Python py3compat.StringIO类代码示例

        def surcor_callback(res,error):
            data = StringIO()
            for item in res:
                data.write(item['data'])

            fileobj['data'] = data.getvalue() 
            callback(fileobj)

def _encode_batched_write_command(
        namespace, operation, command, docs, check_keys, opts, ctx):
    """Encode the next batched insert, update, or delete command.
    """
    buf = StringIO()

    to_send, _ = _batched_write_command(
        namespace, operation, command, docs, check_keys, opts, ctx, buf)
    return buf.getvalue(), to_send

    def readline(self, size=-1):
        """Read one line or up to `size` bytes from the file.

        :Parameters:
         - `size` (optional): the maximum number of bytes to read

        .. versionadded:: 1.9
        """
        if size == 0:
            return b('')

        remainder = int(self.length) - self.__position
        if size < 0 or size > remainder:
            size = remainder

        received = 0
        data = StringIO()
        while received < size:
            chunk_data = self.readchunk()
            pos = chunk_data.find(NEWLN, 0, size)
            if pos != -1:
                size = received + pos + 1

            received += len(chunk_data)
            data.write(chunk_data)
            if pos != -1:
                break

        self.__position -= received - size

        # Return 'size' bytes and store the rest.
        data.seek(size)
        self.__buffer = data.read()
        data.seek(0)
        return data.read(size)

    def read(self, size=-1):
        """Read at most `size` bytes from the file (less if there
        isn't enough data).

        The bytes are returned as an instance of :class:`str` (:class:`bytes`
        in python 3). If `size` is negative or omitted all data is read.

        :Parameters:
          - `size` (optional): the number of bytes to read
        """
        self._ensure_file()

        if size == 0:
            return EMPTY

        remainder = int(self.length) - self.__position
        if size < 0 or size > remainder:
            size = remainder

        received = 0
        data = StringIO()
        while received < size:
            chunk_data = self.readchunk()
            received += len(chunk_data)
            data.write(chunk_data)

        self.__position -= received - size

        # Return 'size' bytes and store the rest.
        data.seek(size)
        self.__buffer = data.read()
        data.seek(0)
        return data.read(size)

        def write_me(s, chunk_size):
            buf = StringIO(s)
            infile = GridIn(self.db.fs)
            while True:
                to_write = buf.read(chunk_size)
                if to_write == b(''):
                    break
                infile.write(to_write)
            infile.close()
            buf.close()

            outfile = GridOut(self.db.fs, infile._id)
            data = outfile.read()
            self.assertEqual(s, data)

    def test_zip(self):
        zf = StringIO()
        z = zipfile.ZipFile(zf, "w")
        z.writestr("test.txt", b"hello world")
        z.close()
        zf.seek(0)

        f = GridIn(self.db.fs, filename="test.zip")
        f.write(zf)
        f.close()
        self.assertEqual(1, self.db.fs.files.count_documents({}))
        self.assertEqual(1, self.db.fs.chunks.count_documents({}))

        g = GridOut(self.db.fs, f._id)
        z = zipfile.ZipFile(g)
        self.assertSequenceEqual(z.namelist(), ["test.txt"])
        self.assertEqual(z.read("test.txt"), b"hello world")

    def read(self, size=-1):
        """Read at most `size` bytes from the file (less if there
        isn't enough data).

        The bytes are returned as an instance of :class:`str` (:class:`bytes`
        in python 3). If `size` is negative or omitted all data is read.

        :Parameters:
          - `size` (optional): the number of bytes to read

        .. versionchanged:: 3.8
           This method now only checks for extra chunks after reading the
           entire file. Previously, this method would check for extra chunks
           on every call.
        """
        self._ensure_file()

        remainder = int(self.length) - self.__position
        if size < 0 or size > remainder:
            size = remainder

        if size == 0:
            return EMPTY

        received = 0
        data = StringIO()
        while received < size:
            chunk_data = self.readchunk()
            received += len(chunk_data)
            data.write(chunk_data)

        # Detect extra chunks after reading the entire file.
        if size == remainder and self.__chunk_iter:
            try:
                self.__chunk_iter.next()
            except StopIteration:
                pass

        self.__position -= received - size

        # Return 'size' bytes and store the rest.
        data.seek(size)
        self.__buffer = data.read()
        data.seek(0)
        return data.read(size)

    def read(self, size=-1):
        """Read at most `size` bytes from the file (less if there
        isn't enough data).

