DB-GPT/dbgpt/core/interface/message.py

from __future__ import annotations

from abc import ABC, abstractmethod
from typing import Dict, List, Tuple, Union, Optional
from datetime import datetime

from dbgpt._private.pydantic import BaseModel, Field

from dbgpt.core.interface.storage import (
    ResourceIdentifier,
    StorageItem,
    StorageInterface,
    InMemoryStorage,
)


class BaseMessage(BaseModel, ABC):
    """Message object."""

    content: str
    index: int = 0
    round_index: int = 0
    """The round index of the message in the conversation"""
    additional_kwargs: dict = Field(default_factory=dict)

    @property
    @abstractmethod
    def type(self) -> str:
        """Type of the message, used for serialization."""

    @property
    def pass_to_model(self) -> bool:
        """Whether the message will be passed to the model"""
        return True

    def to_dict(self) -> Dict:
        """Convert to dict

        Returns:
            Dict: The dict object
        """
        return {
            "type": self.type,
            "data": self.dict(),
            "index": self.index,
            "round_index": self.round_index,
        }


class HumanMessage(BaseMessage):
    """Type of message that is spoken by the human."""

    example: bool = False

    @property
    def type(self) -> str:
        """Type of the message, used for serialization."""
        return "human"


class AIMessage(BaseMessage):
    """Type of message that is spoken by the AI."""

    example: bool = False

    @property
    def type(self) -> str:
        """Type of the message, used for serialization."""
        return "ai"


class ViewMessage(BaseMessage):
    """Type of message that is spoken by the AI."""

    example: bool = False

    @property
    def type(self) -> str:
        """Type of the message, used for serialization."""
        return "view"

    @property
    def pass_to_model(self) -> bool:
        """Whether the message will be passed to the model

        The view message will not be passed to the model
        """
        return False


class SystemMessage(BaseMessage):
    """Type of message that is a system message."""

    @property
    def type(self) -> str:
        """Type of the message, used for serialization."""
        return "system"


class ModelMessageRoleType:
    """ "Type of ModelMessage role"""

    SYSTEM = "system"
    HUMAN = "human"
    AI = "ai"
    VIEW = "view"


class ModelMessage(BaseModel):
    """Type of message that interaction between dbgpt-server and llm-server"""

    """Similar to openai's message format"""
    role: str
    content: str

    @staticmethod
    def from_openai_messages(
        messages: Union[str, List[Dict[str, str]]]
    ) -> List["ModelMessage"]:
        """Openai message format to current ModelMessage format"""
        if isinstance(messages, str):
            return [ModelMessage(role=ModelMessageRoleType.HUMAN, content=messages)]
        result = []
        for message in messages:
            msg_role = message["role"]
            content = message["content"]
            if msg_role == "system":
                result.append(
                    ModelMessage(role=ModelMessageRoleType.SYSTEM, content=content)
                )
            elif msg_role == "user":
                result.append(
                    ModelMessage(role=ModelMessageRoleType.HUMAN, content=content)
                )
            elif msg_role == "assistant":
                result.append(
                    ModelMessage(role=ModelMessageRoleType.AI, content=content)
                )
            else:
                raise ValueError(f"Unknown role: {msg_role}")
        return result

    @staticmethod
    def to_openai_messages(messages: List["ModelMessage"]) -> List[Dict[str, str]]:
        """Convert to OpenAI message format and
        hugggingface [Templates of Chat Models](https://huggingface.co/docs/transformers/v4.34.1/en/chat_templating)
        """
        history = []
        # Add history conversation
        for message in messages:
            if message.role == ModelMessageRoleType.HUMAN:
                history.append({"role": "user", "content": message.content})
            elif message.role == ModelMessageRoleType.SYSTEM:
                history.append({"role": "system", "content": message.content})
            elif message.role == ModelMessageRoleType.AI:
                history.append({"role": "assistant", "content": message.content})
            else:
                pass
        # Move the last user's information to the end
        last_user_input_index = None
        for i in range(len(history) - 1, -1, -1):
            if history[i]["role"] == "user":
                last_user_input_index = i
                break
        if last_user_input_index:
            last_user_input = history.pop(last_user_input_index)
            history.append(last_user_input)
        return history

