How can I save my onw chat history in my TXT flie?

[jeremygan2021]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

# Lib\site-packages\langchain\memory
    def save_context(self, inputs: Dict[str, Any], outputs: Dict[str, str]) -> None:
        """Save context from this conversation to buffer."""
        input_str, output_str = self._get_input_output(inputs, outputs)
        self.chat_memory.add_messages(
            [HumanMessage(content=input_str), AIMessage(content=output_str)]
        )
        A = f"Input: {input_str}"
        B = f"Output:{output_str}"
        print(A)
        print(B)
        return A, B

#my code

	def chat_by_purpose_ET_memory_retrieval(self, who, content, Entity, source="wechat", signature=None):

		self.Echat = ConversationChain(
			llm=self.llm_zhipu_new,
			verbose=True,
			prompt=Wechat(var=content, entity=str(Entity)).Q_BAKE_TEMPLATE_END2END,
			memory=ConversationEntityMemory(llm=self.llm_zhipu_air)
		)

		start_word = self.Q_bake_start.invoke({"name": who,"signature" : signature}).content
		self.wx.SendMsg_hotkey(start_word, who)
		
		while True:
			try:

				friend_name, receive_msg = self.wx.GetAllMessage[-1][0], self.wx.GetAllMessage[-1][1]  
				if (friend_name == who) and (receive_msg != temp_msg):
					print(f'【{who}】send：【{receive_msg}】')
					temp_msg = receive_msg
					# Entity 
					msg = self.Echat.predict(input=receive_msg)
					if "Target_Down" in msg:
						print("finish")
						reply_msg = thanks you"
						self.wx.SendMsg_hotkey(reply_msg, who)
						return temp_msg, last_msg, content
						break
					else:
						reply_msg = msg
						pprint(f"reply_msg:【{reply_msg}】")
						last_msg = reply_msg
						self.wx.SendMsg_hotkey(reply_msg, who)
						store_input = self.Echat.memory.entity_store.store
						first_save = Memory_save(name=who, memory_entity=store_input, source=source)
						print(first_save)

						# Entity 
						self.Echat.predict(response=reply_msg)
						store_output = self.Echat.memory.entity_store.store
						second_save = Memory_save(name=who, memory_entity=store_output, source=source)
						print(second_save)

			except:
				pass

Description

How can I return the A and B from save_context in my code? so I can save this data in my way?

System Info

Entering new ConversationChain chain...
Prompt after formatting:

You are helpful assistant

Context:
{}

Current conversation:

Last line:
human: test
AI：
Input: test
Output:How can I help you?

[dosubot[bot]]

To return the variables A and B from the save_context method so you can save this data in your own way, you can modify the method to return these variables. Here is how you can do it:

# Lib\site-packages\langchain\memory
def save_context(self, inputs: Dict[str, Any], outputs: Dict[str, str]) -> Dict[str, Any]:
    """Save context from this conversation to buffer and return variables A and B."""
    input_str, output_str = self._get_input_output(inputs, outputs)
    self.chat_memory.add_messages(
        [HumanMessage(content=input_str), AIMessage(content=output_str)]
    )
    A = f"Input: {input_str}"
    B = f"Output: {output_str}"
    print(A)
    print(B)
    return {'A': A, 'B': B}

This modification ensures that the save_context method now returns a dictionary containing the variables A and B. You can then use this returned data in your own code to save it in a TXT file or any other format you prefer.

Here is an example of how you might integrate this into your existing code:

# my code

def chat_by_purpose_ET_memory_retrieval(self, who, content, Entity, source="wechat", signature=None):
    self.Echat = ConversationChain(
        llm=self.llm_zhipu_new,
        verbose=True,
        prompt=Wechat(var=content, entity=str(Entity)).Q_BAKE_TEMPLATE_END2END,
        memory=ConversationEntityMemory(llm=self.llm_zhipu_air)
    )

    start_word = self.Q_bake_start.invoke({"name": who, "signature": signature}).content
    self.wx.SendMsg_hotkey(start_word, who)
    
    while True:
        try:
            friend_name, receive_msg = self.wx.GetAllMessage[-1][0], self.wx.GetAllMessage[-1][1]
            if (friend_name == who) and (receive_msg != temp_msg):
                print(f'【{who}】send：【{receive_msg}】')
                temp_msg = receive_msg
                # Entity 
                msg = self.Echat.predict(input=receive_msg)
                if "Target_Down" in msg:
                    print("finish")
                    reply_msg = "thank you"
                    self.wx.SendMsg_hotkey(reply_msg, who)
                    return temp_msg, last_msg, content
                    break
                else:
                    reply_msg = msg
                    pprint(f"reply_msg:【{reply_msg}】")
                    last_msg = reply_msg
                    self.wx.SendMsg_hotkey(reply_msg, who)
                    store_input = self.Echat.memory.entity_store.store
                    first_save = Memory_save(name=who, memory_entity=store_input, source=source)
                    print(first_save)

                    # Save context and get A and B
                    context_data = self.Echat.memory.save_context(inputs={'input': receive_msg}, outputs={'response': reply_msg})
                    A = context_data['A']
                    B = context_data['B']
                    
                    # Save A and B to a TXT file
                    with open('chat_history.txt', 'a') as file:
                        file.write(f"{A}\n{B}\n")

                    # Entity 
                    self.Echat.predict(response=reply_msg)
                    store_output = self.Echat.memory.entity_store.store
                    second_save = Memory_save(name=who, memory_entity=store_output, source=source)
                    print(second_save)

        except Exception as e:
            print(f"An error occurred: {e}")
            pass

In this example, the save_context method is called, and the returned dictionary containing A and B is used to save the chat history to a TXT file. This way, you can store the conversation history in your preferred format ^[1].

To continue talking to Dosu, mention @dosu.

[jeremygan2021]

I try this code, however It dose not work for returning the variables A and B. I should I do?

[jeremygan2021]

Context:
{}

Current conversation:

Last line:
human: test

Input: test
Output: waht can I help you?

Finished chain.

Input: test
Output: waht can I help you?

It dose not return the variables in my code!