        The bytes are returned as an instance of :class:`str` (:class:`bytes`
        in python 3). If `size` is negative or omitted all data is read.

        :Parameters:
          - `size` (optional): the number of bytes to read
        """
        self._ensure_file()

        if size == 0:
            return EMPTY

        remainder = int(self.length) - self.__position
        if size < 0 or size > remainder:
            size = remainder

        received = 0
        data = StringIO()
        while received < size:
            chunk_data = self.readchunk()
            received += len(chunk_data)
            data.write(chunk_data)

        # Detect extra chunks.
        max_chunk_n = math.ceil(self.length / float(self.chunk_size))
        chunk = self.__chunks.find_one({"files_id": self._id,
                                        "n": {"$gte": max_chunk_n}},
                                       session=self._session)
        # According to spec, ignore extra chunks if they are empty.
        if chunk is not None and len(chunk['data']):
            raise CorruptGridFile(
                "Extra chunk found: expected %i chunks but found "
                "chunk with n=%i" % (max_chunk_n, chunk['n']))

        self.__position -= received - size

        # Return 'size' bytes and store the rest.
        data.seek(size)
        self.__buffer = data.read()
        data.seek(0)
        return data.read(size)

    def __init__(self, client, root_collection, **kwargs):
        self.client = client
        self.root_collection = root_collection
        self._files = self.client.connection(files_coll(self.root_collection))
        self._chunks = self.client.connection(chunks_coll(self.root_collection))

        # Handle alternative naming
        if "content_type" in kwargs:
            kwargs["contentType"] = kwargs.pop("content_type")
        if "chunk_size" in kwargs:
            kwargs["chunkSize"] = kwargs.pop("chunk_size")
        
        # Defaults
        kwargs["_id"] = kwargs.get("_id", ObjectId())
        kwargs["chunkSize"] = kwargs.get("chunkSize", DEFAULT_CHUNK_SIZE)

        self._file = kwargs
        self._chunk_number = 0
        self._position = 0
        self._buffer = StringIO()

def _do_batched_write_command(namespace, operation, command,
                              docs, check_keys, opts, ctx):
    """Execute a batch of insert, update, or delete commands.
    """
    max_bson_size = ctx.max_bson_size
    max_write_batch_size = ctx.max_write_batch_size
    # Max BSON object size + 16k - 2 bytes for ending NUL bytes.
    # Server guarantees there is enough room: SERVER-10643.
    max_cmd_size = max_bson_size + _COMMAND_OVERHEAD

    ordered = command.get('ordered', True)

    buf = StringIO()
    # Save space for message length and request id
    buf.write(_ZERO_64)
    # responseTo, opCode
    buf.write(b"\x00\x00\x00\x00\xd4\x07\x00\x00")
    # No options
    buf.write(_ZERO_32)
    # Namespace as C string
    buf.write(b(namespace))
    buf.write(_ZERO_8)
    # Skip: 0, Limit: -1
    buf.write(_SKIPLIM)

    # Where to write command document length
    command_start = buf.tell()
    buf.write(bson.BSON.encode(command))

    # Start of payload
    buf.seek(-1, 2)
    # Work around some Jython weirdness.
    buf.truncate()
    try:
        buf.write(_OP_MAP[operation])
    except KeyError:
        raise InvalidOperation('Unknown command')

    if operation in (_UPDATE, _DELETE):
        check_keys = False

    # Where to write list document length
    list_start = buf.tell() - 4

    to_send = []

    def send_message():
        """Finalize and send the current OP_QUERY message.
        """
        # Close list and command documents
        buf.write(_ZERO_16)

        # Write document lengths and request id
        length = buf.tell()
        buf.seek(list_start)
        buf.write(struct.pack('<i', length - list_start - 1))
        buf.seek(command_start)
        buf.write(struct.pack('<i', length - command_start))
        buf.seek(4)
        request_id = _randint()
        buf.write(struct.pack('<i', request_id))
        buf.seek(0)
        buf.write(struct.pack('<i', length))
        return ctx.write_command(request_id, buf.getvalue(), to_send)

