number, is_repeated, is_packed):
"""Returns an encoder for a message field."""
tag = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
local_EncodeVarint = _EncodeVarint
assert not is_packed
if is_repeated:
def EncodeRepeatedField(write, value):
for element in value:
write(tag)
local_EncodeVarint(write, element.ByteSize())
element._InternalSerialize(write)
return EncodeRepeatedField
else:
def EncodeField(write, value):
write(tag)
local_EncodeVarint(write, value.ByteSize())
return value._InternalSerialize(write)
return EncodeField
# --------------------------------------------------------------------
# As before, MessageSet is special.
def MessageSetItemEncoder(field_number):
"""Encoder for extensions of MessageSet.
The message set message looks like this:
message MessageSet {
repeated group Item = 1 {
required int32 type_id = 2;
required string message = 3;
}
}
"""
start_bytes = "".join([
TagBytes(1, wire_format.WIRETYPE_START_GROUP),
TagBytes(2, wire_format.WIRETYPE_VARINT),
_VarintBytes(field_number),
TagBytes(3, wire_format.WIRETYPE_LENGTH_DELIMITED)])
end_bytes = TagBytes(1, wire_format.WIRETYPE_END_GROUP)
local_EncodeVarint = _EncodeVarint
def EncodeField(write, value):
write(start_bytes)
local_EncodeVarint(write, value.ByteSize())
value._InternalSerialize(write)
return write(end_bytes)
return EncodeField

View File

@ -0,0 +1,78 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Defines a listener interface for observing certain
state transitions on Message objects.
Also defines a null implementation of this interface.
"""
__author__ = 'robinson@google.com (Will Robinson)'
class MessageListener(object):
"""Listens for modifications made to a message. Meant to be registered via
Message._SetListener().
Attributes:
dirty: If True, then calling Modified() would be a no-op. This can be
used to avoid these calls entirely in the common case.
"""
def Modified(self):
"""Called every time the message is modified in such a way that the parent
message may need to be updated. This currently means either:
(a) The message was modified for the first time, so the parent message
should henceforth mark the message as present.
(b) The message's cached byte size became dirty -- i.e. the message was
modified for the first time after a previous call to ByteSize().
Therefore the parent should also mark its byte size as dirty.
Note that (a) implies (b), since new objects start out with a client cached
size (zero). However, we document (a) explicitly because it is important.
Modified() will *only* be called in response to one of these two events --
not every time the sub-message is modified.
Note that if the listener's |dirty| attribute is true, then calling
Modified at the moment would be a no-op, so it can be skipped. Performance-
sensitive callers should check this attribute directly before calling since
it will be true most of the time.
"""
raise NotImplementedError
class NullMessageListener(object):
"""No-op MessageListener implementation."""
def Modified(self):
pass

