Shreyas094
commited on
Commit
•
781b94b
1
Parent(s):
7f18930
Update app.py
Browse files
app.py
CHANGED
@@ -1,102 +1,170 @@
|
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
-
from huggingface_hub import InferenceApi
|
3 |
-
from duckduckgo_search import DDGS
|
4 |
import requests
|
5 |
-
import
|
|
|
|
|
|
|
6 |
from typing import List
|
7 |
from pydantic import BaseModel, Field
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
|
9 |
-
#
|
10 |
huggingface_token = os.environ.get("HUGGINGFACE_TOKEN")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
|
12 |
-
# Function to perform a DuckDuckGo search
|
13 |
def duckduckgo_search(query):
|
14 |
with DDGS() as ddgs:
|
15 |
results = ddgs.text(query, max_results=5)
|
16 |
return results
|
17 |
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
-
def get_response_with_search(query):
|
25 |
-
# Perform the web search
|
26 |
-
search_results = duckduckgo_search(query)
|
27 |
-
|
28 |
-
# Use the search results as context for the model
|
29 |
-
context = "\n".join(f"{result['title']}\n{result['body']}\nSource: {result['href']}\n"
|
30 |
-
for result in search_results if 'body' in result)
|
31 |
-
|
32 |
-
# Prompt formatted for Mistral-7B-Instruct
|
33 |
prompt = f"""<s>[INST] Using the following context:
|
34 |
{context}
|
35 |
Write a detailed and complete research document that fulfills the following user request: '{query}'
|
36 |
After writing the document, please provide a list of sources used in your response. [/INST]"""
|
|
|
|
|
37 |
|
38 |
-
|
39 |
-
API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.3"
|
40 |
|
41 |
-
|
42 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
|
44 |
-
|
45 |
-
|
46 |
-
"
|
47 |
-
"
|
48 |
-
"max_new_tokens": 1000,
|
49 |
-
"temperature": 0.7,
|
50 |
-
"top_p": 0.95,
|
51 |
-
"top_k": 40,
|
52 |
-
"repetition_penalty": 1.1
|
53 |
-
}
|
54 |
-
}
|
55 |
|
56 |
-
|
57 |
-
|
58 |
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
# Split the response into main content and sources
|
70 |
-
parts = generated_text.split("Sources:", 1)
|
71 |
-
main_content = parts[0].strip()
|
72 |
-
sources = parts[1].strip() if len(parts) > 1 else ""
|
73 |
-
|
74 |
-
return main_content, sources
|
75 |
-
else:
|
76 |
-
return f"Unexpected response format: {result}", ""
|
77 |
-
else:
|
78 |
-
return f"Error: API returned status code {response.status_code}", ""
|
79 |
|
80 |
-
def
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
examples=[
|
91 |
-
["What are the latest developments in AI?"],
|
92 |
-
["Tell me about recent updates on GitHub"],
|
93 |
-
["What are the best hotels in Galapagos, Ecuador?"],
|
94 |
-
["Summarize recent advancements in Python programming"],
|
95 |
-
],
|
96 |
-
retry_btn="Retry",
|
97 |
-
undo_btn="Undo",
|
98 |
-
clear_btn="Clear",
|
99 |
-
)
|
100 |
|
101 |
if __name__ == "__main__":
|
102 |
-
|
|
|
1 |
+
import os
|
2 |
+
import json
|
3 |
+
import re
|
4 |
import gradio as gr
|
|
|
|
|
5 |
import requests
|
6 |
+
import random
|
7 |
+
import urllib.parse