    # If there are multiple batches we'll
    # merge results in the caller.
    results = []

    idx = 0
    idx_offset = 0
    has_docs = False
    for doc in docs:
        has_docs = True
        # Encode the current operation
        key = b(str(idx))
        value = bson.BSON.encode(doc, check_keys, opts)
        # Send a batch?
        enough_data = (buf.tell() + len(key) + len(value) + 2) >= max_cmd_size
        enough_documents = (idx >= max_write_batch_size)
        if enough_data or enough_documents:
            if not idx:
                write_op = "insert" if operation == _INSERT else None
                _raise_document_too_large(
                    write_op, len(value), max_bson_size)
            result = send_message()
            results.append((idx_offset, result))
            if ordered and "writeErrors" in result:
                return results

            # Truncate back to the start of list elements
            buf.seek(list_start + 4)
            buf.truncate()
            idx_offset += idx
            idx = 0
            key = b'0'
            to_send = []
        buf.write(_BSONOBJ)
        buf.write(key)
        buf.write(_ZERO_8)
#.........这里部分代码省略.........

class GridIn(object):
    """Class to write data to GridFS.
    """
    def __init__(self, root_collection, **kwargs):
        """Write a file to GridFS

        Application developers should generally not need to
        instantiate this class directly - instead see the methods
        provided by :class:`~gridfs.GridFS`.

        Raises :class:`TypeError` if `root_collection` is not an
        instance of :class:`~pymongo.collection.Collection`.

        Any of the file level options specified in the `GridFS Spec
        <http://dochub.mongodb.org/core/gridfsspec>`_ may be passed as
        keyword arguments. Any additional keyword arguments will be
        set as additional fields on the file document. Valid keyword
        arguments include:

          - ``"_id"``: unique ID for this file (default:
            :class:`~bson.objectid.ObjectId`) - this ``"_id"`` must
            not have already been used for another file

          - ``"filename"``: human name for the file

          - ``"contentType"`` or ``"content_type"``: valid mime-type
            for the file

          - ``"chunkSize"`` or ``"chunk_size"``: size of each of the
            chunks, in bytes (default: 255 kb)

          - ``"encoding"``: encoding used for this file. In Python 2,
            any :class:`unicode` that is written to the file will be
            converted to a :class:`str`. In Python 3, any :class:`str`
            that is written to the file will be converted to
            :class:`bytes`.

        :Parameters:
          - `root_collection`: root collection to write to
          - `**kwargs` (optional): file level options (see above)

        .. versionchanged:: 3.0
           `root_collection` must use an acknowledged
           :attr:`~pymongo.collection.Collection.write_concern`
        """
        if not isinstance(root_collection, Collection):
            raise TypeError("root_collection must be an "
                            "instance of Collection")

        # With w=0, 'filemd5' might run before the final chunks are written.
        if not root_collection.write_concern.acknowledged:
            raise ConfigurationError('root_collection must use '
                                     'acknowledged write_concern')

        # Handle alternative naming
        if "content_type" in kwargs:
            kwargs["contentType"] = kwargs.pop("content_type")
        if "chunk_size" in kwargs:
            kwargs["chunkSize"] = kwargs.pop("chunk_size")

        kwargs['md5'] = md5()
        # Defaults
        kwargs["_id"] = kwargs.get("_id", ObjectId())
        kwargs["chunkSize"] = kwargs.get("chunkSize", DEFAULT_CHUNK_SIZE)
        object.__setattr__(self, "_coll", root_collection)
        object.__setattr__(self, "_chunks", root_collection.chunks)
        object.__setattr__(self, "_file", kwargs)
        object.__setattr__(self, "_buffer", StringIO())
        object.__setattr__(self, "_position", 0)
        object.__setattr__(self, "_chunk_number", 0)
        object.__setattr__(self, "_closed", False)
        object.__setattr__(self, "_ensured_index", False)