    @staticmethod
    def to_dict_list(messages: List["ModelMessage"]) -> List[Dict[str, str]]:
        return list(map(lambda m: m.dict(), messages))

    @staticmethod
    def build_human_message(content: str) -> "ModelMessage":
        return ModelMessage(role=ModelMessageRoleType.HUMAN, content=content)


def _message_to_dict(message: BaseMessage) -> Dict:
    return message.to_dict()


def _messages_to_dict(messages: List[BaseMessage]) -> List[Dict]:
    return [_message_to_dict(m) for m in messages]


def _message_from_dict(message: Dict) -> BaseMessage:
    _type = message["type"]
    if _type == "human":
        return HumanMessage(**message["data"])
    elif _type == "ai":
        return AIMessage(**message["data"])
    elif _type == "system":
        return SystemMessage(**message["data"])
    elif _type == "view":
        return ViewMessage(**message["data"])
    else:
        raise ValueError(f"Got unexpected type: {_type}")


def _messages_from_dict(messages: List[Dict]) -> List[BaseMessage]:
    return [_message_from_dict(m) for m in messages]


def parse_model_messages(
    messages: List[ModelMessage],
) -> Tuple[str, List[str], List[List[str, str]]]:
    """
    Parse model messages to extract the user prompt, system messages, and a history of conversation.

    This function analyzes a list of ModelMessage objects, identifying the role of each message (e.g., human, system, ai)
    and categorizes them accordingly. The last message is expected to be from the user (human), and it's treated as
    the current user prompt. System messages are extracted separately, and the conversation history is compiled into
    pairs of human and AI messages.

    Args:
        messages (List[ModelMessage]): List of messages from a chat conversation.

    Returns:
        tuple: A tuple containing the user prompt, list of system messages, and the conversation history.
               The conversation history is a list of message pairs, each containing a user message and the corresponding AI response.

    Examples:
        .. code-block:: python

            # Example 1: Single round of conversation
            messages = [
                ModelMessage(role="human", content="Hello"),
                ModelMessage(role="ai", content="Hi there!"),
                ModelMessage(role="human", content="How are you?"),
            ]
            user_prompt, system_messages, history = parse_model_messages(messages)
            # user_prompt: "How are you?"
            # system_messages: []
            # history: [["Hello", "Hi there!"]]

            # Example 2: Conversation with system messages
            messages = [
                ModelMessage(role="system", content="System initializing..."),
                ModelMessage(role="human", content="Is it sunny today?"),
                ModelMessage(role="ai", content="Yes, it's sunny."),
                ModelMessage(role="human", content="Great!"),
            ]
            user_prompt, system_messages, history = parse_model_messages(messages)
            # user_prompt: "Great!"
            # system_messages: ["System initializing..."]
            # history: [["Is it sunny today?", "Yes, it's sunny."]]

            # Example 3: Multiple rounds with system message
            messages = [
                ModelMessage(role="human", content="Hi"),
                ModelMessage(role="ai", content="Hello!"),
                ModelMessage(role="system", content="Error 404"),
                ModelMessage(role="human", content="What's the error?"),
                ModelMessage(role="ai", content="Just a joke."),
                ModelMessage(role="human", content="Funny!"),
            ]
            user_prompt, system_messages, history = parse_model_messages(messages)
            # user_prompt: "Funny!"
            # system_messages: ["Error 404"]
            # history: [["Hi", "Hello!"], ["What's the error?", "Just a joke."]]
    """

    system_messages: List[str] = []
    history_messages: List[List[str]] = [[]]

    for message in messages[:-1]:
        if message.role == "human":
            history_messages[-1].append(message.content)
        elif message.role == "system":
            system_messages.append(message.content)
        elif message.role == "ai":
            history_messages[-1].append(message.content)
            history_messages.append([])
    if messages[-1].role != "human":
        raise ValueError("Hi! What do you want to talk about？")
    # Keep message a pair of [user message, assistant message]
    history_messages = list(filter(lambda x: len(x) == 2, history_messages))
    user_prompt = messages[-1].content
    return user_prompt, system_messages, history_messages


class OnceConversation:
    """All the information of a conversation, the current single service in memory,
    can expand cache and database support distributed services.

