Spaces:
Runtime error
Runtime error
JahanavDixit
commited on
Commit
β’
62174f8
1
Parent(s):
80f1af7
Upload 5 files
Browse files- 48lawsofpower.pdf +0 -0
- chainlit.md +14 -0
- dockerfile.txt +11 -0
- pdf_qa.py +98 -0
- requirements.txt +8 -0
48lawsofpower.pdf
ADDED
Binary file (105 kB). View file
|
|
chainlit.md
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Welcome to Chainlit! ππ€
|
2 |
+
|
3 |
+
Hi there, Developer! π We're excited to have you on board. Chainlit is a powerful tool designed to help you prototype, debug and share applications built on top of LLMs.
|
4 |
+
|
5 |
+
## Useful Links π
|
6 |
+
|
7 |
+
- **Documentation:** Get started with our comprehensive [Chainlit Documentation](https://docs.chainlit.io) π
|
8 |
+
- **Discord Community:** Join our friendly [Chainlit Discord](https://discord.gg/k73SQ3FyUh) to ask questions, share your projects, and connect with other developers! π¬
|
9 |
+
|
10 |
+
We can't wait to see what you create with Chainlit! Happy coding! π»π
|
11 |
+
|
12 |
+
## Welcome screen
|
13 |
+
|
14 |
+
To modify the welcome screen, edit the `chainlit.md` file at the root of your project. If you do not want a welcome screen, just leave this file empty.
|
dockerfile.txt
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
FROM python:3.9
|
2 |
+
|
3 |
+
WORKDIR /code
|
4 |
+
|
5 |
+
COPY ./requirements.txt /code/requirements.txt
|
6 |
+
|
7 |
+
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
|
8 |
+
|
9 |
+
COPY ./* /code/
|
10 |
+
|
11 |
+
CMD ["chainlit", "run", "app.py", "--port", "7860"]
|
pdf_qa.py
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Import necessary modules and define env variables
|
2 |
+
|
3 |
+
from langchain.chains import RetrievalQA
|
4 |
+
from langchain_community.document_loaders import PyPDFLoader
|
5 |
+
from langchain_community.vectorstores import FAISS
|
6 |
+
from langchain.prompts.chat import (
|
7 |
+
ChatPromptTemplate
|
8 |
+
)
|
9 |
+
from langchain_community.llms import HuggingFaceHub
|
10 |
+
import tempfile
|
11 |
+
from langchain_community.embeddings import HuggingFaceEmbeddings
|
12 |
+
import os
|
13 |
+
import io
|
14 |
+
import chainlit as cl
|
15 |
+
import PyPDF2
|
16 |
+
|
17 |
+
#os.environ["HUGGINGFACEHUB_API_TOKEN"] = ""
|
18 |
+
template = """Answer the question based only on the following context from the book 48 Laws of Power:
|
19 |
+
{context}
|
20 |
+
|
21 |
+
Question: {question}
|
22 |
+
"""
|
23 |
+
prompt = ChatPromptTemplate.from_template(template)
|
24 |
+
chain_type_kwargs = {"prompt": prompt}
|
25 |
+
|
26 |
+
from langchain.text_splitter import SpacyTextSplitter
|
27 |
+
|
28 |
+
text_splitter = SpacyTextSplitter(chunk_size=1000)
|
29 |
+
|
30 |
+
@cl.on_chat_start
|
31 |
+
async def on_chat_start():
|
32 |
+
await cl.Message(content="Hello there, Welcome to Laws of Power chat app!").send()
|
33 |
+
msg = cl.Message(content=f"Processing Laws of Power...")
|
34 |
+
await msg.send()
|
35 |
+
loader = PyPDFLoader('./48lawsofpower.pdf')
|
36 |
+
pages = loader.load_and_split()
|
37 |
+
|
38 |
+
# Create a Chroma vector store
|
39 |
+
embeddings = HuggingFaceEmbeddings()
|
40 |
+
faiss_index = FAISS.from_documents(pages, embeddings)
|
41 |
+
|
42 |
+
# Clean up the temporary file
|
43 |
+
pdf = PyPDF2.PdfReader('./48lawsofpower.pdf')
|
44 |
+
pdf_text = ""
|
45 |
+
for page in pdf.pages:
|
46 |
+
pdf_text += page.extract_text()
|
47 |
+
|
48 |
+
# Split the text into chunks
|
49 |
+
texts = text_splitter.split_text(pdf_text)
|
50 |
+
|
51 |
+
# Create metadata for each chunk
|
52 |
+
metadatas = [{"source": f"{i}-pl"} for i in range(len(texts))]
|
53 |
+
|
54 |
+
repo_id = "HuggingFaceH4/zephyr-7b-beta"
|
55 |
+
|
56 |
+
chain_type_kwargs = {"prompt": prompt}
|
57 |
+
|
58 |
+
llm = HuggingFaceHub(
|
59 |
+
repo_id=repo_id, model_kwargs={"temperature": 0.1, "max_new_tokens":1024, "max_length": 728}
|
60 |
+
)
|
61 |
+
|
62 |
+
# Create a chain that uses the Chroma vector store
|
63 |
+
chain = RetrievalQA.from_chain_type(
|
64 |
+
llm,
|
65 |
+
chain_type="stuff",
|
66 |
+
retriever=faiss_index.as_retriever(),
|
67 |
+
)
|
68 |
+
|
69 |
+
|
70 |
+
# Save the metadata and texts in the user session
|
71 |
+
cl.user_session.set("metadatas", metadatas)
|
72 |
+
cl.user_session.set("texts", texts)
|
73 |
+
|
74 |
+
# Let the user know that the system is ready
|
75 |
+
msg.content = f"Processing Laws of Power done. You can now ask questions!"
|
76 |
+
await msg.update()
|
77 |
+
|
78 |
+
cl.user_session.set("chain", chain)
|
79 |
+
|
80 |
+
|
81 |
+
@cl.on_message
|
82 |
+
async def main(message:str):
|
83 |
+
message = message.content
|
84 |
+
print("This" , message)
|
85 |
+
chain = cl.user_session.get("chain")
|
86 |
+
cb = cl.AsyncLangchainCallbackHandler(
|
87 |
+
stream_final_answer=True, answer_prefix_tokens=["FINAL", "ANSWER"]
|
88 |
+
)
|
89 |
+
cb.answer_reached = True
|
90 |
+
res = await chain.acall(message, callbacks=[cb])
|
91 |
+
|
92 |
+
answer = res['result']
|
93 |
+
source_elements = []
|
94 |
+
if cb.has_streamed_final_answer:
|
95 |
+
cb.final_stream.elements = source_elements
|
96 |
+
await cb.final_stream.update()
|
97 |
+
else:
|
98 |
+
await cl.Message(content=answer, elements=source_elements).send()
|
requirements.txt
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
langchain
|
2 |
+
chainlit
|
3 |
+
transformers
|
4 |
+
huggingface_hub
|
5 |
+
faiss_cpu
|
6 |
+
tiktoken
|
7 |
+
spacy
|
8 |
+
PyPDF2
|