File diff suppressed because it is too large Load Diff

View File

@ -0,0 +1,286 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Provides type checking routines.
This module defines type checking utilities in the forms of dictionaries:
VALUE_CHECKERS: A dictionary of field types and a value validation object.
TYPE_TO_BYTE_SIZE_FN: A dictionary with field types and a size computing
function.
TYPE_TO_SERIALIZE_METHOD: A dictionary with field types and serialization
function.
FIELD_TYPE_TO_WIRE_TYPE: A dictionary with field typed and their
coresponding wire types.
TYPE_TO_DESERIALIZE_METHOD: A dictionary with field types and deserialization
function.
"""
__author__ = 'robinson@google.com (Will Robinson)'
from google.protobuf.internal import decoder
from google.protobuf.internal import encoder
from google.protobuf.internal import wire_format
from google.protobuf import descriptor
_FieldDescriptor = descriptor.FieldDescriptor
def GetTypeChecker(cpp_type, field_type):
"""Returns a type checker for a message field of the specified types.
Args:
cpp_type: C++ type of the field (see descriptor.py).
field_type: Protocol message field type (see descriptor.py).
Returns:
An instance of TypeChecker which can be used to verify the types
of values assigned to a field of the specified type.
"""
if (cpp_type == _FieldDescriptor.CPPTYPE_STRING and
field_type == _FieldDescriptor.TYPE_STRING):
return UnicodeValueChecker()
return _VALUE_CHECKERS[cpp_type]
# None of the typecheckers below make any attempt to guard against people
# subclassing builtin types and doing weird things. We're not trying to
# protect against malicious clients here, just people accidentally shooting
# themselves in the foot in obvious ways.
class TypeChecker(object):
"""Type checker used to catch type errors as early as possible
when the client is setting scalar fields in protocol messages.
"""
def __init__(self, *acceptable_types):
self._acceptable_types = acceptable_types
def CheckValue(self, proposed_value):
if not isinstance(proposed_value, self._acceptable_types):
message = ('%.1024r has type %s, but expected one of: %s' %
(proposed_value, type(proposed_value), self._acceptable_types))
raise TypeError(message)
# IntValueChecker and its subclasses perform integer type-checks
# and bounds-checks.
class IntValueChecker(object):
"""Checker used for integer fields. Performs type-check and range check."""
def CheckValue(self, proposed_value):
if not isinstance(proposed_value, (int, long)):
message = ('%.1024r has type %s, but expected one of: %s' %
(proposed_value, type(proposed_value), (int, long)))
raise TypeError(message)
if not self._MIN <= proposed_value <= self._MAX:
raise ValueError('Value out of range: %d' % proposed_value)
class UnicodeValueChecker(object):
"""Checker used for string fields."""
def CheckValue(self, proposed_value):
if not isinstance(proposed_value, (str, unicode)):
message = ('%.1024r has type %s, but expected one of: %s' %
(proposed_value, type(proposed_value), (str, unicode)))
raise TypeError(message)
# If the value is of type 'str' make sure that it is in 7-bit ASCII
# encoding.
if isinstance(proposed_value, str):
try:
unicode(proposed_value, 'ascii')
except UnicodeDecodeError:
raise ValueError('%.1024r has type str, but isn\'t in 7-bit ASCII '
'encoding. Non-ASCII strings must be converted to '
'unicode objects before being added.' %
(proposed_value))
class Int32ValueChecker(IntValueChecker):
# We're sure to use ints instead of longs here since comparison may be more
# efficient.
_MIN = -2147483648
_MAX = 2147483647
class Uint32ValueChecker(IntValueChecker):
_MIN = 0
_MAX = (1 << 32) - 1
class Int64ValueChecker(IntValueChecker):
_MIN = -(1 << 63)
_MAX = (1 << 63) - 1
class Uint64ValueChecker(IntValueChecker):
_MIN = 0
_MAX = (1 << 64) - 1
# Type-checkers for all scalar CPPTYPEs.
_VALUE_CHECKERS = {
_FieldDescriptor.CPPTYPE_INT32: Int32ValueChecker(),
_FieldDescriptor.CPPTYPE_INT64: Int64ValueChecker(),
_FieldDescriptor.CPPTYPE_UINT32: Uint32ValueChecker(),
_FieldDescriptor.CPPTYPE_UINT64: Uint64ValueChecker(),
_FieldDescriptor.CPPTYPE_DOUBLE: TypeChecker(
float, int, long),
_FieldDescriptor.CPPTYPE_FLOAT: TypeChecker(
float, int, long),
_FieldDescriptor.CPPTYPE_BOOL: TypeChecker(bool, int),
_FieldDescriptor.CPPTYPE_ENUM: Int32ValueChecker(),
_FieldDescriptor.CPPTYPE_STRING: TypeChecker(str),
}
# Map from field type to a function F, such that F(field_num, value)
# gives the total byte size for a value of the given type. This
# byte size includes tag information and any other additional space
# associated with serializing "value".
TYPE_TO_BYTE_SIZE_FN = {
_FieldDescriptor.TYPE_DOUBLE: wire_format.DoubleByteSize,
_FieldDescriptor.TYPE_FLOAT: wire_format.FloatByteSize,
_FieldDescriptor.TYPE_INT64: wire_format.Int64ByteSize,
_FieldDescriptor.TYPE_UINT64: wire_format.UInt64ByteSize,
_FieldDescriptor.TYPE_INT32: wire_format.Int32ByteSize,
_FieldDescriptor.TYPE_FIXED64: wire_format.Fixed64ByteSize,
_FieldDescriptor.TYPE_FIXED32: wire_format.Fixed32ByteSize,
_FieldDescriptor.TYPE_BOOL: wire_format.BoolByteSize,
_FieldDescriptor.TYPE_STRING: wire_format.StringByteSize,
_FieldDescriptor.TYPE_GROUP: wire_format.GroupByteSize,
_FieldDescriptor.TYPE_MESSAGE: wire_format.MessageByteSize,
_FieldDescriptor.TYPE_BYTES: wire_format.BytesByteSize,
_FieldDescriptor.TYPE_UINT32: wire_format.UInt32ByteSize,
_FieldDescriptor.TYPE_ENUM: wire_format.EnumByteSize,
_FieldDescriptor.TYPE_SFIXED32: wire_format.SFixed32ByteSize,
_FieldDescriptor.TYPE_SFIXED64: wire_format.SFixed64ByteSize,
_FieldDescriptor.TYPE_SINT32: wire_format.SInt32ByteSize,
_FieldDescriptor.TYPE_SINT64: wire_format.SInt64ByteSize
}
# Maps from field types to encoder constructors.
TYPE_TO_ENCODER = {
_FieldDescriptor.TYPE_DOUBLE: encoder.DoubleEncoder,
_FieldDescriptor.TYPE_FLOAT: encoder.FloatEncoder,
_FieldDescriptor.TYPE_INT64: encoder.Int64Encoder,
_FieldDescriptor.TYPE_UINT64: encoder.UInt64Encoder,
_FieldDescriptor.TYPE_INT32: encoder.Int32Encoder,
_FieldDescriptor.TYPE_FIXED64: encoder.Fixed64Encoder,
_FieldDescriptor.TYPE_FIXED32: encoder.Fixed32Encoder,
_FieldDescriptor.TYPE_BOOL: encoder.BoolEncoder,
_FieldDescriptor.TYPE_STRING: encoder.StringEncoder,
_FieldDescriptor.TYPE_GROUP: encoder.GroupEncoder,
_FieldDescriptor.TYPE_MESSAGE: encoder.MessageEncoder,
_FieldDescriptor.TYPE_BYTES: encoder.BytesEncoder,
_FieldDescriptor.TYPE_UINT32: encoder.UInt32Encoder,
_FieldDescriptor.TYPE_ENUM: encoder.EnumEncoder,
_FieldDescriptor.TYPE_SFIXED32: encoder.SFixed32Encoder,
_FieldDescriptor.TYPE_SFIXED64: encoder.SFixed64Encoder,
_FieldDescriptor.TYPE_SINT32: encoder.SInt32Encoder,
_FieldDescriptor.TYPE_SINT64: encoder.SInt64Encoder,
}
# Maps from field types to sizer constructors.
TYPE_TO_SIZER = {
_FieldDescriptor.TYPE_DOUBLE: encoder.DoubleSizer,
_FieldDescriptor.TYPE_FLOAT: encoder.FloatSizer,
_FieldDescriptor.TYPE_INT64: encoder.Int64Sizer,
_FieldDescriptor.TYPE_UINT64: encoder.UInt64Sizer,
_FieldDescriptor.TYPE_INT32: encoder.Int32Sizer,
_FieldDescriptor.TYPE_FIXED64: encoder.Fixed64Sizer,
_FieldDescriptor.TYPE_FIXED32: encoder.Fixed32Sizer,
_FieldDescriptor.TYPE_BOOL: encoder.BoolSizer,
_FieldDescriptor.TYPE_STRING: encoder.StringSizer,
_FieldDescriptor.TYPE_GROUP: encoder.GroupSizer,
_FieldDescriptor.TYPE_MESSAGE: encoder.MessageSizer,
_FieldDescriptor.TYPE_BYTES: encoder.BytesSizer,
_FieldDescriptor.TYPE_UINT32: encoder.UInt32Sizer,
_FieldDescriptor.TYPE_ENUM: encoder.EnumSizer,
_FieldDescriptor.TYPE_SFIXED32: encoder.SFixed32Sizer,
_FieldDescriptor.TYPE_SFIXED64: encoder.SFixed64Sizer,
_FieldDescriptor.TYPE_SINT32: encoder.SInt32Sizer,
_FieldDescriptor.TYPE_SINT64: encoder.SInt64Sizer,
}
# Maps from field type to a decoder constructor.
TYPE_TO_DECODER = {
_FieldDescriptor.TYPE_DOUBLE: decoder.DoubleDecoder,
_FieldDescriptor.TYPE_FLOAT: decoder.FloatDecoder,
_FieldDescriptor.TYPE_INT64: decoder.Int64Decoder,
_FieldDescriptor.TYPE_UINT64: decoder.UInt64Decoder,
_FieldDescriptor.TYPE_INT32: decoder.Int32Decoder,
_FieldDescriptor.TYPE_FIXED64: decoder.Fixed64Decoder,
_FieldDescriptor.TYPE_FIXED32: decoder.Fixed32Decoder,
_FieldDescriptor.TYPE_BOOL: decoder.BoolDecoder,
_FieldDescriptor.TYPE_STRING: decoder.StringDecoder,
_FieldDescriptor.TYPE_GROUP: decoder.GroupDecoder,
_FieldDescriptor.TYPE_MESSAGE: decoder.MessageDecoder,
_FieldDescriptor.TYPE_BYTES: decoder.BytesDecoder,
_FieldDescriptor.TYPE_UINT32: decoder.UInt32Decoder,
_FieldDescriptor.TYPE_ENUM: decoder.EnumDecoder,
_FieldDescriptor.TYPE_SFIXED32: decoder.SFixed32Decoder,
_FieldDescriptor.TYPE_SFIXED64: decoder.SFixed64Decoder,
_FieldDescriptor.TYPE_SINT32: decoder.SInt32Decoder,
_FieldDescriptor.TYPE_SINT64: decoder.SInt64Decoder,
}
# Maps from field type to expected wiretype.
FIELD_TYPE_TO_WIRE_TYPE = {
_FieldDescriptor.TYPE_DOUBLE: wire_format.WIRETYPE_FIXED64,
_FieldDescriptor.TYPE_FLOAT: wire_format.WIRETYPE_FIXED32,
_FieldDescriptor.TYPE_INT64: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_UINT64: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_INT32: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_FIXED64: wire_format.WIRETYPE_FIXED64,
_FieldDescriptor.TYPE_FIXED32: wire_format.WIRETYPE_FIXED32,
_FieldDescriptor.TYPE_BOOL: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_STRING:
wire_format.WIRETYPE_LENGTH_DELIMITED,
_FieldDescriptor.TYPE_GROUP: wire_format.WIRETYPE_START_GROUP,
_FieldDescriptor.TYPE_MESSAGE:
wire_format.WIRETYPE_LENGTH_DELIMITED,
_FieldDescriptor.TYPE_BYTES:
wire_format.WIRETYPE_LENGTH_DELIMITED,
_FieldDescriptor.TYPE_UINT32: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_ENUM: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_SFIXED32: wire_format.WIRETYPE_FIXED32,
_FieldDescriptor.TYPE_SFIXED64: wire_format.WIRETYPE_FIXED64,
_FieldDescriptor.TYPE_SINT32: wire_format.WIRETYPE_VARINT,
_FieldDescriptor.TYPE_SINT64: wire_format.WIRETYPE_VARINT,
}

