ConversationBufferMemory¶

In [1]:

from langchain.memory import CassandraChatMessageHistory
from langchain.memory import ConversationBufferMemory

from langchain.memory import CassandraChatMessageHistory from langchain.memory import ConversationBufferMemory

In [2]:

from cqlsession import getCQLSession, getCQLKeyspace
astraSession = getCQLSession()
astraKeyspace = getCQLKeyspace()

from cqlsession import getCQLSession, getCQLKeyspace astraSession = getCQLSession() astraKeyspace = getCQLKeyspace()

/home/stefano/.virtualenvs/langchain-cassio-3.10/lib/python3.10/site-packages/cassandra/io/asyncorereactor.py:34: DeprecationWarning: The asyncore module is deprecated and will be removed in Python 3.12. The recommended replacement is asyncio
  import asyncore
/home/stefano/.virtualenvs/langchain-cassio-3.10/lib/python3.10/site-packages/cassandra/datastax/cloud/__init__.py:173: DeprecationWarning: ssl.PROTOCOL_TLS is deprecated
  ssl_context = SSLContext(PROTOCOL_TLS)
/home/stefano/.virtualenvs/langchain-cassio-3.10/lib/python3.10/site-packages/cassandra/io/asyncorereactor.py:347: DeprecationWarning: ssl.match_hostname() is deprecated
  self._connect_socket()

In [3]:

message_history = CassandraChatMessageHistory(
    session_id='conversation-0123',
    session=astraSession,
    keyspace='langchain',
    ttl_seconds = 3600,
)
message_history.clear()

message_history = CassandraChatMessageHistory( session_id='conversation-0123', session=astraSession, keyspace='langchain', ttl_seconds = 3600, ) message_history.clear()

Use in a ConversationChain¶

Create a Memory¶

The Cassandra message history, and the corresponding template variable name for filling prompts, are specified:

In [8]:

cassBuffMemory = ConversationBufferMemory(
    chat_memory=message_history,
)

cassBuffMemory = ConversationBufferMemory( chat_memory=message_history, )

Language model¶

We need an LLM

In [9]:

from langchain.llms import OpenAI
llm = OpenAI()

from langchain.llms import OpenAI llm = OpenAI()

Create a chain¶

In [10]:

from langchain.chains import ConversationChain

conversation = ConversationChain(
    llm=llm, 
    verbose=True, 
    memory=cassBuffMemory,
)

from langchain.chains import ConversationChain conversation = ConversationChain( llm=llm, verbose=True, memory=cassBuffMemory, )

In [12]:

conversation.predict(input="Hello, how can I roast an apple?")

conversation.predict(input="Hello, how can I roast an apple?")

> Entering new ConversationChain chain...
Prompt after formatting:
The following is a friendly conversation between a human and an AI. The AI is talkative and provides lots of specific details from its context. If the AI does not know the answer to a question, it truthfully says it does not know.

Current conversation:

Human: Hello, how can I roast an apple?
AI:

> Finished chain.

Out[12]:

" Hi there! To roast an apple, you'll need to preheat your oven to 375 degrees Fahrenheit. Cut the apple into slices or wedges and spread them out on a baking sheet. Sprinkle with cinnamon and brown sugar. Bake for 30 minutes, or until the apple slices are tender. Enjoy your roasted apples!"

In [13]:

conversation.predict(input="Can I do it on a bonfire?")

conversation.predict(input="Can I do it on a bonfire?")

> Entering new ConversationChain chain...
Prompt after formatting:
The following is a friendly conversation between a human and an AI. The AI is talkative and provides lots of specific details from its context. If the AI does not know the answer to a question, it truthfully says it does not know.

Current conversation:
Human: Hello, how can I roast an apple?
AI:  Hi there! To roast an apple, you'll need to preheat your oven to 375 degrees Fahrenheit. Cut the apple into slices or wedges and spread them out on a baking sheet. Sprinkle with cinnamon and brown sugar. Bake for 30 minutes, or until the apple slices are tender. Enjoy your roasted apples!
Human: Can I do it on a bonfire?
AI:

> Finished chain.

Out[13]:

" I'm sorry, I don't know the answer to that question. Roasting an apple requires an oven."

In [15]:

conversation.predict(input="What about a microwave, would the apple taste good?")

conversation.predict(input="What about a microwave, would the apple taste good?")

> Entering new ConversationChain chain...
Prompt after formatting:
The following is a friendly conversation between a human and an AI. The AI is talkative and provides lots of specific details from its context. If the AI does not know the answer to a question, it truthfully says it does not know.

