week2assignment_shakes_final.py

# -*- coding: utf-8 -*-
"""week2assignment_shakes.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1OfNmkHMwkuJONUG4yQHDwJiuzLJvF1kJ
"""

!pip install openai langchain python-dotenv -q

!echo openai_api_key="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" > .env

import os
import openai
from dotenv import load_dotenv

load_dotenv(".env")

openai.api_key = os.environ.get("openai_api_key")

from IPython.display import display, Markdown

def disp_markdown(text: str) -> None:
  display(Markdown(text))

from langchain.chat_models import ChatOpenAI
from langchain.schema import HumanMessage

chat_model = ChatOpenAI(model_name="gpt-3.5-turbo", openai_api_key=os.environ.get("openai_api_key"))

from langchain.schema import (
    AIMessage,
    HumanMessage,
    SystemMessage
)

# The SystemMessage is associated with the system role
system_message = SystemMessage(content="You are a food critic.")

# The HumanMessage is associated with the user role
user_message = HumanMessage(content="Do you think Kraft Dinner constitues fine dining?")

# The AIMessage is associated with the assistant role
assistant_message = AIMessage(content="Egads! No, it most certainly does not!")

second_user_message = HumanMessage(content="What about Red Lobster, surely that is fine dining!")

# create the list of prompts
list_of_prompts = [
    system_message,
    user_message,
    assistant_message,
    second_user_message
]

# we can just call our chat_model on the list of prompts!
chat_model(list_of_prompts)

from langchain.prompts.chat import (
    ChatPromptTemplate,
    SystemMessagePromptTemplate,
    HumanMessagePromptTemplate
)

# we can signify variables we want access to by wrapping them in {}
system_prompt_template = "You are an expert in {SUBJECT}, and you're currently feeling {MOOD}"
system_prompt_template = SystemMessagePromptTemplate.from_template(system_prompt_template)

user_prompt_template = "{CONTENT}"
user_prompt_template = HumanMessagePromptTemplate.from_template(user_prompt_template)

# put them together into a ChatPromptTemplate
chat_prompt = ChatPromptTemplate.from_messages([system_prompt_template, user_prompt_template])

formatted_chat_prompt = chat_prompt.format_prompt(SUBJECT="cheeses", MOOD="quite tired", CONTENT="Hi, what are the finest cheeses?").to_messages()

disp_markdown(chat_model(formatted_chat_prompt).content)

from langchain.chains import LLMChain

chain = LLMChain(llm=chat_model, prompt=chat_prompt)

disp_markdown(chain.run(SUBJECT="classic cars", MOOD="angry", CONTENT="Is the 67 Chevrolet Impala a good vehicle?"))

!wget https://erki.lap.ee/failid/raamatud/guide1.txt

with open("guide1.txt") as f:
    hitchhikersguide = f.read()

from langchain.text_splitter import CharacterTextSplitter

text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0, separator = "\n")
texts = text_splitter.split_text(hitchhikersguide)

from langchain.embeddings.openai import OpenAIEmbeddings

os.environ["OPENAI_API_KEY"] = openai.api_key

embeddings = OpenAIEmbeddings()

# !pip install chromadb==0.3.22 tiktoken -q

# !pip install chromadb -U

!pip install pydantic -q
import chromadb

from langchain.vectorstores.chroma import Chroma
docsearch = Chroma.from_texts(texts, embeddings, metadatas=[{"source": str(i)} for i in range(len(texts))]).as_retriever()

query = "What makes towels important?"
docs = docsearch.get_relevant_documents(query)

docs[0]

from langchain.chains.question_answering import load_qa_chain
from langchain.llms import OpenAI

chain = load_qa_chain(OpenAI(temperature=0), chain_type="stuff")
query = "What makes towels important?"
chain.run(input_documents=docs, question=query)

"""# Assignment 2

"""

!git clone https://github.com/TheMITTech/shakespeare

from glob import glob

files = glob("./shakespeare/**/*.html")

import shutil
import os

os.mkdir('./data')
destination_folder = './data/'

for html_file in files:
  shutil.move(html_file, destination_folder + html_file.split("/")[-1])