View File

@ -0,0 +1,268 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Constants and static functions to support protocol buffer wire format."""
__author__ = 'robinson@google.com (Will Robinson)'
import struct
from google.protobuf import descriptor
from google.protobuf import message
TAG_TYPE_BITS = 3 # Number of bits used to hold type info in a proto tag.
TAG_TYPE_MASK = (1 << TAG_TYPE_BITS) - 1 # 0x7
# These numbers identify the wire type of a protocol buffer value.
# We use the least-significant TAG_TYPE_BITS bits of the varint-encoded
# tag-and-type to store one of these WIRETYPE_* constants.
# These values must match WireType enum in google/protobuf/wire_format.h.
WIRETYPE_VARINT = 0
WIRETYPE_FIXED64 = 1
WIRETYPE_LENGTH_DELIMITED = 2
WIRETYPE_START_GROUP = 3
WIRETYPE_END_GROUP = 4
WIRETYPE_FIXED32 = 5
_WIRETYPE_MAX = 5
# Bounds for various integer types.
INT32_MAX = int((1 << 31) - 1)
INT32_MIN = int(-(1 << 31))
UINT32_MAX = (1 << 32) - 1
INT64_MAX = (1 << 63) - 1
INT64_MIN = -(1 << 63)
UINT64_MAX = (1 << 64) - 1
# "struct" format strings that will encode/decode the specified formats.
FORMAT_UINT32_LITTLE_ENDIAN = '<I'
FORMAT_UINT64_LITTLE_ENDIAN = '<Q'
FORMAT_FLOAT_LITTLE_ENDIAN = '<f'
FORMAT_DOUBLE_LITTLE_ENDIAN = '<d'
# We'll have to provide alternate implementations of AppendLittleEndian*() on
# any architectures where these checks fail.
if struct.calcsize(FORMAT_UINT32_LITTLE_ENDIAN) != 4:
raise AssertionError('Format "I" is not a 32-bit number.')
if struct.calcsize(FORMAT_UINT64_LITTLE_ENDIAN) != 8:
raise AssertionError('Format "Q" is not a 64-bit number.')
def PackTag(field_number, wire_type):
"""Returns an unsigned 32-bit integer that encodes the field number and
wire type information in standard protocol message wire format.
Args:
field_number: Expected to be an integer in the range [1, 1 << 29)
wire_type: One of the WIRETYPE_* constants.
"""
if not 0 <= wire_type <= _WIRETYPE_MAX:
raise message.EncodeError('Unknown wire type: %d' % wire_type)
return (field_number << TAG_TYPE_BITS) | wire_type
def UnpackTag(tag):
"""The inverse of PackTag(). Given an unsigned 32-bit number,
returns a (field_number, wire_type) tuple.
"""
return (tag >> TAG_TYPE_BITS), (tag & TAG_TYPE_MASK)
def ZigZagEncode(value):
"""ZigZag Transform: Encodes signed integers so that they can be
effectively used with varint encoding. See wire_format.h for
more details.
"""
if value >= 0:
return value << 1
return (value << 1) ^ (~0)
def ZigZagDecode(value):
"""Inverse of ZigZagEncode()."""
if not value & 0x1:
return value >> 1
return (value >> 1) ^ (~0)
# The *ByteSize() functions below return the number of bytes required to
# serialize "field number + type" information and then serialize the value.
def Int32ByteSize(field_number, int32):
return Int64ByteSize(field_number, int32)
def Int32ByteSizeNoTag(int32):
return _VarUInt64ByteSizeNoTag(0xffffffffffffffff & int32)
def Int64ByteSize(field_number, int64):
# Have to convert to uint before calling UInt64ByteSize().
return UInt64ByteSize(field_number, 0xffffffffffffffff & int64)
def UInt32ByteSize(field_number, uint32):
return UInt64ByteSize(field_number, uint32)
def UInt64ByteSize(field_number, uint64):
return TagByteSize(field_number) + _VarUInt64ByteSizeNoTag(uint64)
def SInt32ByteSize(field_number, int32):
return UInt32ByteSize(field_number, ZigZagEncode(int32))
def SInt64ByteSize(field_number, int64):
return UInt64ByteSize(field_number, ZigZagEncode(int64))
def Fixed32ByteSize(field_number, fixed32):
return TagByteSize(field_number) + 4
def Fixed64ByteSize(field_number, fixed64):
return TagByteSize(field_number) + 8
def SFixed32ByteSize(field_number, sfixed32):
return TagByteSize(field_number) + 4
def SFixed64ByteSize(field_number, sfixed64):
return TagByteSize(field_number) + 8
def FloatByteSize(field_number, flt):
return TagByteSize(field_number) + 4
def DoubleByteSize(field_number, double):
return TagByteSize(field_number) + 8
def BoolByteSize(field_number, b):
return TagByteSize(field_number) + 1
def EnumByteSize(field_number, enum):
return UInt32ByteSize(field_number, enum)
def StringByteSize(field_number, string):
return BytesByteSize(field_number, string.encode('utf-8'))
def BytesByteSize(field_number, b):
return (TagByteSize(field_number)
+ _VarUInt64ByteSizeNoTag(len(b))
+ len(b))
def GroupByteSize(field_number, message):
return (2 * TagByteSize(field_number) # START and END group.
+ message.ByteSize())
def MessageByteSize(field_number, message):
return (TagByteSize(field_number)
+ _VarUInt64ByteSizeNoTag(message.ByteSize())
+ message.ByteSize())
def MessageSetItemByteSize(field_number, msg):
# First compute the sizes of the tags.
# There are 2 tags for the beginning and ending of the repeated group, that
# is field number 1, one with field number 2 (type_id) and one with field
# number 3 (message).
total_size = (2 * TagByteSize(1) + TagByteSize(2) + TagByteSize(3))
# Add the number of bytes for type_id.
total_size += _VarUInt64ByteSizeNoTag(field_number)
message_size = msg.ByteSize()
# The number of bytes for encoding the length of the message.
total_size += _VarUInt64ByteSizeNoTag(message_size)
# The size of the message.
total_size += message_size
return total_size
def TagByteSize(field_number):
"""Returns the bytes required to serialize a tag with this field number."""
# Just pass in type 0, since the type won't affect the tag+type size.
return _VarUInt64ByteSizeNoTag(PackTag(field_number, 0))
# Private helper function for the *ByteSize() functions above.
def _VarUInt64ByteSizeNoTag(uint64):
"""Returns the number of bytes required to serialize a single varint
using boundary value comparisons. (unrolled loop optimization -WPierce)
uint64 must be unsigned.
"""
if uint64 <= 0x7f: return 1
if uint64 <= 0x3fff: return 2
if uint64 <= 0x1fffff: return 3
if uint64 <= 0xfffffff: return 4
if uint64 <= 0x7ffffffff: return 5
if uint64 <= 0x3ffffffffff: return 6
if uint64 <= 0x1ffffffffffff: return 7
if uint64 <= 0xffffffffffffff: return 8
if uint64 <= 0x7fffffffffffffff: return 9
if uint64 > UINT64_MAX:
raise message.EncodeError('Value out of range: %d' % uint64)
return 10
NON_PACKABLE_TYPES = (
descriptor.FieldDescriptor.TYPE_STRING,
descriptor.FieldDescriptor.TYPE_GROUP,
descriptor.FieldDescriptor.TYPE_MESSAGE,
descriptor.FieldDescriptor.TYPE_BYTES
)
def IsTypePackable(field_type):
"""Return true iff packable = true is valid for fields of this type.
Args:
field_type: a FieldDescriptor::Type value.
Returns:
True iff fields of this type are packable.
"""
return field_type not in NON_PACKABLE_TYPES

