dfasd
commited on
Update app.py
Browse files
app.py
CHANGED
@@ -1,118 +1,78 @@
|
|
1 |
from dotenv import load_dotenv
|
2 |
import os
|
3 |
-
|
4 |
-
from langchain_community.vectorstores import Chroma
|
5 |
-
from langchain_text_splitters import CharacterTextSplitter
|
6 |
from langchain_community.document_loaders import PyPDFLoader
|
7 |
-
from
|
8 |
from langchain_openai import OpenAIEmbeddings
|
|
|
|
|
9 |
from langchain_openai import ChatOpenAI
|
10 |
-
from langchain.chains.combine_documents import create_stuff_documents_chain
|
11 |
-
from langchain.chains import create_retrieval_chain
|
12 |
from langchain import hub
|
13 |
-
from langchain_core.
|
14 |
-
from langchain.chains.question_answering import load_qa_chain
|
15 |
-
from langchain.prompts import PromptTemplate
|
16 |
|
17 |
-
|
18 |
load_dotenv()
|
19 |
-
|
20 |
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
|
21 |
|
22 |
-
#
|
23 |
-
|
24 |
-
|
25 |
-
|
|
|
26 |
|
27 |
-
|
|
|
|
|
|
|
28 |
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
db = request.get_json().get("db")
|
34 |
-
|
35 |
-
# if title == "search":
|
36 |
-
# response = tavily.search(query=prompt, include_images=True, include_answer=True, max_results=5)
|
37 |
-
|
38 |
-
# output = response['answer'] + "\n"
|
39 |
-
# for res in response['results']:
|
40 |
-
# output += f"\nTitle: {res['title']}\nURL: {res['url']}\nContent: {res['content']}\n"
|
41 |
-
|
42 |
-
# data = {"success": "ok", "response": output, "images": response['images']}
|
43 |
|
44 |
-
|
|
|
|
|
|
|
45 |
|
46 |
-
|
47 |
-
|
48 |
-
template = """Please answer to human's input based on context. If the input is not mentioned in context, output something like 'I don't know'.
|
49 |
-
Context: {context}
|
50 |
-
Human: {human_input}
|
51 |
-
Your Response as Chatbot:"""
|
52 |
-
|
53 |
-
prompt_s = PromptTemplate(
|
54 |
-
input_variables=["human_input", "context"],
|
55 |
-
template=template
|
56 |
-
)
|
57 |
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
llm = ChatOpenAI(model="gpt-4-1106-preview", api_key=OPENAI_API_KEY)
|
63 |
|
64 |
-
|
65 |
-
|
66 |
|
67 |
-
|
|
|
|
|
|
|
|
|
|
|
68 |
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
data = {"success": "ok", "response": "Please select database."}
|
74 |
|
75 |
-
|
76 |
-
|
77 |
-
def uploadDocuments():
|
78 |
-
# uploaded_files = request.files.getlist('files[]')
|
79 |
-
uploaded_files = ['annualreport2223.pdf', 'Airbus-Annual-Report-2023.pdf']
|
80 |
-
dbname = request.form.get('dbname')
|
81 |
-
if dbname == "":
|
82 |
-
return {"success": "db"}
|
83 |
-
|
84 |
-
if len(uploaded_files) > 0:
|
85 |
-
for file in uploaded_files:
|
86 |
-
file.save(f"uploads/{file.filename}")
|
87 |
-
|
88 |
-
if file.filename.endswith(".txt"):
|
89 |
-
loader = TextLoader(f"uploads/{file.filename}", encoding='utf-8')
|
90 |
-
else:
|
91 |
-
loader = PyPDFLoader(f"uploads/{file.filename}")
|
92 |
-
|
93 |
-
data = loader.load()
|
94 |
-
texts = text_splitter.split_documents(data)
|
95 |
-
|
96 |
-
Chroma.from_documents(texts, embeddings, persist_directory=os.path.join(vectordb_path, dbname))
|
97 |
-
|
98 |
-
return {'success': "ok"}
|
99 |
|
100 |
-
|
101 |
-
|
|
|
|
|
102 |
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
import gradio as gr
|
113 |
-
chatbot = gr.Chatbot(avatar_images=["user.png", "bot.jpg"], height=600)
|
114 |
-
clear_but = gr.Button(value="Clear Chat")
|
