File size: 5,487 Bytes
173eacf |
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 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 |
"""
creator: Lewis Kamau Kimaru
Function: chat with pdf documents in different languages
best version yet
"""
from langchain.text_splitter import CharacterTextSplitter
from langchain.embeddings import HuggingFaceBgeEmbeddings
from langchain.vectorstores import FAISS
from langchain.chat_models import ChatOpenAI
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationalRetrievalChain
from langchain.llms import HuggingFaceHub
from typing import Union
from dotenv import load_dotenv
from PyPDF2 import PdfReader
import streamlit as st
import requests
import json
import os
# set this key as an environment variable
os.environ["HUGGINGFACEHUB_API_TOKEN"] = st.secrets['huggingface_token']
# Page configuration
st.set_page_config(page_title="SemaNaPDF", page_icon="📚",)
# Sema Translator
Public_Url = 'https://lewiskimaru-helloworld.hf.space' #endpoint
def translate(userinput, target_lang, source_lang=None):
if source_lang:
url = f"{Public_Url}/translate_enter/"
data = {
"userinput": userinput,
"source_lang": source_lang,
"target_lang": target_lang,
}
response = requests.post(url, json=data)
result = response.json()
print(type(result))
source_lange = source_lang
translation = result['translated_text']
else:
url = f"{Public_Url}/translate_detect/"
data = {
"userinput": userinput,
"target_lang": target_lang,
}
response = requests.post(url, json=data)
result = response.json()
source_lange = result['source_language']
translation = result['translated_text']
return source_lange, translation
def get_pdf_text(pdf : Union[str, bytes, bytearray]) -> str:
reader = PdfReader(pdf)
pdf_text = ''
for page in (reader.pages):
text = page.extract_text()
if text:
pdf_text += text
return text
def get_text_chunks(text:str) ->list:
text_splitter = CharacterTextSplitter(
separator="\n", chunk_size=1500, chunk_overlap=300, length_function=len
)
chunks = text_splitter.split_text(text)
return chunks
def get_vectorstore(text_chunks : list) -> FAISS:
model = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
encode_kwargs = {
"normalize_embeddings": True
} # set True to compute cosine similarity
embeddings = HuggingFaceBgeEmbeddings(
model_name=model, encode_kwargs=encode_kwargs, model_kwargs={"device": "cpu"}
)
vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
return vectorstore
def get_conversation_chain(vectorstore:FAISS) -> ConversationalRetrievalChain:
llm = HuggingFaceHub(
repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1",
#repo_id="TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF"
model_kwargs={"temperature": 0.5, "max_length": 1048},
)
memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
conversation_chain = ConversationalRetrievalChain.from_llm(
llm=llm, retriever=vectorstore.as_retriever(), memory=memory
)
return conversation_chain
st.markdown ("""
<style> div.stSpinner > div {
text-align:center;
text-align:center;
align-items: center;
justify-content: center;
}
</style>""", unsafe_allow_html=True)
def main():
st.title("SemaNaPDF📚")
# upload file
pdf = st.file_uploader("Upload a PDF Document", type="pdf")
if pdf is not None:
with st.spinner(""):
# get pdf text
raw_text = get_pdf_text(pdf)
# get the text chunks
text_chunks = get_text_chunks(raw_text)
# create vector store
vectorstore = get_vectorstore(text_chunks)
# create conversation chain
st.session_state.conversation = get_conversation_chain(vectorstore)
st.info("done")
#user_question = st.text_input("chat with your pdf ...")
# show user input
if "messages" not in st.session_state:
st.session_state.messages = []
for message in st.session_state.messages:
with st.chat_message(message["role"]):
st.markdown(message["content"])
if user_question := st.chat_input("Ask your document anything ......?"):
with st.chat_message("user"):
st.markdown(user_question)
user_langd, Queryd = translate(user_question, 'eng_Latn')
st.session_state.messages.append({"role": "user", "content": user_question})
response = st.session_state.conversation({"question": Queryd}) #Queryd
st.session_state.chat_history = response["chat_history"]
output = translate(response['answer'], user_langd, 'eng_Latn')[1] # translated response
with st.chat_message("assistant"):
#st.markdown(response['answer'])
st.markdown(output)
st.session_state.messages.append({"role": "assistant", "content": response['answer']})
# Signature
st.markdown(
"""
<div style="position: fixed; bottom: 0; right: 0; padding: 10px;">
<a href="https://kamaukimaru.vercel.app" target="_blank" style="font-size: 12px; color: #269129; text-decoration: none;">©2023 Lewis Kimaru. All rights reserved.</a>
</div>
""",
unsafe_allow_html=True
)
if __name__ == '__main__':
main()
|