View File

@ -0,0 +1,268 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
# TODO(robinson): We should just make these methods all "pure-virtual" and move
# all implementation out, into reflection.py for now.
"""Contains an abstract base class for protocol messages."""
__author__ = 'robinson@google.com (Will Robinson)'
class Error(Exception): pass
class DecodeError(Error): pass
class EncodeError(Error): pass
class Message(object):
"""Abstract base class for protocol messages.
Protocol message classes are almost always generated by the protocol
compiler. These generated types subclass Message and implement the methods
shown below.
TODO(robinson): Link to an HTML document here.
TODO(robinson): Document that instances of this class will also
have an Extensions attribute with __getitem__ and __setitem__.
Again, not sure how to best convey this.
TODO(robinson): Document that the class must also have a static
RegisterExtension(extension_field) method.
Not sure how to best express at this point.
"""
# TODO(robinson): Document these fields and methods.
__slots__ = []
DESCRIPTOR = None
def __deepcopy__(self, memo=None):
clone = type(self)()
clone.MergeFrom(self)
return clone
def __eq__(self, other_msg):
raise NotImplementedError
def __ne__(self, other_msg):
# Can't just say self != other_msg, since that would infinitely recurse. :)
return not self == other_msg
def __hash__(self):
raise TypeError('unhashable object')
def __str__(self):
raise NotImplementedError
def __unicode__(self):
raise NotImplementedError
def MergeFrom(self, other_msg):
"""Merges the contents of the specified message into current message.
This method merges the contents of the specified message into the current
message. Singular fields that are set in the specified message overwrite
the corresponding fields in the current message. Repeated fields are
appended. Singular sub-messages and groups are recursively merged.
Args:
other_msg: Message to merge into the current message.
"""
raise NotImplementedError
def CopyFrom(self, other_msg):
"""Copies the content of the specified message into the current message.
The method clears the current message and then merges the specified
message using MergeFrom.
Args:
other_msg: Message to copy into the current one.
"""
if self is other_msg:
return
self.Clear()
self.MergeFrom(other_msg)
def Clear(self):
"""Clears all data that was set in the message."""
raise NotImplementedError
def SetInParent(self):
"""Mark this as present in the parent.
This normally happens automatically when you assign a field of a
sub-message, but sometimes you want to make the sub-message
present while keeping it empty. If you find yourself using this,
you may want to reconsider your design."""
raise NotImplementedError
def IsInitialized(self):
"""Checks if the message is initialized.
Returns:
The method returns True if the message is initialized (i.e. all of its
required fields are set).
"""
raise NotImplementedError
# TODO(robinson): MergeFromString() should probably return None and be
# implemented in terms of a helper that returns the # of bytes read. Our
# deserialization routines would use the helper when recursively
# deserializing, but the end user would almost always just want the no-return
# MergeFromString().
def MergeFromString(self, serialized):
"""Merges serialized protocol buffer data into this message.
When we find a field in |serialized| that is already present
in this message:
- If it's a "repeated" field, we append to the end of our list.
- Else, if it's a scalar, we overwrite our field.
- Else, (it's a nonrepeated composite), we recursively merge
into the existing composite.
TODO(robinson): Document handling of unknown fields.
Args:
serialized: Any object that allows us to call buffer(serialized)
to access a string of bytes using the buffer interface.
TODO(robinson): When we switch to a helper, this will return None.
Returns:
The number of bytes read from |serialized|.
For non-group messages, this will always be len(serialized),
but for messages which are actually groups, this will
generally be less than len(serialized), since we must
stop when we reach an END_GROUP tag. Note that if
we *do* stop because of an END_GROUP tag, the number
of bytes returned does not include the bytes
for the END_GROUP tag information.
"""
raise NotImplementedError
def ParseFromString(self, serialized):
"""Like MergeFromString(), except we clear the object first."""
self.Clear()
self.MergeFromString(serialized)
def SerializeToString(self):
"""Serializes the protocol message to a binary string.
Returns:
A binary string representation of the message if all of the required
fields in the message are set (i.e. the message is initialized).
Raises:
message.EncodeError if the message isn't initialized.
"""
raise NotImplementedError
def SerializePartialToString(self):
"""Serializes the protocol message to a binary string.
This method is similar to SerializeToString but doesn't check if the
message is initialized.
Returns:
A string representation of the partial message.
"""
raise NotImplementedError
# TODO(robinson): Decide whether we like these better
# than auto-generated has_foo() and clear_foo() methods
# on the instances themselves. This way is less consistent
# with C++, but it makes reflection-type access easier and
# reduces the number of magically autogenerated things.
#
# TODO(robinson): Be sure to document (and test) exactly
# which field names are accepted here. Are we case-sensitive?
# What do we do with fields that share names with Python keywords
# like 'lambda' and 'yield'?
#
# nnorwitz says:
# """
# Typically (in python), an underscore is appended to names that are
# keywords. So they would become lambda_ or yield_.
# """
def ListFields(self):
"""Returns a list of (FieldDescriptor, value) tuples for all
fields in the message which are not empty. A singular field is non-empty
if HasField() would return true, and a repeated field is non-empty if
it contains at least one element. The fields are ordered by field
number"""
raise NotImplementedError
def HasField(self, field_name):
"""Checks if a certain field is set for the message. Note if the
field_name is not defined in the message descriptor, ValueError will be
raised."""
raise NotImplementedError
def ClearField(self, field_name):
raise NotImplementedError
def HasExtension(self, extension_handle):
raise NotImplementedError
def ClearExtension(self, extension_handle):
raise NotImplementedError
def ByteSize(self):
"""Returns the serialized size of this message.
Recursively calls ByteSize() on all contained messages.
"""
raise NotImplementedError
def _SetListener(self, message_listener):
"""Internal method used by the protocol message implementation.
Clients should not call this directly.
Sets a listener that this message will call on certain state transitions.
The purpose of this method is to register back-edges from children to
parents at runtime, for the purpose of setting "has" bits and
byte-size-dirty bits in the parent and ancestor objects whenever a child or
descendant object is modified.
If the client wants to disconnect this Message from the object tree, she
explicitly sets callback to None.
If message_listener is None, unregisters any existing listener. Otherwise,
message_listener must implement the MessageListener interface in
internal/message_listener.py, and we discard any listener registered
via a previous _SetListener() call.
"""
raise NotImplementedError

View File

@ -0,0 +1,142 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
# This code is meant to work on Python 2.4 and above only.
"""Contains a metaclass and helper functions used to create
protocol message classes from Descriptor objects at runtime.
Recall that a metaclass is the "type" of a class.
(A class is to a metaclass what an instance is to a class.)
In this case, we use the GeneratedProtocolMessageType metaclass
to inject all the useful functionality into the classes
output by the protocol compiler at compile-time.
The upshot of all this is that the real implementation
details for ALL pure-Python protocol buffers are *here in
this file*.
"""
__author__ = 'robinson@google.com (Will Robinson)'
from google.protobuf.internal import api_implementation
from google.protobuf import descriptor as descriptor_mod
_FieldDescriptor = descriptor_mod.FieldDescriptor
if api_implementation.Type() == 'cpp':
from google.protobuf.internal import cpp_message
_NewMessage = cpp_message.NewMessage
_InitMessage = cpp_message.InitMessage
else:
from google.protobuf.internal import python_message
_NewMessage = python_message.NewMessage
_InitMessage = python_message.InitMessage
class GeneratedProtocolMessageType(type):
"""Metaclass for protocol message classes created at runtime from Descriptors.
We add implementations for all methods described in the Message class. We
also create properties to allow getting/setting all fields in the protocol
message. Finally, we create slots to prevent users from accidentally
"setting" nonexistent fields in the protocol message, which then wouldn't get
serialized / deserialized properly.
The protocol compiler currently uses this metaclass to create protocol
message classes at runtime. Clients can also manually create their own
classes at runtime, as in this example:
mydescriptor = Descriptor(.....)
class MyProtoClass(Message):
__metaclass__ = GeneratedProtocolMessageType
DESCRIPTOR = mydescriptor
myproto_instance = MyProtoClass()
myproto.foo_field = 23
...
"""
# Must be consistent with the protocol-compiler code in
# proto2/compiler/internal/generator.*.
_DESCRIPTOR_KEY = 'DESCRIPTOR'
def __new__(cls, name, bases, dictionary):
"""Custom allocation for runtime-generated class types.
We override __new__ because this is apparently the only place
where we can meaningfully set __slots__ on the class we're creating(?).
(The interplay between metaclasses and slots is not very well-documented).
Args:
name: Name of the class (ignored, but required by the
metaclass protocol).
bases: Base classes of the class we're constructing.
(Should be message.Message). We ignore this field, but
it's required by the metaclass protocol
dictionary: The class dictionary of the class we're
constructing. dictionary[_DESCRIPTOR_KEY] must contain
a Descriptor object describing this protocol message
type.
Returns:
Newly-allocated class.
"""
descriptor = dictionary[GeneratedProtocolMessageType._DESCRIPTOR_KEY]
_NewMessage(descriptor, dictionary)
superclass = super(GeneratedProtocolMessageType, cls)
new_class = superclass.__new__(cls, name, bases, dictionary)
setattr(descriptor, '_concrete_class', new_class)
return new_class
def __init__(cls, name, bases, dictionary):
"""Here we perform the majority of our work on the class.
We add enum getters, an __init__ method, implementations
of all Message methods, and properties for all fields
in the protocol type.
Args:
name: Name of the class (ignored, but required by the
metaclass protocol).
bases: Base classes of the class we're constructing.
(Should be message.Message). We ignore this field, but
it's required by the metaclass protocol
dictionary: The class dictionary of the class we're
constructing. dictionary[_DESCRIPTOR_KEY] must contain
a Descriptor object describing this protocol message
type.
"""
descriptor = dictionary[GeneratedProtocolMessageType._DESCRIPTOR_KEY]
_InitMessage(descriptor, cls)
superclass = super(GeneratedProtocolMessageType, cls)
superclass.__init__(name, bases, dictionary)