    """

    def __init__(
        self,
        chat_mode: str,
        user_name: str = None,
        sys_code: str = None,
        summary: str = None,
        **kwargs,
    ):
        self.chat_mode: str = chat_mode
        self.user_name: str = user_name
        self.sys_code: str = sys_code
        self.summary: str = summary

        self.messages: List[BaseMessage] = kwargs.get("messages", [])
        self.start_date: str = kwargs.get("start_date", "")
        # After each complete round of dialogue, the current value will be increased by 1
        self.chat_order: int = int(kwargs.get("chat_order", 0))
        self.model_name: str = kwargs.get("model_name", "")
        self.param_type: str = kwargs.get("param_type", "")
        self.param_value: str = kwargs.get("param_value", "")
        self.cost: int = int(kwargs.get("cost", 0))
        self.tokens: int = int(kwargs.get("tokens", 0))
        self._message_index: int = int(kwargs.get("message_index", 0))

    def _append_message(self, message: BaseMessage) -> None:
        index = self._message_index
        self._message_index += 1
        message.index = index
        message.round_index = self.chat_order
        self.messages.append(message)

    def start_new_round(self) -> None:
        """Start a new round of conversation

        Example:
            >>> conversation = OnceConversation()
            >>> # The chat order will be 0, then we start a new round of conversation
            >>> assert conversation.chat_order == 0
            >>> conversation.start_new_round()
            >>> # Now the chat order will be 1
            >>> assert conversation.chat_order == 1
            >>> conversation.add_user_message("hello")
            >>> conversation.add_ai_message("hi")
            >>> conversation.end_current_round()
            >>> # Now the chat order will be 1, then we start a new round of conversation
            >>> conversation.start_new_round()
            >>> # Now the chat order will be 2
            >>> assert conversation.chat_order == 2
            >>> conversation.add_user_message("hello")
            >>> conversation.add_ai_message("hi")
            >>> conversation.end_current_round()
            >>> assert conversation.chat_order == 2
        """
        self.chat_order += 1

    def end_current_round(self) -> None:
        """End the current round of conversation

        We do noting here, just for the interface
        """
        pass

    def add_user_message(
        self, message: str, check_duplicate_type: Optional[bool] = False
    ) -> None:
        """Add a user message to the conversation

        Args:
            message (str): The message content
            check_duplicate_type (bool): Whether to check the duplicate message type

        Raises:
            ValueError: If the message is duplicate and check_duplicate_type is True
        """
        if check_duplicate_type:
            has_message = any(
                isinstance(instance, HumanMessage) for instance in self.messages
            )
            if has_message:
                raise ValueError("Already Have Human message")
        self._append_message(HumanMessage(content=message))

    def add_ai_message(
        self, message: str, update_if_exist: Optional[bool] = False
    ) -> None:
        """Add an AI message to the conversation

        Args:
            message (str): The message content
            update_if_exist (bool): Whether to update the message if the message type is duplicate
        """
        if not update_if_exist:
            self._append_message(AIMessage(content=message))
            return
        has_message = any(isinstance(instance, AIMessage) for instance in self.messages)
        if has_message:
            self._update_ai_message(message)
        else:
            self._append_message(AIMessage(content=message))

    def _update_ai_message(self, new_message: str) -> None:
        """
        stream out message update
        Args:
            new_message:

        Returns:

        """

        for item in self.messages:
            if item.type == "ai":
                item.content = new_message

    def add_view_message(self, message: str) -> None:
        """Add an AI message to the store"""
        self._append_message(ViewMessage(content=message))

    def add_system_message(self, message: str) -> None:
        """Add a system message to the store"""
        self._append_message(SystemMessage(content=message))

    def set_start_time(self, datatime: datetime):
        dt_str = datatime.strftime("%Y-%m-%d %H:%M:%S")
        self.start_date = dt_str

    def clear(self) -> None:
        """Remove all messages from the store"""
        self.messages.clear()

    def get_latest_user_message(self) -> Optional[HumanMessage]:
        """Get the latest user message"""
        for message in self.messages[::-1]:
            if isinstance(message, HumanMessage):
                return message
        return None

