Spaces:
Runtime error
Runtime error
File size: 4,481 Bytes
eda774c 5abdd75 eda774c d0397f4 1f61739 1b2762e eda774c 1b2762e eda774c 164769e 3a1e5b8 acf1331 3a1e5b8 5abdd75 3a1e5b8 8e1c3e9 3a1e5b8 1f61739 8e1c3e9 5abdd75 05bcb80 5abdd75 05bcb80 8e1c3e9 164769e bf3cf08 c386109 c928c71 bf3cf08 5464ccc |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 |
from langchain.chains import RetrievalQA
from langchain.chains import RetrievalQAWithSourcesChain
from langchain.document_loaders import TextLoader
from langchain.docstore.document import Document
import openai
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.llms import OpenAI
import cohere
from langchain.embeddings.cohere import CohereEmbeddings
from langchain.llms import Cohere
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores import Chroma
import os
import pickle
from tqdm import tqdm
import gradio as gr
# from langchain.memory import ConversationSummaryMemory
# from langchain.chains import ConversationalRetrievalChain
# from langchain.text_splitter import RecursiveCharacterTextSplitter
# from langchain.chains import LLMChain
# from langchain.prompts import (
# ChatPromptTemplate,
# HumanMessagePromptTemplate,
# MessagesPlaceholder,
# SystemMessagePromptTemplate,
# )
from langchain.schema import AIMessage,HumanMessage
documents=[]
path='./bios/'
for file in os.listdir(path):
loader = TextLoader(f'{path}{file}',encoding='unicode_escape')
# loader.load()[0].metadata['category']='biography'
# print(loader.load()[0].metadata)
documents += loader.load()
# print(documents)
print(len(documents))
'''This is the code used for without memory chat'''
text_splitter = CharacterTextSplitter(chunk_size=500, chunk_overlap=0)
texts = text_splitter.split_documents(documents)
embeddings = CohereEmbeddings(model='embed-english-v3.0')
docsearch = Chroma.from_documents(texts, embeddings)
qa = RetrievalQA.from_chain_type(llm=Cohere(model='command'), chain_type="stuff", \
retriever=docsearch.as_retriever(search_kwargs={'k':1}),return_source_documents=True)
btuTuples=pickle.load(open('./bookTitleUrlTuples.pkl','rb'))
bookTitleUrlDict={x:y for x,y in btuTuples}
def predict(message, history):
'''experimenation with memory and conversation retrieval chain has resulted in less
performance, usefulness, and more halucination. Hence, this chat bot provides one
shot answers with zero memory. You can use the code in github notebooks to do this
experimentation. github.com/mehrdad-es/Amazon-But-Better'''
# history_langchain_format = []
# for human, ai in history:
# history_langchain_format.append(HumanMessage(content=human))
# history_langchain_format.append(AIMessage(content=ai))
# history_langchain_format.append(HumanMessage(content=message))
# gpt_response = qa({'query':history_langchain_format})
# return gpt_response['result']
# gpt_response = qa({'query':''.join(history)+f'.\n given the previous conversation respond using the following prompt:{message}'})
# # print(gpt_response)
# history.append((f'HumanMessage:{message}',f'AIMessage: {gpt_response},'))
# # history=history_langchain_format
# return gpt_response['result']
message="you are a language model that gives book recommendation based on your context"+message+\
'just give the book title and author'
result = qa({"query": message})
# r1=docsearch.similarity_search_with_score(query=q,k=3)
# print([(item[-2].metadata,item[-1]) for item in r1],\
# '\n\n',result['result'],f'|| {result["source_documents"][0].metadata}','\n*****\n')
if result['result'] not in ["I don't know","I don't know."]:
bookNamePath=result["source_documents"][0].metadata["source"]
return result['result']+f'\n---\n***Ignore the description below if the chatbot was unsure about its response \
or if the response is not about the book shown below***\nAmazon Kindle ebook description is:\n {result["source_documents"][0].page_content}'+\
f'\n**from this file:** {bookNamePath}\n'+\
f'**link==>** {bookTitleUrlDict[bookNamePath.split("/")[-1][:-4]]}'
else:
return result['result']
gr.ChatInterface(predict,
chatbot=gr.Chatbot(height='auto'),
textbox=gr.Textbox(placeholder="Recommend a book on someone who..."),
title="Amazon But Better",
description="Amazon started out with selling books. However, searching books on \
Amazon is tedious and inaccurate if you don't know what you are exactly looking for. **Why not \
make it faster and easier with LLMs:).** This chatbot's context is based on all the non-sponsored \
Kindle ebooks found in the biography section of amazon.ca (1195 items).").launch()
# gr.ChatInterface(predict).launch()
|