View File

@ -0,0 +1,226 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""DEPRECATED: Declares the RPC service interfaces.
This module declares the abstract interfaces underlying proto2 RPC
services. These are intended to be independent of any particular RPC
implementation, so that proto2 services can be used on top of a variety
of implementations. Starting with version 2.3.0, RPC implementations should
not try to build on these, but should instead provide code generator plugins
which generate code specific to the particular RPC implementation. This way
the generated code can be more appropriate for the implementation in use
and can avoid unnecessary layers of indirection.
"""
__author__ = 'petar@google.com (Petar Petrov)'
class RpcException(Exception):
"""Exception raised on failed blocking RPC method call."""
pass
class Service(object):
"""Abstract base interface for protocol-buffer-based RPC services.
Services themselves are abstract classes (implemented either by servers or as
stubs), but they subclass this base interface. The methods of this
interface can be used to call the methods of the service without knowing
its exact type at compile time (analogous to the Message interface).
"""
def GetDescriptor():
"""Retrieves this service's descriptor."""
raise NotImplementedError
def CallMethod(self, method_descriptor, rpc_controller,
request, done):
"""Calls a method of the service specified by method_descriptor.
If "done" is None then the call is blocking and the response
message will be returned directly. Otherwise the call is asynchronous
and "done" will later be called with the response value.
In the blocking case, RpcException will be raised on error.
Preconditions:
* method_descriptor.service == GetDescriptor
* request is of the exact same classes as returned by
GetRequestClass(method).
* After the call has started, the request must not be modified.
* "rpc_controller" is of the correct type for the RPC implementation being
used by this Service. For stubs, the "correct type" depends on the
RpcChannel which the stub is using.
Postconditions:
* "done" will be called when the method is complete. This may be
before CallMethod() returns or it may be at some point in the future.
* If the RPC failed, the response value passed to "done" will be None.
Further details about the failure can be found by querying the
RpcController.
"""
raise NotImplementedError
def GetRequestClass(self, method_descriptor):
"""Returns the class of the request message for the specified method.
CallMethod() requires that the request is of a particular subclass of
Message. GetRequestClass() gets the default instance of this required
type.
Example:
method = service.GetDescriptor().FindMethodByName("Foo")
request = stub.GetRequestClass(method)()
request.ParseFromString(input)
service.CallMethod(method, request, callback)
"""
raise NotImplementedError
def GetResponseClass(self, method_descriptor):
"""Returns the class of the response message for the specified method.
This method isn't really needed, as the RpcChannel's CallMethod constructs
the response protocol message. It's provided anyway in case it is useful
for the caller to know the response type in advance.
"""
raise NotImplementedError
class RpcController(object):
"""An RpcController mediates a single method call.
The primary purpose of the controller is to provide a way to manipulate
settings specific to the RPC implementation and to find out about RPC-level
errors. The methods provided by the RpcController interface are intended
to be a "least common denominator" set of features which we expect all
implementations to support. Specific implementations may provide more
advanced features (e.g. deadline propagation).
"""
# Client-side methods below
def Reset(self):
"""Resets the RpcController to its initial state.
After the RpcController has been reset, it may be reused in
a new call. Must not be called while an RPC is in progress.
"""
raise NotImplementedError
def Failed(self):
"""Returns true if the call failed.
After a call has finished, returns true if the call failed. The possible
reasons for failure depend on the RPC implementation. Failed() must not
be called before a call has finished. If Failed() returns true, the
contents of the response message are undefined.
"""
raise NotImplementedError
def ErrorText(self):
"""If Failed is true, returns a human-readable description of the error."""
raise NotImplementedError
def StartCancel(self):
"""Initiate cancellation.
Advises the RPC system that the caller desires that the RPC call be
canceled. The RPC system may cancel it immediately, may wait awhile and
then cancel it, or may not even cancel the call at all. If the call is
canceled, the "done" callback will still be called and the RpcController
will indicate that the call failed at that time.
"""
raise NotImplementedError
# Server-side methods below
def SetFailed(self, reason):
"""Sets a failure reason.
Causes Failed() to return true on the client side. "reason" will be
incorporated into the message returned by ErrorText(). If you find
you need to return machine-readable information about failures, you
should incorporate it into your response protocol buffer and should
NOT call SetFailed().
"""
raise NotImplementedError
def IsCanceled(self):
"""Checks if the client cancelled the RPC.
If true, indicates that the client canceled the RPC, so the server may
as well give up on replying to it. The server should still call the
final "done" callback.
"""
raise NotImplementedError
def NotifyOnCancel(self, callback):
"""Sets a callback to invoke on cancel.
Asks that the given callback be called when the RPC is canceled. The
callback will always be called exactly once. If the RPC completes without
being canceled, the callback will be called after completion. If the RPC
has already been canceled when NotifyOnCancel() is called, the callback
will be called immediately.
NotifyOnCancel() must be called no more than once per request.
"""
raise NotImplementedError
class RpcChannel(object):
"""Abstract interface for an RPC channel.
An RpcChannel represents a communication line to a service which can be used
to call that service's methods. The service may be running on another
machine. Normally, you should not use an RpcChannel directly, but instead
construct a stub {@link Service} wrapping it. Example:
Example:
RpcChannel channel = rpcImpl.Channel("remotehost.example.com:1234")
RpcController controller = rpcImpl.Controller()
MyService service = MyService_Stub(channel)
service.MyMethod(controller, request, callback)
"""
def CallMethod(self, method_descriptor, rpc_controller,
request, response_class, done):
"""Calls the method identified by the descriptor.
Call the given method of the remote service. The signature of this
procedure looks the same as Service.CallMethod(), but the requirements
are less strict in one important way: the request object doesn't have to
be of any specific class as long as its descriptor is method.input_type.
"""
raise NotImplementedError