    def get_system_messages(self) -> List[SystemMessage]:
        """Get the latest user message"""
        return list(filter(lambda x: isinstance(x, SystemMessage), self.messages))

    def _to_dict(self) -> Dict:
        return _conversation_to_dict(self)

    def from_conversation(self, conversation: OnceConversation) -> None:
        """Load the conversation from the storage"""
        self.chat_mode = conversation.chat_mode
        self.messages = conversation.messages
        self.start_date = conversation.start_date
        self.chat_order = conversation.chat_order
        self.model_name = conversation.model_name
        self.param_type = conversation.param_type
        self.param_value = conversation.param_value
        self.cost = conversation.cost
        self.tokens = conversation.tokens
        self.user_name = conversation.user_name
        self.sys_code = conversation.sys_code

    def get_messages_by_round(self, round_index: int) -> List[BaseMessage]:
        """Get the messages by round index

        Args:
            round_index (int): The round index

        Returns:
            List[BaseMessage]: The messages
        """
        return list(filter(lambda x: x.round_index == round_index, self.messages))

    def get_latest_round(self) -> List[BaseMessage]:
        """Get the latest round messages

        Returns:
            List[BaseMessage]: The messages
        """
        return self.get_messages_by_round(self.chat_order)

    def get_messages_with_round(self, round_count: int) -> List[BaseMessage]:
        """Get the messages with round count

        If the round count is 1, the history messages will not be included.

        Example:
            .. code-block:: python
                conversation = OnceConversation()
                conversation.start_new_round()
                conversation.add_user_message("hello, this is the first round")
                conversation.add_ai_message("hi")
                conversation.end_current_round()
                conversation.start_new_round()
                conversation.add_user_message("hello, this is the second round")
                conversation.add_ai_message("hi")
                conversation.end_current_round()
                conversation.start_new_round()
                conversation.add_user_message("hello, this is the third round")
                conversation.add_ai_message("hi")
                conversation.end_current_round()

                assert len(conversation.get_messages_with_round(1)) == 2
                assert conversation.get_messages_with_round(1)[0].content == "hello, this is the third round"
                assert conversation.get_messages_with_round(1)[1].content == "hi"

                assert len(conversation.get_messages_with_round(2)) == 4
                assert conversation.get_messages_with_round(2)[0].content == "hello, this is the second round"
                assert conversation.get_messages_with_round(2)[1].content == "hi"

        Args:
            round_count (int): The round count

        Returns:
            List[BaseMessage]: The messages
        """
        latest_round_index = self.chat_order
        start_round_index = max(1, latest_round_index - round_count + 1)
        messages = []
        for round_index in range(start_round_index, latest_round_index + 1):
            messages.extend(self.get_messages_by_round(round_index))
        return messages

    def get_model_messages(self) -> List[ModelMessage]:
        """Get the model messages

        Model messages just include human, ai and system messages.
        Model messages maybe include the history messages, The order of the messages is the same as the order of
        the messages in the conversation, the last message is the latest message.

        If you want to hand the message with your own logic, you can override this method.

        Examples:
            If you not need the history messages, you can override this method like this:
            .. code-block:: python
                def get_model_messages(self) -> List[ModelMessage]:
                    messages = []
                    for message in self.get_latest_round():
                        if message.pass_to_model:
                            messages.append(
                                ModelMessage(role=message.type, content=message.content)
                            )
                    return messages

            If you want to add the one round history messages, you can override this method like this:
            .. code-block:: python
                def get_model_messages(self) -> List[ModelMessage]:
                    messages = []
                    latest_round_index = self.chat_order
                    round_count = 1
                    start_round_index = max(1, latest_round_index - round_count + 1)
                    for round_index in range(start_round_index, latest_round_index + 1):
                        for message in self.get_messages_by_round(round_index):
                            if message.pass_to_model:
                                messages.append(
                                    ModelMessage(role=message.type, content=message.content)
                                )
                    return messages