    def __create_index(self, collection, index, unique):
        doc = collection.find_one(projection={"_id": 1})
        if doc is None:
            try:
                indexes = list(collection.list_indexes())
            except OperationFailure:
                indexes = []
            if index not in indexes:
                collection.create_index(index, unique=unique)

    def __ensure_indexes(self):
        if not object.__getattribute__(self, "_ensured_index"):
            self.__create_index(self._coll.files, _F_INDEX, False)
            self.__create_index(self._coll.chunks, _C_INDEX, True)
            object.__setattr__(self, "_ensured_index", True)

    def abort(self):
        """Remove all chunks/files that may have been uploaded and close.
        """
        self._coll.chunks.delete_many({"files_id": self._file['_id']})
        self._coll.files.delete_one({"_id": self._file['_id']})
        object.__setattr__(self, "_closed", True)


    @property
    def closed(self):
        """Is this file closed?
#.........这里部分代码省略.........

 def __flush_buffer(self):
     """Flush the buffer contents out to a chunk.
     """
     self.__flush_data(self._buffer.getvalue())
     self._buffer.close()
     self._buffer = StringIO()

    def test_download_to_stream_by_name(self):
        file1 = StringIO(b"hello world")
        # Test with one chunk.
        oid = self.fs.upload_from_stream("one_chunk", file1)
        self.assertEqual(1, self.db.fs.chunks.count())
        file2 = StringIO()
        self.fs.download_to_stream_by_name("one_chunk", file2)
        file1.seek(0)
        file2.seek(0)
        self.assertEqual(file1.read(), file2.read())

        # Test with many chunks.
        self.db.drop_collection("fs.files")
        self.db.drop_collection("fs.chunks")
        file1.seek(0)
        self.fs.upload_from_stream("many_chunks", file1, chunk_size_bytes=1)
        self.assertEqual(11, self.db.fs.chunks.count())

        file2 = StringIO()
        self.fs.download_to_stream_by_name("many_chunks", file2)
        file1.seek(0)
        file2.seek(0)
        self.assertEqual(file1.read(), file2.read())

class GridIn(object):
    """Class to write data to GridFS.
    """
    def __init__(self, root_collection, **kwargs):
        """Write a file to GridFS

        Application developers should generally not need to
        instantiate this class directly - instead see the methods
        provided by :class:`~gridfs.GridFS`.

        Raises :class:`TypeError` if `root_collection` is not an
        instance of :class:`~pymongo.collection.Collection`.

        Any of the file level options specified in the `GridFS Spec
        <http://dochub.mongodb.org/core/gridfsspec>`_ may be passed as
        keyword arguments. Any additional keyword arguments will be
        set as additional fields on the file document. Valid keyword
        arguments include:

          - ``"_id"``: unique ID for this file (default:
            :class:`~bson.objectid.ObjectId`) - this ``"_id"`` must
            not have already been used for another file

          - ``"filename"``: human name for the file

          - ``"contentType"`` or ``"content_type"``: valid mime-type
            for the file

          - ``"chunkSize"`` or ``"chunk_size"``: size of each of the
            chunks, in bytes (default: 256 kb)

          - ``"encoding"``: encoding used for this file. In Python 2,
            any :class:`unicode` that is written to the file will be
            converted to a :class:`str`. In Python 3, any :class:`str`
            that is written to the file will be converted to
            :class:`bytes`.

        If you turn off write-acknowledgment for performance reasons, it is
        critical to wrap calls to :meth:`write` and :meth:`close` within a
        single request:

           >>> from pymongo import MongoClient
           >>> from gridfs import GridFS
           >>> client = MongoClient(w=0) # turn off write acknowledgment
           >>> fs = GridFS(client.database)
           >>> gridin = fs.new_file()
           >>> request = client.start_request()
           >>> try:
           ...     for i in range(10):
           ...         gridin.write('foo')
           ...     gridin.close()
           ... finally:
           ...     request.end()

        In Python 2.5 and later this code can be simplified with a
        with-statement, see :doc:`/examples/requests` for more information.