View File

@ -0,0 +1,284 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Contains metaclasses used to create protocol service and service stub
classes from ServiceDescriptor objects at runtime.
The GeneratedServiceType and GeneratedServiceStubType metaclasses are used to
inject all useful functionality into the classes output by the protocol
compiler at compile-time.
"""
__author__ = 'petar@google.com (Petar Petrov)'
class GeneratedServiceType(type):
"""Metaclass for service classes created at runtime from ServiceDescriptors.
Implementations for all methods described in the Service class are added here
by this class. We also create properties to allow getting/setting all fields
in the protocol message.
The protocol compiler currently uses this metaclass to create protocol service
classes at runtime. Clients can also manually create their own classes at
runtime, as in this example:
mydescriptor = ServiceDescriptor(.....)
class MyProtoService(service.Service):
__metaclass__ = GeneratedServiceType
DESCRIPTOR = mydescriptor
myservice_instance = MyProtoService()
...
"""
_DESCRIPTOR_KEY = 'DESCRIPTOR'
def __init__(cls, name, bases, dictionary):
"""Creates a message service class.
Args:
name: Name of the class (ignored, but required by the metaclass
protocol).
bases: Base classes of the class being constructed.
dictionary: The class dictionary of the class being constructed.
dictionary[_DESCRIPTOR_KEY] must contain a ServiceDescriptor object
describing this protocol service type.
"""
# Don't do anything if this class doesn't have a descriptor. This happens
# when a service class is subclassed.
if GeneratedServiceType._DESCRIPTOR_KEY not in dictionary:
return
descriptor = dictionary[GeneratedServiceType._DESCRIPTOR_KEY]
service_builder = _ServiceBuilder(descriptor)
service_builder.BuildService(cls)
class GeneratedServiceStubType(GeneratedServiceType):
"""Metaclass for service stubs created at runtime from ServiceDescriptors.
This class has similar responsibilities as GeneratedServiceType, except that
it creates the service stub classes.
"""
_DESCRIPTOR_KEY = 'DESCRIPTOR'
def __init__(cls, name, bases, dictionary):
"""Creates a message service stub class.
Args:
name: Name of the class (ignored, here).
bases: Base classes of the class being constructed.
dictionary: The class dictionary of the class being constructed.
dictionary[_DESCRIPTOR_KEY] must contain a ServiceDescriptor object
describing this protocol service type.
"""
super(GeneratedServiceStubType, cls).__init__(name, bases, dictionary)
# Don't do anything if this class doesn't have a descriptor. This happens
# when a service stub is subclassed.
if GeneratedServiceStubType._DESCRIPTOR_KEY not in dictionary:
return
descriptor = dictionary[GeneratedServiceStubType._DESCRIPTOR_KEY]
service_stub_builder = _ServiceStubBuilder(descriptor)
service_stub_builder.BuildServiceStub(cls)
class _ServiceBuilder(object):
"""This class constructs a protocol service class using a service descriptor.
Given a service descriptor, this class constructs a class that represents
the specified service descriptor. One service builder instance constructs
exactly one service class. That means all instances of that class share the
same builder.
"""
def __init__(self, service_descriptor):
"""Initializes an instance of the service class builder.
Args:
service_descriptor: ServiceDescriptor to use when constructing the
service class.
"""
self.descriptor = service_descriptor
def BuildService(self, cls):
"""Constructs the service class.
Args:
cls: The class that will be constructed.
"""
# CallMethod needs to operate with an instance of the Service class. This
# internal wrapper function exists only to be able to pass the service
# instance to the method that does the real CallMethod work.
def _WrapCallMethod(srvc, method_descriptor,
rpc_controller, request, callback):
return self._CallMethod(srvc, method_descriptor,
rpc_controller, request, callback)
self.cls = cls
cls.CallMethod = _WrapCallMethod
cls.GetDescriptor = staticmethod(lambda: self.descriptor)
cls.GetDescriptor.__doc__ = "Returns the service descriptor."
cls.GetRequestClass = self._GetRequestClass
cls.GetResponseClass = self._GetResponseClass
for method in self.descriptor.methods:
setattr(cls, method.name, self._GenerateNonImplementedMethod(method))
def _CallMethod(self, srvc, method_descriptor,
rpc_controller, request, callback):
"""Calls the method described by a given method descriptor.
Args:
srvc: Instance of the service for which this method is called.
method_descriptor: Descriptor that represent the method to call.
rpc_controller: RPC controller to use for this method's execution.
request: Request protocol message.
callback: A callback to invoke after the method has completed.
"""
if method_descriptor.containing_service != self.descriptor:
raise RuntimeError(
'CallMethod() given method descriptor for wrong service type.')
method = getattr(srvc, method_descriptor.name)
return method(rpc_controller, request, callback)
def _GetRequestClass(self, method_descriptor):
"""Returns the class of the request protocol message.
Args:
method_descriptor: Descriptor of the method for which to return the
request protocol message class.
Returns:
A class that represents the input protocol message of the specified
method.
"""
if method_descriptor.containing_service != self.descriptor:
raise RuntimeError(
'GetRequestClass() given method descriptor for wrong service type.')
return method_descriptor.input_type._concrete_class
def _GetResponseClass(self, method_descriptor):
"""Returns the class of the response protocol message.
Args:
method_descriptor: Descriptor of the method for which to return the
response protocol message class.
Returns:
A class that represents the output protocol message of the specified
method.
"""
if method_descriptor.containing_service != self.descriptor:
raise RuntimeError(
'GetResponseClass() given method descriptor for wrong service type.')
return method_descriptor.output_type._concrete_class
def _GenerateNonImplementedMethod(self, method):
"""Generates and returns a method that can be set for a service methods.
Args:
method: Descriptor of the service method for which a method is to be
generated.
Returns:
A method that can be added to the service class.
"""
return lambda inst, rpc_controller, request, callback: (
self._NonImplementedMethod(method.name, rpc_controller, callback))
def _NonImplementedMethod(self, method_name, rpc_controller, callback):
"""The body of all methods in the generated service class.
Args:
method_name: Name of the method being executed.
rpc_controller: RPC controller used to execute this method.
callback: A callback which will be invoked when the method finishes.
"""
rpc_controller.SetFailed('Method %s not implemented.' % method_name)
callback(None)
class _ServiceStubBuilder(object):
"""Constructs a protocol service stub class using a service descriptor.
Given a service descriptor, this class constructs a suitable stub class.
A stub is just a type-safe wrapper around an RpcChannel which emulates a
local implementation of the service.
One service stub builder instance constructs exactly one class. It means all
instances of that class share the same service stub builder.
"""
def __init__(self, service_descriptor):
"""Initializes an instance of the service stub class builder.
Args:
service_descriptor: ServiceDescriptor to use when constructing the
stub class.
"""
self.descriptor = service_descriptor
def BuildServiceStub(self, cls):
"""Constructs the stub class.
Args:
cls: The class that will be constructed.
"""
def _ServiceStubInit(stub, rpc_channel):
stub.rpc_channel = rpc_channel
self.cls = cls
cls.__init__ = _ServiceStubInit
for method in self.descriptor.methods:
setattr(cls, method.name, self._GenerateStubMethod(method))
def _GenerateStubMethod(self, method):
return (lambda inst, rpc_controller, request, callback=None:
self._StubMethod(inst, method, rpc_controller, request, callback))
def _StubMethod(self, stub, method_descriptor,
rpc_controller, request, callback):
"""The body of all service methods in the generated stub class.
Args:
stub: Stub instance.
method_descriptor: Descriptor of the invoked method.
rpc_controller: Rpc controller to execute the method.
request: Request protocol message.
callback: A callback to execute when the method finishes.
Returns:
Response message (in case of blocking call).
"""
return stub.rpc_channel.CallMethod(
method_descriptor, rpc_controller, request,
method_descriptor.output_type._concrete_class, callback)