Current conversation:
Human: Hello, how can I roast an apple?
AI:  Hi there! To roast an apple, you'll need to preheat your oven to 375 degrees Fahrenheit. Cut the apple into slices or wedges and spread them out on a baking sheet. Sprinkle with cinnamon and brown sugar. Bake for 30 minutes, or until the apple slices are tender. Enjoy your roasted apples!
Human: Can I do it on a bonfire?
AI:  I'm sorry, I don't know the answer to that question. Roasting an apple requires an oven.
Human: What about a microwave, would the apple taste good?
AI:

> Finished chain.

Out[15]:

' No, microwaving an apple is not recommended. The result would be a soggy apple with no texture. Roasting is the best way to get a delicious, caramelized apple.'

In [16]:

message_history.messages

message_history.messages

Out[16]:

[HumanMessage(content='Hello, how can I roast an apple?', additional_kwargs={}, example=False),
 AIMessage(content=" Hi there! To roast an apple, you'll need to preheat your oven to 375 degrees Fahrenheit. Cut the apple into slices or wedges and spread them out on a baking sheet. Sprinkle with cinnamon and brown sugar. Bake for 30 minutes, or until the apple slices are tender. Enjoy your roasted apples!", additional_kwargs={}, example=False),
 HumanMessage(content='Can I do it on a bonfire?', additional_kwargs={}, example=False),
 AIMessage(content=" I'm sorry, I don't know the answer to that question. Roasting an apple requires an oven.", additional_kwargs={}, example=False),
 HumanMessage(content='What about a microwave, would the apple taste good?', additional_kwargs={}, example=False),
 AIMessage(content=' No, microwaving an apple is not recommended. The result would be a soggy apple with no texture. Roasting is the best way to get a delicious, caramelized apple.', additional_kwargs={}, example=False)]

Manually tinkering with the prompt¶

In [17]:

from langchain import LLMChain, PromptTemplate

from langchain import LLMChain, PromptTemplate

In [18]:

template = """You are a quirky chatbot having a
conversation with a human, riddled with puns and silly jokes.

{chat_history}
Human: {human_input}
Chatbot:"""

prompt = PromptTemplate(
    input_variables=["chat_history", "human_input"], 
    template=template
)

template = """You are a quirky chatbot having a conversation with a human, riddled with puns and silly jokes. {chat_history} Human: {human_input} Chatbot:""" prompt = PromptTemplate( input_variables=["chat_history", "human_input"], template=template )

In [20]:

f_message_history = CassandraChatMessageHistory(
    session_id='conversation-funny-a001',
    session=astraSession,
    keyspace='langchain',
)
f_message_history.clear()

f_message_history = CassandraChatMessageHistory( session_id='conversation-funny-a001', session=astraSession, keyspace='langchain', ) f_message_history.clear()

In [21]:

f_memory = ConversationBufferMemory(
    memory_key="chat_history",
    chat_memory=f_message_history,
)

f_memory = ConversationBufferMemory( memory_key="chat_history", chat_memory=f_message_history, )

In [22]:

llm_chain = LLMChain(
    llm=llm, 
    prompt=prompt, 
    verbose=True, 
    memory=f_memory,
)

llm_chain = LLMChain( llm=llm, prompt=prompt, verbose=True, memory=f_memory, )

In [23]:

llm_chain.predict(human_input="Tell me about springs")

llm_chain.predict(human_input="Tell me about springs")

> Entering new LLMChain chain...
Prompt after formatting:
You are a quirky chatbot having a
conversation with a human, riddled with puns and silly jokes.


Human: Tell me about springs
Chatbot:

> Finished chain.

Out[23]:

' Springs are full of surprises! They bounce back with lots of energy and are always ready to take on a new challenge. Plus, they provide a great way to relax and have fun!'

In [24]:

llm_chain.predict(human_input='Er ... I mean the other type actually.')

llm_chain.predict(human_input='Er ... I mean the other type actually.')

> Entering new LLMChain chain...
Prompt after formatting:
You are a quirky chatbot having a
conversation with a human, riddled with puns and silly jokes.

Human: Tell me about springs
AI:  Springs are full of surprises! They bounce back with lots of energy and are always ready to take on a new challenge. Plus, they provide a great way to relax and have fun!
Human: Er ... I mean the other type actually.
Chatbot:

> Finished chain.

Out[24]:

" Oh, you mean the season? Well, spring is a season of renewal and growth. Everything is coming back to life after a cold winter and the world starts to bloom and blossom. It's definitely a time of joy and celebration!"