Can Fast-DDS-python do raw serialize/deserialize?

[jordancharest]

My main application is in C++ and we log all messages. I have some custom log analysis tools written in C++ and I would like to start using python instead because it's substantially easier to add visualizations. For every message, I write to the log:
metadata | message size in bytes | serialized message

Note that the serialized message is the actual message, not the associated PubSubType. I serialize like so:

template <typename MsgType>
uint16_t MessageWriter::serialize(char* buffer, size_t buffer_size, const MsgType& data) {
    eprosima::fastcdr::FastBuffer cdr_buffer(buffer, buffer_size);
    eprosima::fastcdr::Cdr serializer(cdr_buffer);
    eprosima::fastcdr::serialize(serializer, data);
    return static_cast<uint16_t>(serializer.get_serialized_data_length());
}

Metadata is of the form:

struct MsgMetadata {
    MsgHeader header; // timestamp
    long type_id;       // ID identifying the data type of the message.
    string<256> topic;  // The topic the message was sent on.
};

So with that and the size of the serialized message, it's easy to deserialize:

#define CASE(topic_name, MsgType, id) \
                case id: {                        \
                    MsgType msg;                  \
                    deserialize(buffer, msg);     \
                    return msg;                   \
                }
EXPAND_ALL_TOPICS(CASE)

template <typename MsgType>
void deserialize(std::string& buffer, MsgType& msg) {
    eprosima::fastcdr::FastBuffer msg_buffer(buffer.data(), buffer.size());
    eprosima::fastcdr::Cdr msg_deserializer(msg_buffer);
    eprosima::fastcdr::deserialize(msg_deserializer, msg);
}

How do I replicate deserialize in Python? Given a topic name, the message type, the size of the serialized message, and a serialized buffer, can it be done?