|
8 |
+
from tempfile import NamedTemporaryFile
|
9 |
+
from bs4 import BeautifulSoup
|
10 |
from typing import List
|
11 |
from pydantic import BaseModel, Field
|
12 |
+
from huggingface_hub import InferenceApi
|
13 |
+
from duckduckgo_search import DDGS
|
14 |
+
from langchain_community.vectorstores import FAISS
|
15 |
+
from langchain_community.document_loaders import PyPDFLoader
|
16 |
+
from langchain_community.embeddings import HuggingFaceEmbeddings
|
17 |
+
from langchain_community.llms import HuggingFaceHub
|
18 |
+
from langchain_core.documents import Document
|
19 |
+
from sentence_transformers import SentenceTransformer
|
20 |
+
from llama_parse import LlamaParse
|
21 |
|
22 |
+
# Environment variables and configurations
|
23 |
huggingface_token = os.environ.get("HUGGINGFACE_TOKEN")
|
24 |
+
llama_cloud_api_key = os.environ.get("LLAMA_CLOUD_API_KEY")
|
25 |
+
|
26 |
+
# Initialize SentenceTransformer and LlamaParse
|
27 |
+
sentence_model = SentenceTransformer('paraphrase-MiniLM-L6-v2')
|
28 |
+
llama_parser = LlamaParse(
|
29 |
+
api_key=llama_cloud_api_key,
|
30 |
+
result_type="markdown",
|
31 |
+
num_workers=4,
|
32 |
+
verbose=True,
|
33 |
+
language="en",
|
34 |
+
)
|
35 |
+
|
36 |
+
def load_document(file: NamedTemporaryFile, parser: str = "pypdf") -> List[Document]:
|
37 |
+
if parser == "pypdf":
|
38 |
+
loader = PyPDFLoader(file.name)
|
39 |
+
return loader.load_and_split()
|
40 |
+
elif parser == "llamaparse":
|
41 |
+
try:
|
42 |
+
documents = llama_parser.load_data(file.name)
|
43 |
+
return [Document(page_content=doc.text, metadata={"source": file.name}) for doc in documents]
|
44 |
+
except Exception as e:
|
45 |
+
print(f"Error using Llama Parse: {str(e)}")
|
46 |
+
print("Falling back to PyPDF parser")
|
47 |
+
loader = PyPDFLoader(file.name)
|
48 |
+
return loader.load_and_split()
|
49 |
+
else:
|
50 |
+
raise ValueError("Invalid parser specified. Use 'pypdf' or 'llamaparse'.")
|
51 |
+
|
52 |
+
def update_vectors(files, parser):
|
53 |
+
if not files:
|
54 |
+
return "Please upload at least one PDF file."
|
55 |
+
|
56 |
+
embed = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
|
57 |
+
total_chunks = 0
|
58 |
+
|
59 |
+
all_data = []
|
60 |
+
for file in files:
|
61 |
+
data = load_document(file, parser)
|
62 |
+
all_data.extend(data)
|
63 |
+
total_chunks += len(data)
|
64 |
+
|
65 |
+
if os.path.exists("faiss_database"):
|
66 |
+
database = FAISS.load_local("faiss_database", embed, allow_dangerous_deserialization=True)
|
67 |
+
database.add_documents(all_data)
|
68 |
+
else:
|
69 |
+
database = FAISS.from_documents(all_data, embed)
|
70 |
+
|
71 |
+
database.save_local("faiss_database")
|
72 |
+
|
73 |
+
return f"Vector store updated successfully. Processed {total_chunks} chunks from {len(files)} files using {parser}."
|
74 |
+
|
75 |
+
def clear_cache():
|
76 |
+
if os.path.exists("faiss_database"):
|
77 |
+
os.remove("faiss_database")
|
78 |
+
return "Cache cleared successfully."
|
79 |
+
else:
|
80 |
+
return "No cache to clear."