        :Parameters:
          - `root_collection`: root collection to write to
          - `**kwargs` (optional): file level options (see above)
        """
        if not isinstance(root_collection, Collection):
            raise TypeError("root_collection must be an "
                            "instance of Collection")

        # Handle alternative naming
        if "content_type" in kwargs:
            kwargs["contentType"] = kwargs.pop("content_type")
        if "chunk_size" in kwargs:
            kwargs["chunkSize"] = kwargs.pop("chunk_size")

        # Defaults
        kwargs["_id"] = kwargs.get("_id", ObjectId())
        kwargs["chunkSize"] = kwargs.get("chunkSize", DEFAULT_CHUNK_SIZE)
        object.__setattr__(self, "_coll", root_collection)
        object.__setattr__(self, "_chunks", root_collection.chunks)
        object.__setattr__(self, "_file", kwargs)
        object.__setattr__(self, "_buffer", StringIO())
        object.__setattr__(self, "_position", 0)
        object.__setattr__(self, "_chunk_number", 0)
        object.__setattr__(self, "_closed", False)
        object.__setattr__(self, "_ensured_index", False)

    def _ensure_index(self):
        if not object.__getattribute__(self, "_ensured_index"):
            self._coll.chunks.ensure_index(
                [("files_id", ASCENDING), ("n", ASCENDING)],
                unique=True)
            object.__setattr__(self, "_ensured_index", True)

    @property
    def closed(self):
        """Is this file closed?
        """
        return self._closed

    _id = _create_property("_id", "The ``'_id'`` value for this file.",
                            read_only=True)
    filename = _create_property("filename", "Name of this file.")
    name = _create_property("filename", "Alias for `filename`.")
#.........这里部分代码省略.........

def _do_batched_write_command(namespace, operation, command,
                              docs, check_keys, opts, ctx):
    """Create the next batched insert, update, or delete command.
    """
    max_bson_size = ctx.max_bson_size
    max_write_batch_size = ctx.max_write_batch_size
    # Max BSON object size + 16k - 2 bytes for ending NUL bytes.
    # Server guarantees there is enough room: SERVER-10643.
    max_cmd_size = max_bson_size + _COMMAND_OVERHEAD

    buf = StringIO()
    # Save space for message length and request id
    buf.write(_ZERO_64)
    # responseTo, opCode
    buf.write(b"\x00\x00\x00\x00\xd4\x07\x00\x00")
    # No options
    buf.write(_ZERO_32)
    # Namespace as C string
    buf.write(b(namespace))
    buf.write(_ZERO_8)
    # Skip: 0, Limit: -1
    buf.write(_SKIPLIM)

    # Where to write command document length
    command_start = buf.tell()
    buf.write(bson.BSON.encode(command))

    # Start of payload
    buf.seek(-1, 2)
    # Work around some Jython weirdness.
    buf.truncate()
    try:
        buf.write(_OP_MAP[operation])
    except KeyError:
        raise InvalidOperation('Unknown command')

    if operation in (_UPDATE, _DELETE):
        check_keys = False

    # Where to write list document length
    list_start = buf.tell() - 4
    to_send = []
    idx = 0
    for doc in docs:
        # Encode the current operation
        key = b(str(idx))
        value = bson.BSON.encode(doc, check_keys, opts)
        # Is there enough room to add this document? max_cmd_size accounts for
        # the two trailing null bytes.
        enough_data = (buf.tell() + len(key) + len(value)) >= max_cmd_size
        enough_documents = (idx >= max_write_batch_size)
        if enough_data or enough_documents:
            if not idx:
                write_op = "insert" if operation == _INSERT else None
                _raise_document_too_large(
                    write_op, len(value), max_bson_size)
            break
        buf.write(_BSONOBJ)
        buf.write(key)
        buf.write(_ZERO_8)
        buf.write(value)
        to_send.append(doc)
        idx += 1

    # Finalize the current OP_QUERY message.
    # Close list and command documents
    buf.write(_ZERO_16)

    # Write document lengths and request id
    length = buf.tell()
    buf.seek(list_start)
    buf.write(struct.pack('<i', length - list_start - 1))
    buf.seek(command_start)
    buf.write(struct.pack('<i', length - command_start))
    buf.seek(4)
    request_id = _randint()
    buf.write(struct.pack('<i', request_id))
    buf.seek(0)
    buf.write(struct.pack('<i', length))

    return request_id, buf.getvalue(), to_send