View File

@ -0,0 +1,691 @@
# Protocol Buffers - Google's data interchange format
# Copyright 2008 Google Inc. All rights reserved.
# http://code.google.com/p/protobuf/
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following disclaimer
# in the documentation and/or other materials provided with the
# distribution.
# * Neither the name of Google Inc. nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Contains routines for printing protocol messages in text format."""
__author__ = 'kenton@google.com (Kenton Varda)'
import cStringIO
import re
from collections import deque
from google.protobuf.internal import type_checkers
from google.protobuf import descriptor
__all__ = [ 'MessageToString', 'PrintMessage', 'PrintField',
'PrintFieldValue', 'Merge' ]
# Infinity and NaN are not explicitly supported by Python pre-2.6, and
# float('inf') does not work on Windows (pre-2.6).
_INFINITY = 1e10000 # overflows, thus will actually be infinity.
_NAN = _INFINITY * 0
class ParseError(Exception):
"""Thrown in case of ASCII parsing error."""
def MessageToString(message, as_utf8=False, as_one_line=False):
out = cStringIO.StringIO()
PrintMessage(message, out, as_utf8=as_utf8, as_one_line=as_one_line)
result = out.getvalue()
out.close()
if as_one_line:
return result.rstrip()
return result
def PrintMessage(message, out, indent=0, as_utf8=False, as_one_line=False):
for field, value in message.ListFields():
if field.label == descriptor.FieldDescriptor.LABEL_REPEATED:
for element in value:
PrintField(field, element, out, indent, as_utf8, as_one_line)
else:
PrintField(field, value, out, indent, as_utf8, as_one_line)
def PrintField(field, value, out, indent=0, as_utf8=False, as_one_line=False):
"""Print a single field name/value pair. For repeated fields, the value
should be a single element."""
out.write(' ' * indent);
if field.is_extension:
out.write('[')
if (field.containing_type.GetOptions().message_set_wire_format and
field.type == descriptor.FieldDescriptor.TYPE_MESSAGE and
field.message_type == field.extension_scope and
field.label == descriptor.FieldDescriptor.LABEL_OPTIONAL):
out.write(field.message_type.full_name)
else:
out.write(field.full_name)
out.write(']')
elif field.type == descriptor.FieldDescriptor.TYPE_GROUP:
# For groups, use the capitalized name.
out.write(field.message_type.name)
else:
out.write(field.name)
if field.cpp_type != descriptor.FieldDescriptor.CPPTYPE_MESSAGE:
# The colon is optional in this case, but our cross-language golden files
# don't include it.
out.write(': ')
PrintFieldValue(field, value, out, indent, as_utf8, as_one_line)
if as_one_line:
out.write(' ')
else:
out.write('\n')
def PrintFieldValue(field, value, out, indent=0,
as_utf8=False, as_one_line=False):
"""Print a single field value (not including name). For repeated fields,
the value should be a single element."""
if field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_MESSAGE:
if as_one_line:
out.write(' { ')
PrintMessage(value, out, indent, as_utf8, as_one_line)
out.write('}')
else:
out.write(' {\n')
PrintMessage(value, out, indent + 2, as_utf8, as_one_line)
out.write(' ' * indent + '}')
elif field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_ENUM:
out.write(field.enum_type.values_by_number[value].name)
elif field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_STRING:
out.write('\"')
if type(value) is unicode:
out.write(_CEscape(value.encode('utf-8'), as_utf8))
else:
out.write(_CEscape(value, as_utf8))
out.write('\"')
elif field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_BOOL:
if value:
out.write("true")
else:
out.write("false")
else:
out.write(str(value))
def Merge(text, message):
"""Merges an ASCII representation of a protocol message into a message.
Args:
text: Message ASCII representation.
message: A protocol buffer message to merge into.
Raises:
ParseError: On ASCII parsing problems.
"""
tokenizer = _Tokenizer(text)
while not tokenizer.AtEnd():
_MergeField(tokenizer, message)
def _MergeField(tokenizer, message):
"""Merges a single protocol message field into a message.
Args:
tokenizer: A tokenizer to parse the field name and values.
message: A protocol message to record the data.
Raises:
ParseError: In case of ASCII parsing problems.
"""
message_descriptor = message.DESCRIPTOR
if tokenizer.TryConsume('['):
name = [tokenizer.ConsumeIdentifier()]
while tokenizer.TryConsume('.'):
name.append(tokenizer.ConsumeIdentifier())
name = '.'.join(name)
if not message_descriptor.is_extendable:
raise tokenizer.ParseErrorPreviousToken(
'Message type "%s" does not have extensions.' %
message_descriptor.full_name)
field = message.Extensions._FindExtensionByName(name)
if not field:
raise tokenizer.ParseErrorPreviousToken(
'Extension "%s" not registered.' % name)
elif message_descriptor != field.containing_type:
raise tokenizer.ParseErrorPreviousToken(
'Extension "%s" does not extend message type "%s".' % (
name, message_descriptor.full_name))
tokenizer.Consume(']')
else:
name = tokenizer.ConsumeIdentifier()
field = message_descriptor.fields_by_name.get(name, None)
# Group names are expected to be capitalized as they appear in the
# .proto file, which actually matches their type names, not their field
# names.
if not field:
field = message_descriptor.fields_by_name.get(name.lower(), None)
if field and field.type != descriptor.FieldDescriptor.TYPE_GROUP:
field = None
if (field and field.type == descriptor.FieldDescriptor.TYPE_GROUP and
field.message_type.name != name):
field = None
if not field:
raise tokenizer.ParseErrorPreviousToken(
'Message type "%s" has no field named "%s".' % (
message_descriptor.full_name, name))
if field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_MESSAGE:
tokenizer.TryConsume(':')
if tokenizer.TryConsume('<'):
end_token = '>'
else:
tokenizer.Consume('{')
end_token = '}'
if field.label == descriptor.FieldDescriptor.LABEL_REPEATED:
if field.is_extension:
sub_message = message.Extensions[field].add()
else:
sub_message = getattr(message, field.name).add()
else:
if field.is_extension:
sub_message = message.Extensions[field]
else:
sub_message = getattr(message, field.name)
sub_message.SetInParent()
while not tokenizer.TryConsume(end_token):
if tokenizer.AtEnd():
raise tokenizer.ParseErrorPreviousToken('Expected "%s".' % (end_token))
_MergeField(tokenizer, sub_message)
else:
_MergeScalarField(tokenizer, message, field)
def _MergeScalarField(tokenizer, message, field):
"""Merges a single protocol message scalar field into a message.
Args:
tokenizer: A tokenizer to parse the field value.
message: A protocol message to record the data.
field: The descriptor of the field to be merged.
Raises:
ParseError: In case of ASCII parsing problems.
RuntimeError: On runtime errors.
"""
tokenizer.Consume(':')
value = None
if field.type in (descriptor.FieldDescriptor.TYPE_INT32,
descriptor.FieldDescriptor.TYPE_SINT32,
descriptor.FieldDescriptor.TYPE_SFIXED32):
value = tokenizer.ConsumeInt32()
elif field.type in (descriptor.FieldDescriptor.TYPE_INT64,
descriptor.FieldDescriptor.TYPE_SINT64,
descriptor.FieldDescriptor.TYPE_SFIXED64):
value = tokenizer.ConsumeInt64()
elif field.type in (descriptor.FieldDescriptor.TYPE_UINT32,
descriptor.FieldDescriptor.TYPE_FIXED32):
value = tokenizer.ConsumeUint32()
elif field.type in (descriptor.FieldDescriptor.TYPE_UINT64,
descriptor.FieldDescriptor.TYPE_FIXED64):
value = tokenizer.ConsumeUint64()
elif field.type in (descriptor.FieldDescriptor.TYPE_FLOAT,
descriptor.FieldDescriptor.TYPE_DOUBLE):
value = tokenizer.ConsumeFloat()
elif field.type == descriptor.FieldDescriptor.TYPE_BOOL:
value = tokenizer.ConsumeBool()
elif field.type == descriptor.FieldDescriptor.TYPE_STRING:
value = tokenizer.ConsumeString()
elif field.type == descriptor.FieldDescriptor.TYPE_BYTES:
value = tokenizer.ConsumeByteString()
elif field.type == descriptor.FieldDescriptor.TYPE_ENUM:
# Enum can be specified by a number (the enum value), or by
# a string literal (the enum name).
enum_descriptor = field.enum_type
if tokenizer.LookingAtInteger():
number = tokenizer.ConsumeInt32()
enum_value = enum_descriptor.values_by_number.get(number, None)
if enum_value is None:
raise tokenizer.ParseErrorPreviousToken(
'Enum type "%s" has no value with number %d.' % (
enum_descriptor.full_name, number))
else:
identifier = tokenizer.ConsumeIdentifier()
enum_value = enum_descriptor.values_by_name.get(identifier, None)
if enum_value is None:
raise tokenizer.ParseErrorPreviousToken(
'Enum type "%s" has no value named %s.' % (
enum_descriptor.full_name, identifier))
value = enum_value.number
else:
raise RuntimeError('Unknown field type %d' % field.type)
if field.label == descriptor.FieldDescriptor.LABEL_REPEATED:
if field.is_extension:
message.Extensions[field].append(value)
else:
getattr(message, field.name).append(value)
else:
if field.is_extension:
message.Extensions[field] = value
else:
setattr(message, field.name, value)
class _Tokenizer(object):
"""Protocol buffer ASCII representation tokenizer.