115 |
-
demo = gr.ChatInterface(fn="", title="Mediate.com Chatbot Prototype", multimodal=False, retry_btn=None, undo_btn=None, clear_btn=clear_but, chatbot=chatbot)
|
116 |
|
117 |
-
if __name__ ==
|
118 |
-
demo.launch(debug=True)
|
|
|
1 |
from dotenv import load_dotenv
|
2 |
import os
|
3 |
+
import gradio as gr
|
|
|
|
|
4 |
from langchain_community.document_loaders import PyPDFLoader
|
5 |
+
from langchain_text_splitters import CharacterTextSplitter
|
6 |
from langchain_openai import OpenAIEmbeddings
|
7 |
+
from langchain_community.vectorstores import Chroma
|
8 |
+
from langchain_core.runnables import RunnablePassthrough
|
9 |
from langchain_openai import ChatOpenAI
|
|
|
|
|
10 |
from langchain import hub
|
11 |
+
from langchain_core.output_parsers import StrOutputParser
|
|
|
|
|
12 |
|
13 |
+
# Load environment variables
|
14 |
load_dotenv()
|
|
|
15 |
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
|
16 |
|
17 |
+
# Initialize components
|
18 |
+
text_splitter = CharacterTextSplitter(separator="\n", chunk_size=1000, chunk_overlap=200, length_function=len)
|
19 |
+
embeddings = OpenAIEmbeddings(openai_api_key=OPENAI_API_KEY)
|
20 |
+
llm = ChatOpenAI(model="gpt-4-1106-preview", api_key=OPENAI_API_KEY)
|
21 |
+
vectordb_path = './vector_db'
|
22 |
|
23 |
+
# Load and process documents
|
24 |
+
uploaded_files = ['airbus.pdf', 'annualreport2223.pdf']
|
25 |
+
dbname = 'vector_db'
|
26 |
+
vectorstore = None
|
27 |
|
28 |
+
for file in uploaded_files:
|
29 |
+
loader = PyPDFLoader(file)
|
30 |
+
data = loader.load()
|
31 |
+
texts = text_splitter.split_documents(data)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
|
33 |
+
if vectorstore is None:
|
34 |
+
vectorstore = Chroma.from_documents(documents=texts, embedding=embeddings, persist_directory=os.path.join(vectordb_path, dbname))
|
35 |
+
else:
|
36 |
+
vectorstore.add_documents(texts)
|
37 |
|
38 |
+
vectorstore.persist()
|
39 |
+
retriever = vectorstore.as_retriever()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
+
# Load prompt template
|
42 |
+
prompt = hub.pull("rlm/rag-prompt")
|
43 |
+
print(prompt)
|
|
|
|
|
44 |
|
45 |
+
def format_docs(docs):
|
46 |
+
return "\n\n".join(doc.page_content for doc in docs)
|
47 |
|
48 |
+
rag_chain = (
|
49 |
+
{"context": retriever | format_docs, "question": RunnablePassthrough()}
|
50 |
+
| prompt
|
51 |
+
| llm
|
52 |
+
| StrOutputParser()
|
53 |
+
)
|
54 |
|
55 |
+
# Gradio interface
|
56 |
+
def rag_bot(query, chat_history):
|
57 |
+
response = rag_chain.invoke({"input": query, "chat_history": chat_history})
|
58 |
+
return response
|
|
|
59 |
|
60 |
+
chatbot = gr.Chatbot(avatar_images=["user.jpg", "bot.png"], height=600)
|
61 |
+
clear_but = gr.Button(value="Clear Chat")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
62 |
|
63 |
+
def chat(query, chat_history):
|
64 |
+
response = rag_bot(query, chat_history)
|
65 |
+
chat_history.append((query, response))
|
66 |
+
return chat_history, chat_history
|
67 |
|
68 |
+
demo = gr.Interface(
|
69 |
+
fn=chat,
|
70 |
+
inputs=["text", "state"],
|
71 |
+
outputs=["chatbot", "state"],
|
72 |
+
title="RAG Chatbot Prototype",
|
73 |
+
description="A Chatbot using Retrieval-Augmented Generation (RAG) with PDF files.",
|
74 |
+
allow_flagging="never",
|
75 |
+
)
|
|
|
|
|
|
|
|
|
|
|
76 |
|
77 |
+
if __name__ == '__main__':
|
78 |
+
demo.launch(debug=True, share=True)
|