|
81 |
+
|
82 |
+
def get_model(temperature, top_p, repetition_penalty):
|
83 |
+
return HuggingFaceHub(
|
84 |
+
repo_id="mistralai/Mistral-7B-Instruct-v0.3",
|
85 |
+
model_kwargs={
|
86 |
+
"temperature": temperature,
|
87 |
+
"top_p": top_p,
|
88 |
+
"repetition_penalty": repetition_penalty,
|
89 |
+
"max_length": 1000
|
90 |
+
},
|
91 |
+
huggingfacehub_api_token=huggingface_token
|
92 |
+
)
|
93 |
|
|
|
94 |
def duckduckgo_search(query):
|
95 |
with DDGS() as ddgs:
|
96 |
results = ddgs.text(query, max_results=5)
|
97 |
return results
|
98 |
|
99 |
+
def get_response_with_search(query, temperature, top_p, repetition_penalty, use_pdf=False):
|
100 |
+
model = get_model(temperature, top_p, repetition_penalty)
|
101 |
+
embed = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
|
102 |
+
|
103 |
+
if use_pdf and os.path.exists("faiss_database"):
|
104 |
+
database = FAISS.load_local("faiss_database", embed, allow_dangerous_deserialization=True)
|
105 |
+
retriever = database.as_retriever()
|
106 |
+
relevant_docs = retriever.get_relevant_documents(query)
|
107 |
+
context = "\n".join([f"Content: {doc.page_content}\nSource: {doc.metadata['source']}\n" for doc in relevant_docs])
|
108 |
+
else:
|
109 |
+
search_results = duckduckgo_search(query)
|
110 |
+
context = "\n".join(f"{result['title']}\n{result['body']}\nSource: {result['href']}\n"
|
111 |
+
for result in search_results if 'body' in result)
|
112 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
113 |
prompt = f"""<s>[INST] Using the following context:
|
114 |
{context}
|
115 |
Write a detailed and complete research document that fulfills the following user request: '{query}'
|
116 |
After writing the document, please provide a list of sources used in your response. [/INST]"""
|
117 |
+
|
118 |
+
response = model(prompt)
|
119 |
|
120 |
+
main_content, sources = split_response(response)
|
|
|
121 |
|
122 |
+
return main_content, sources
|
123 |
+
|
124 |
+
def split_response(response):
|
125 |
+
parts = response.split("Sources:", 1)
|
126 |
+
main_content = parts[0].strip()
|
127 |
+
sources = parts[1].strip() if len(parts) > 1 else ""
|
128 |
+
return main_content, sources
|
129 |
+
|
130 |
+
def chatbot_interface(message, history, temperature, top_p, repetition_penalty, use_pdf):
|
131 |
+
main_content, sources = get_response_with_search(message, temperature, top_p, repetition_penalty, use_pdf)
|
132 |
+
formatted_response = f"{main_content}\n\nSources:\n{sources}"
|
133 |
+
return formatted_response
|
134 |
+
|
135 |
+
# Gradio interface
|
136 |
+
with gr.Blocks() as demo:
|
137 |
+
gr.Markdown("# AI-powered Web Search and PDF Chat Assistant")
|
138 |
|
139 |
+
with gr.Row():
|
140 |
+
file_input = gr.Files(label="Upload your PDF documents", file_types=[".pdf"])
|
141 |
+
parser_dropdown = gr.Dropdown(choices=["pypdf", "llamaparse"], label="Select PDF Parser", value="pypdf")
|
142 |
+
update_button = gr.Button("Upload PDF")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
143 |
|
144 |
+
update_output = gr.Textbox(label="Update Status")
|
145 |
+
update_button.click(update_vectors, inputs=[file_input, parser_dropdown], outputs=update_output)
|
146 |
|
147 |
+
with gr.Row():
|
148 |
+
with gr.Column(scale=2):
|
149 |
+
chatbot = gr.Chatbot(label="Conversation")
|
150 |
+
msg = gr.Textbox(label="Ask a question")
|
151 |
+
submit_button = gr.Button("Submit")
|
152 |
+
with gr.Column(scale=1):
|
153 |
+
temperature = gr.Slider(label="Temperature", minimum=0.0, maximum=1.0, value=0.7, step=0.1)
|
154 |
+
top_p = gr.Slider(label="Top P", minimum=0.0, maximum=1.0, value=0.95, step=0.05)
|
155 |
+
repetition_penalty = gr.Slider(label="Repetition Penalty", minimum=1.0, maximum=2.0, value=1.1, step=0.1)
|
156 |
+
use_pdf = gr.Checkbox(label="Use PDF Documents", value=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
157 |
|
158 |
+
def respond(message, chat_history, temperature, top_p, repetition_penalty, use_pdf):
|
159 |
+
bot_message = chatbot_interface(message, chat_history, temperature, top_p, repetition_penalty, use_pdf)
|
160 |
+
chat_history.append((message, bot_message))
|
161 |
+
return "", chat_history
|
162 |
|
163 |
+
submit_button.click(respond, inputs=[msg, chatbot, temperature, top_p, repetition_penalty, use_pdf], outputs=[msg, chatbot])
|
164 |
+
|
165 |
+
clear_button = gr.Button("Clear Cache")
|
166 |
+
clear_output = gr.Textbox(label="Cache Status")
|
167 |
+
clear_button.click(clear_cache, inputs=[], outputs=clear_output)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
168 |
|
169 |
if __name__ == "__main__":
|
170 |
+
demo.launch()
|