This class handles the lower level string parsing by splitting it into
meaningful tokens.
It was directly ported from the Java protocol buffer API.
"""
_WHITESPACE = re.compile('(\\s|(#.*$))+', re.MULTILINE)
_TOKEN = re.compile(
'[a-zA-Z_][0-9a-zA-Z_+-]*|' # an identifier
'[0-9+-][0-9a-zA-Z_.+-]*|' # a number
'\"([^\"\n\\\\]|\\\\.)*(\"|\\\\?$)|' # a double-quoted string
'\'([^\'\n\\\\]|\\\\.)*(\'|\\\\?$)') # a single-quoted string
_IDENTIFIER = re.compile('\w+')
_INTEGER_CHECKERS = [type_checkers.Uint32ValueChecker(),
type_checkers.Int32ValueChecker(),
type_checkers.Uint64ValueChecker(),
type_checkers.Int64ValueChecker()]
_FLOAT_INFINITY = re.compile('-?inf(inity)?f?', re.IGNORECASE)
_FLOAT_NAN = re.compile("nanf?", re.IGNORECASE)
def __init__(self, text_message):
self._text_message = text_message
self._position = 0
self._line = -1
self._column = 0
self._token_start = None
self.token = ''
self._lines = deque(text_message.split('\n'))
self._current_line = ''
self._previous_line = 0
self._previous_column = 0
self._SkipWhitespace()
self.NextToken()
def AtEnd(self):
"""Checks the end of the text was reached.
Returns:
True iff the end was reached.
"""
return self.token == ''
def _PopLine(self):
while len(self._current_line) <= self._column:
if not self._lines:
self._current_line = ''
return
self._line += 1
self._column = 0
self._current_line = self._lines.popleft()
def _SkipWhitespace(self):
while True:
self._PopLine()
match = self._WHITESPACE.match(self._current_line, self._column)
if not match:
break
length = len(match.group(0))
self._column += length
def TryConsume(self, token):
"""Tries to consume a given piece of text.
Args:
token: Text to consume.
Returns:
True iff the text was consumed.
"""
if self.token == token:
self.NextToken()
return True
return False
def Consume(self, token):
"""Consumes a piece of text.
Args:
token: Text to consume.
Raises:
ParseError: If the text couldn't be consumed.
"""
if not self.TryConsume(token):
raise self._ParseError('Expected "%s".' % token)
def LookingAtInteger(self):
"""Checks if the current token is an integer.
Returns:
True iff the current token is an integer.
"""
if not self.token:
return False
c = self.token[0]
return (c >= '0' and c <= '9') or c == '-' or c == '+'
def ConsumeIdentifier(self):
"""Consumes protocol message field identifier.
Returns:
Identifier string.
Raises:
ParseError: If an identifier couldn't be consumed.
"""
result = self.token
if not self._IDENTIFIER.match(result):
raise self._ParseError('Expected identifier.')
self.NextToken()
return result
def ConsumeInt32(self):
"""Consumes a signed 32bit integer number.
Returns:
The integer parsed.
Raises:
ParseError: If a signed 32bit integer couldn't be consumed.
"""
try:
result = self._ParseInteger(self.token, is_signed=True, is_long=False)
except ValueError, e:
raise self._IntegerParseError(e)
self.NextToken()
return result
def ConsumeUint32(self):
"""Consumes an unsigned 32bit integer number.
Returns:
The integer parsed.
Raises:
ParseError: If an unsigned 32bit integer couldn't be consumed.
"""
try:
result = self._ParseInteger(self.token, is_signed=False, is_long=False)
except ValueError, e:
raise self._IntegerParseError(e)
self.NextToken()
return result
def ConsumeInt64(self):
"""Consumes a signed 64bit integer number.
Returns:
The integer parsed.
Raises:
ParseError: If a signed 64bit integer couldn't be consumed.
"""
try:
result = self._ParseInteger(self.token, is_signed=True, is_long=True)
except ValueError, e:
raise self._IntegerParseError(e)
self.NextToken()
return result
def ConsumeUint64(self):
"""Consumes an unsigned 64bit integer number.
Returns:
The integer parsed.
Raises:
ParseError: If an unsigned 64bit integer couldn't be consumed.
"""
try:
result = self._ParseInteger(self.token, is_signed=False, is_long=True)
except ValueError, e:
raise self._IntegerParseError(e)
self.NextToken()
return result
def ConsumeFloat(self):
"""Consumes an floating point number.
Returns:
The number parsed.
Raises:
ParseError: If a floating point number couldn't be consumed.
"""
text = self.token
if self._FLOAT_INFINITY.match(text):
self.NextToken()
if text.startswith('-'):
return -_INFINITY
return _INFINITY
if self._FLOAT_NAN.match(text):
self.NextToken()
return _NAN
try:
result = float(text)
except ValueError, e:
raise self._FloatParseError(e)
self.NextToken()
return result
def ConsumeBool(self):
"""Consumes a boolean value.
Returns:
The bool parsed.
Raises:
ParseError: If a boolean value couldn't be consumed.
"""
if self.token in ('true', 't', '1'):
self.NextToken()
return True
elif self.token in ('false', 'f', '0'):
self.NextToken()
return False
else:
raise self._ParseError('Expected "true" or "false".')
def ConsumeString(self):
"""Consumes a string value.
Returns:
The string parsed.
Raises:
ParseError: If a string value couldn't be consumed.
"""
bytes = self.ConsumeByteString()
try:
return unicode(bytes, 'utf-8')
except UnicodeDecodeError, e:
raise self._StringParseError(e)
def ConsumeByteString(self):
"""Consumes a byte array value.
Returns:
The array parsed (as a string).
Raises:
ParseError: If a byte array value couldn't be consumed.
"""
list = [self._ConsumeSingleByteString()]
while len(self.token) > 0 and self.token[0] in ('\'', '"'):
list.append(self._ConsumeSingleByteString())
return "".join(list)
def _ConsumeSingleByteString(self):
"""Consume one token of a string literal.
String literals (whether bytes or text) can come in multiple adjacent
tokens which are automatically concatenated, like in C or Python. This
method only consumes one token.
"""
text = self.token
if len(text) < 1 or text[0] not in ('\'', '"'):
raise self._ParseError('Exptected string.')
if len(text) < 2 or text[-1] != text[0]:
raise self._ParseError('String missing ending quote.')
try:
result = _CUnescape(text[1:-1])
except ValueError, e:
raise self._ParseError(str(e))
self.NextToken()
return result
def _ParseInteger(self, text, is_signed=False, is_long=False):
"""Parses an integer.
Args:
text: The text to parse.
is_signed: True if a signed integer must be parsed.
is_long: True if a long integer must be parsed.
Returns:
The integer value.
Raises:
ValueError: Thrown Iff the text is not a valid integer.
"""
pos = 0
if text.startswith('-'):
pos += 1
base = 10
if text.startswith('0x', pos) or text.startswith('0X', pos):
base = 16
elif text.startswith('0', pos):
base = 8
# Do the actual parsing. Exception handling is propagated to caller.
result = int(text, base)
# Check if the integer is sane. Exceptions handled by callers.
checker = self._INTEGER_CHECKERS[2 * int(is_long) + int(is_signed)]
checker.CheckValue(result)
return result
def ParseErrorPreviousToken(self, message):
"""Creates and *returns* a ParseError for the previously read token.
Args:
message: A message to set for the exception.
Returns:
A ParseError instance.
"""
return ParseError('%d:%d : %s' % (
self._previous_line + 1, self._previous_column + 1, message))
def _ParseError(self, message):
"""Creates and *returns* a ParseError for the current token."""
return ParseError('%d:%d : %s' % (
self._line + 1, self._column - len(self.token) + 1, message))
def _IntegerParseError(self, e):
return self._ParseError('Couldn\'t parse integer: ' + str(e))
def _FloatParseError(self, e):
return self._ParseError('Couldn\'t parse number: ' + str(e))
def _StringParseError(self, e):
return self._ParseError('Couldn\'t parse string: ' + str(e))
def NextToken(self):
"""Reads the next meaningful token."""
self._previous_line = self._line
self._previous_column = self._column
self._column += len(self.token)
self._SkipWhitespace()
if not self._lines and len(self._current_line) <= self._column:
self.token = ''
return
match = self._TOKEN.match(self._current_line, self._column)
if match:
token = match.group(0)
self.token = token
else:
self.token = self._current_line[self._column]
# text.encode('string_escape') does not seem to satisfy our needs as it
# encodes unprintable characters using two-digit hex escapes whereas our
# C++ unescaping function allows hex escapes to be any length. So,
# "\0011".encode('string_escape') ends up being "\\x011", which will be
# decoded in C++ as a single-character string with char code 0x11.
def _CEscape(text, as_utf8):
def escape(c):
o = ord(c)
if o == 10: return r"\n" # optional escape
if o == 13: return r"\r" # optional escape
if o == 9: return r"\t" # optional escape
if o == 39: return r"\'" # optional escape
if o == 34: return r'\"' # necessary escape
if o == 92: return r"\\" # necessary escape
# necessary escapes
if not as_utf8 and (o >= 127 or o < 32): return "\\%03o" % o
return c
return "".join([escape(c) for c in text])
_CUNESCAPE_HEX = re.compile('\\\\x([0-9a-fA-F]{2}|[0-9a-fA-F])')
def _CUnescape(text):
def ReplaceHex(m):
return chr(int(m.group(0)[2:], 16))
# This is required because the 'string_escape' encoding doesn't
# allow single-digit hex escapes (like '\xf').
result = _CUNESCAPE_HEX.sub(ReplaceHex, text)
return result.decode('string_escape')