        Returns:
            List[ModelMessage]: The model messages
        """
        messages = []
        for message in self.messages:
            if message.pass_to_model:
                messages.append(
                    ModelMessage(role=message.type, content=message.content)
                )
        return messages


class ConversationIdentifier(ResourceIdentifier):
    """Conversation identifier"""

    def __init__(self, conv_uid: str, identifier_type: str = "conversation"):
        self.conv_uid = conv_uid
        self.identifier_type = identifier_type

    @property
    def str_identifier(self) -> str:
        return f"{self.identifier_type}:{self.conv_uid}"

    def to_dict(self) -> Dict:
        return {"conv_uid": self.conv_uid, "identifier_type": self.identifier_type}


class MessageIdentifier(ResourceIdentifier):
    """Message identifier"""

    identifier_split = "___"

    def __init__(self, conv_uid: str, index: int, identifier_type: str = "message"):
        self.conv_uid = conv_uid
        self.index = index
        self.identifier_type = identifier_type

    @property
    def str_identifier(self) -> str:
        return f"{self.identifier_type}{self.identifier_split}{self.conv_uid}{self.identifier_split}{self.index}"

    @staticmethod
    def from_str_identifier(str_identifier: str) -> MessageIdentifier:
        """Convert from str identifier

        Args:
            str_identifier (str): The str identifier

        Returns:
            MessageIdentifier: The message identifier
        """
        parts = str_identifier.split(MessageIdentifier.identifier_split)
        if len(parts) != 3:
            raise ValueError(f"Invalid str identifier: {str_identifier}")
        return MessageIdentifier(parts[1], int(parts[2]))

    def to_dict(self) -> Dict:
        return {
            "conv_uid": self.conv_uid,
            "index": self.index,
            "identifier_type": self.identifier_type,
        }


class MessageStorageItem(StorageItem):
    @property
    def identifier(self) -> MessageIdentifier:
        return self._id

    def __init__(self, conv_uid: str, index: int, message_detail: Dict):
        self.conv_uid = conv_uid
        self.index = index
        self.message_detail = message_detail
        self._id = MessageIdentifier(conv_uid, index)

    def to_dict(self) -> Dict:
        return {
            "conv_uid": self.conv_uid,
            "index": self.index,
            "message_detail": self.message_detail,
        }

    def to_message(self) -> BaseMessage:
        """Convert to message object
        Returns:
            BaseMessage: The message object

        Raises:
            ValueError: If the message type is not supported
        """
        return _message_from_dict(self.message_detail)

    def merge(self, other: "StorageItem") -> None:
        """Merge the other message to self

        Args:
            other (StorageItem): The other message
        """
        if not isinstance(other, MessageStorageItem):
            raise ValueError(f"Can not merge {other} to {self}")
        self.message_detail = other.message_detail


class StorageConversation(OnceConversation, StorageItem):
    """All the information of a conversation, the current single service in memory,
    can expand cache and database support distributed services.

    """

    @property
    def identifier(self) -> ConversationIdentifier:
        return self._id

    def to_dict(self) -> Dict:
        dict_data = self._to_dict()
        messages: Dict = dict_data.pop("messages")
        message_ids = []
        index = 0
        for message in messages:
            if "index" in message:
                message_idx = message["index"]
            else:
                message_idx = index
                index += 1
            message_ids.append(
                MessageIdentifier(self.conv_uid, message_idx).str_identifier
            )
        # Replace message with message ids
        dict_data["conv_uid"] = self.conv_uid
        dict_data["message_ids"] = message_ids
        dict_data["save_message_independent"] = self.save_message_independent
        return dict_data

    def merge(self, other: "StorageItem") -> None:
        """Merge the other conversation to self

        Args:
            other (StorageItem): The other conversation
        """
        if not isinstance(other, StorageConversation):
            raise ValueError(f"Can not merge {other} to {self}")
        self.from_conversation(other)

    def __init__(
        self,
        conv_uid: str,
        chat_mode: str = None,
        user_name: str = None,
        sys_code: str = None,
        message_ids: List[str] = None,
        summary: str = None,
        save_message_independent: Optional[bool] = True,
        conv_storage: StorageInterface = None,
        message_storage: StorageInterface = None,
        **kwargs,
    ):
        super().__init__(chat_mode, user_name, sys_code, summary, **kwargs)
        self.conv_uid = conv_uid
        self._message_ids = message_ids
        # Record the message index last time saved to the storage,
        # next time save messages which index is _has_stored_message_index + 1
        self._has_stored_message_index = (
            len(kwargs["messages"]) - 1 if "messages" in kwargs else -1
        )
        self.save_message_independent = save_message_independent
        self._id = ConversationIdentifier(conv_uid)
        if conv_storage is None:
            conv_storage = InMemoryStorage()
        if message_storage is None:
            message_storage = InMemoryStorage()
        self.conv_storage = conv_storage
        self.message_storage = message_storage
        # Load from storage
        self.load_from_storage(self.conv_storage, self.message_storage)

