-
Notifications
You must be signed in to change notification settings - Fork 0
/
app.py
147 lines (98 loc) · 3.83 KB
/
app.py
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
import json
import os
import sys
import boto3
import streamlit as st
## We will be suing Titan Embeddings Model To generate Embedding
from langchain_community.embeddings import BedrockEmbeddings
from langchain.llms.bedrock import Bedrock
## Data Ingestion
import numpy as np
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.document_loaders import PyPDFDirectoryLoader
# Vector Embedding And Vector Store
from langchain.vectorstores import FAISS
## LLm Models
from langchain.prompts import PromptTemplate
from langchain.chains import RetrievalQA
## Bedrock Clients
bedrock=boto3.client(service_name="bedrock-runtime")
bedrock_embeddings=BedrockEmbeddings(model_id="amazon.titan-embed-text-v1",client=bedrock)
## Data ingestion
def data_ingestion():
loader=PyPDFDirectoryLoader("data")
documents=loader.load()
# - in our testing Character split works better with this PDF data set
text_splitter=RecursiveCharacterTextSplitter(chunk_size=10000,
chunk_overlap=1000)
docs=text_splitter.split_documents(documents)
return docs
## Vector Embedding and vector store
def get_vector_store(docs):
vectorstore_faiss=FAISS.from_documents(
docs,
bedrock_embeddings
)
vectorstore_faiss.save_local("faiss_index")
def get_claude_llm():
##create the Anthropic Model
llm=Bedrock(model_id="ai21.j2-mid-v1",client=bedrock,
model_kwargs={'maxTokens':512})
return llm
def get_llama2_llm():
##create the Anthropic Model
llm=Bedrock(model_id="meta.llama2-70b-chat-v1",client=bedrock,
model_kwargs={'max_gen_len':512})
return llm
prompt_template = """
Human: Use the following pieces of context to provide a
concise answer to the question at the end but usse atleast summarize with
250 words with detailed explaantions. If you don't know the answer,
just say that you don't know, don't try to make up an answer.
<context>
{context}
</context
Question: {question}
Assistant:"""
PROMPT = PromptTemplate(
template=prompt_template, input_variables=["context", "question"]
)
def get_response_llm(llm,vectorstore_faiss,query):
qa = RetrievalQA.from_chain_type(
llm=llm,
chain_type="stuff",
retriever=vectorstore_faiss.as_retriever(
search_type="similarity", search_kwargs={"k": 3}
),
return_source_documents=True,
chain_type_kwargs={"prompt": PROMPT}
)
answer=qa({"query":query})
return answer['result']
def main():
st.set_page_config("Chat PDF")
st.header("Chat with PDF using AWS Bedrock💁")
user_question = st.text_input("Ask a Question from the PDF Files")
with st.sidebar:
st.title("Update Or Create Vector Store:")
if st.button("Vectors Update"):
with st.spinner("Processing..."):
docs = data_ingestion()
get_vector_store(docs)
st.success("Done")
if st.button("Claude Output"):
with st.spinner("Processing..."):
faiss_index = FAISS.load_local("faiss_index", bedrock_embeddings)
llm=get_claude_llm()
#faiss_index = get_vector_store(docs)
st.write(get_response_llm(llm,faiss_index,user_question))
st.success("Done")
if st.button("Llama2 Output"):
with st.spinner("Processing..."):
faiss_index = FAISS.load_local("faiss_index", bedrock_embeddings)
llm=get_llama2_llm()
#faiss_index = get_vector_store(docs)
st.write(get_response_llm(llm,faiss_index,user_question))
st.success("Done")
if __name__ == "__main__":
main()