!pip install beautifulsoup4 -q

from langchain.document_loaders import BSHTMLLoader, DirectoryLoader

bshtml_dir_loader = DirectoryLoader('./data/', loader_cls=BSHTMLLoader)

data = bshtml_dir_loader.load()

from langchain.text_splitter import RecursiveCharacterTextSplitter

text_splitter = RecursiveCharacterTextSplitter(
    chunk_size = 1000,
    chunk_overlap  = 20,
    length_function = len,
)

documents = text_splitter.split_documents(data)

persist_directory = "vector_db"

vectordb = Chroma.from_documents(documents=documents, embedding=embeddings, persist_directory=persist_directory)

vectordb.persist()
vectordb = None

vectordb = Chroma(persist_directory=persist_directory, embedding_function=embeddings)

llm = ChatOpenAI(temperature=0, model="gpt-4")

doc_retriever = vectordb.as_retriever()

from langchain.chains import RetrievalQA

shakespeare_qa = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=doc_retriever)

shakespeare_qa.run("Who is Hamlet'?")

!pip install google-search-results -q

os.environ["SERPAPI_API_KEY"] = "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"

from langchain.utilities import SerpAPIWrapper

search = SerpAPIWrapper()

from langchain.agents import initialize_agent, Tool
from langchain.agents import AgentType
from langchain.tools import BaseTool
from langchain.llms import OpenAI
from langchain import LLMMathChain, SerpAPIWrapper

tools = [
    Tool(
        name = "Shakespeare QA System",
        func=shakespeare_qa.run,
        description="useful for when you need to answer questions about Shakespeare's works. Input should be a fully formed question."
    ),
    Tool(
        name = "SERP API Search",
        func=search.run,
        description="useful for when you need to answer questions about ruff (a python linter). Input should be a fully formed question."
    ),
]

agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)

agent.run("What is Hamlet and more importantly who is hamlet?")

from langchain.memory import ConversationBufferMemory, ReadOnlySharedMemory

memory = ConversationBufferMemory(memory_key="chat_history")
readonlymemory = ReadOnlySharedMemory(memory=memory)

shakespeare_qa = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=doc_retriever, memory=readonlymemory)

tools = [
    Tool(
        name = "Shakespeare QA System",
        func=shakespeare_qa.run,
        description="useful for when you need to answer questions about Shakespeare's works. Input should be a fully formed question."
    ),
    Tool(
        name = "SERP API Search",
        func=search.run,
        description="useful for when you need to answer questions about ruff (a python linter). Input should be a fully formed question."
    ),
]

from langchain.agents import ZeroShotAgent, Tool, AgentExecutor

prefix = """Have a conversation with a human, answering the following questions as best you can. You have access to the following tools:"""
suffix = """Begin!"

{chat_history}
Question: {input}
{agent_scratchpad}"""

prompt = ZeroShotAgent.create_prompt(
    tools,
    prefix=prefix,
    suffix=suffix,
    input_variables=["input", "chat_history", "agent_scratchpad"]
)

from langchain import OpenAI, LLMChain, PromptTemplate

llm_chain = LLMChain(llm=llm, prompt=prompt)

agent = ZeroShotAgent(llm_chain=llm_chain, tools=tools, verbose=True)
agent_chain = AgentExecutor.from_agent_and_tools(agent=agent, tools=tools, verbose=True, memory=memory)

agent_chain.run(input="Who is Hamlet and What is Hamlet?")

agent_chain.run(input="What age was he in the play?")

agent_chain.run(input="Did he live through the play?")

agent_chain.run(input="What age did you think he was if you approximate without directly reading it from the play? You make the inference on his acts and ages of people in his life")

!pip install gradio

import gradio as gr

def the_app(text):
  return agent_chain.run(input=text)

# x=the_app('who is gertrude?')



# iface = gr.Interface(fn= the_app, inputs= "text", outputs="text",title="Shakespearean")

# iface.launch()