    @property
    def message_ids(self) -> List[str]:
        """Get the message ids

        Returns:
            List[str]: The message ids
        """
        return self._message_ids if self._message_ids else []

    def end_current_round(self) -> None:
        """End the current round of conversation

        Save the conversation to the storage after a round of conversation
        """
        self.save_to_storage()

    def _get_message_items(self) -> List[MessageStorageItem]:
        return [
            MessageStorageItem(self.conv_uid, message.index, message.to_dict())
            for message in self.messages
        ]

    def save_to_storage(self) -> None:
        """Save the conversation to the storage"""
        # Save messages first
        message_list = self._get_message_items()
        self._message_ids = [
            message.identifier.str_identifier for message in message_list
        ]
        messages_to_save = message_list[self._has_stored_message_index + 1 :]
        self._has_stored_message_index = len(message_list) - 1
        self.message_storage.save_list(messages_to_save)
        # Save conversation
        self.conv_storage.save_or_update(self)

    def load_from_storage(
        self, conv_storage: StorageInterface, message_storage: StorageInterface
    ) -> None:
        """Load the conversation from the storage

        Warning: This will overwrite the current conversation.

        Args:
            conv_storage (StorageInterface): The storage interface
            message_storage (StorageInterface): The storage interface
        """
        # Load conversation first
        conversation: StorageConversation = conv_storage.load(
            self._id, StorageConversation
        )
        if conversation is None:
            return
        message_ids = conversation._message_ids or []

        # Load messages
        message_list = message_storage.load_list(
            [
                MessageIdentifier.from_str_identifier(message_id)
                for message_id in message_ids
            ],
            MessageStorageItem,
        )
        messages = [message.to_message() for message in message_list]
        conversation.messages = messages
        self._message_ids = message_ids
        self._has_stored_message_index = len(messages) - 1
        self.from_conversation(conversation)


def _conversation_to_dict(once: OnceConversation) -> Dict:
    start_str: str = ""
    if hasattr(once, "start_date") and once.start_date:
        if isinstance(once.start_date, datetime):
            start_str = once.start_date.strftime("%Y-%m-%d %H:%M:%S")
        else:
            start_str = once.start_date

    return {
        "chat_mode": once.chat_mode,
        "model_name": once.model_name,
        "chat_order": once.chat_order,
        "start_date": start_str,
        "cost": once.cost if once.cost else 0,
        "tokens": once.tokens if once.tokens else 0,
        "messages": _messages_to_dict(once.messages),
        "param_type": once.param_type,
        "param_value": once.param_value,
        "user_name": once.user_name,
        "sys_code": once.sys_code,
        "summary": once.summary if once.summary else "",
    }


def _conversations_to_dict(conversations: List[OnceConversation]) -> List[dict]:
    return [_conversation_to_dict(m) for m in conversations]


def _conversation_from_dict(once: dict) -> OnceConversation:
    conversation = OnceConversation(
        once.get("chat_mode"), once.get("user_name"), once.get("sys_code")
    )
    conversation.cost = once.get("cost", 0)
    conversation.chat_mode = once.get("chat_mode", "chat_normal")
    conversation.tokens = once.get("tokens", 0)
    conversation.start_date = once.get("start_date", "")
    conversation.chat_order = int(once.get("chat_order"))
    conversation.param_type = once.get("param_type", "")
    conversation.param_value = once.get("param_value", "")
    conversation.model_name = once.get("model_name", "proxyllm")
    print(once.get("messages"))
    conversation.messages = _messages_from_dict(once.get("messages", []))
    return conversation