Spaces:
Sleeping
Sleeping
MuntasirHossain
commited on
Commit
•
43ac004
1
Parent(s):
5a6dca8
Update app.py
Browse files
app.py
CHANGED
@@ -1,10 +1,215 @@
|
|
1 |
-
import os
|
2 |
import gradio as gr
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
|
4 |
-
|
5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
|
7 |
-
# with gr.Blocks() as app:
|
8 |
-
# gr.load("MuntasirHossain/RAG-PDF-Chatbot", hf_token=os.getenv("HF_TOKEN"), src="spaces").render()
|
9 |
|
10 |
-
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
+
import os
|
3 |
+
api_token = os.getenv("HF_TOKEN")
|
4 |
+
|
5 |
+
|
6 |
+
from langchain_community.vectorstores import FAISS
|
7 |
+
from langchain_community.document_loaders import PyPDFLoader
|
8 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
9 |
+
from langchain_community.vectorstores import Chroma
|
10 |
+
from langchain.chains import ConversationalRetrievalChain
|
11 |
+
from langchain_community.embeddings import HuggingFaceEmbeddings
|
12 |
+
from langchain_community.llms import HuggingFacePipeline
|
13 |
+
from langchain.chains import ConversationChain
|
14 |
+
from langchain.memory import ConversationBufferMemory
|
15 |
+
from langchain_community.llms import HuggingFaceEndpoint
|
16 |
+
import torch
|
17 |
+
|
18 |
+
list_llm = ["meta-llama/Meta-Llama-3-8B-Instruct", "mistralai/Mistral-7B-Instruct-v0.2"]
|
19 |
+
list_llm_simple = [os.path.basename(llm) for llm in list_llm]
|
20 |
+
|
21 |
+
# Load and split PDF document
|
22 |
+
def load_doc(list_file_path):
|
23 |
+
# Processing for one document only
|
24 |
+
# loader = PyPDFLoader(file_path)
|
25 |
+
# pages = loader.load()
|
26 |
+
loaders = [PyPDFLoader(x) for x in list_file_path]
|
27 |
+
pages = []
|
28 |
+
for loader in loaders:
|
29 |
+
pages.extend(loader.load())
|
30 |
+
text_splitter = RecursiveCharacterTextSplitter(
|
31 |
+
chunk_size = 1024,
|
32 |
+
chunk_overlap = 64
|
33 |
+
)
|
34 |
+
doc_splits = text_splitter.split_documents(pages)
|
35 |
+
return doc_splits
|
36 |
+
|
37 |
+
# Create vector database
|
38 |
+
def create_db(splits):
|
39 |
+
embeddings = HuggingFaceEmbeddings()
|
40 |
+
vectordb = FAISS.from_documents(splits, embeddings)
|
41 |
+
return vectordb
|
42 |
+
|
43 |
+
|
44 |
+
# Initialize langchain LLM chain
|
45 |
+
def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, progress=gr.Progress()):
|
46 |
+
if llm_model == "meta-llama/Meta-Llama-3-8B-Instruct":
|
47 |
+
llm = HuggingFaceEndpoint(
|
48 |
+
repo_id=llm_model,
|
49 |
+
huggingfacehub_api_token = api_token,
|
50 |
+
temperature = temperature,
|
51 |
+
max_new_tokens = max_tokens,
|
52 |
+
top_k = top_k,
|
53 |
+
)
|
54 |
+
else:
|
55 |
+
llm = HuggingFaceEndpoint(
|
56 |
+
huggingfacehub_api_token = api_token,
|
57 |
+
repo_id=llm_model,
|
58 |
+
temperature = temperature,
|
59 |
+
max_new_tokens = max_tokens,
|
60 |
+
top_k = top_k,
|
61 |
+
)
|
62 |
+
|
63 |
+
memory = ConversationBufferMemory(
|
64 |
+
memory_key="chat_history",
|
65 |
+
output_key='answer',
|
66 |
+
return_messages=True
|
67 |
+
)
|
68 |
+
|
69 |
+
retriever=vector_db.as_retriever()
|
70 |
+
qa_chain = ConversationalRetrievalChain.from_llm(
|
71 |
+
llm,
|
72 |
+
retriever=retriever,
|
73 |
+
chain_type="stuff",
|
74 |
+
memory=memory,
|
75 |
+
return_source_documents=True,
|
76 |
+
verbose=False,
|
77 |
+
)
|
78 |
+
return qa_chain
|
79 |
+
|
80 |
+
# Initialize database
|
81 |
+
def initialize_database(list_file_obj, progress=gr.Progress()):
|
82 |
+
# Create a list of documents (when valid)
|
83 |
+
list_file_path = [x.name for x in list_file_obj if x is not None]
|
84 |
+
# Load document and create splits
|
85 |
+
doc_splits = load_doc(list_file_path)
|
86 |
+
# Create or load vector database
|
87 |
+
vector_db = create_db(doc_splits)
|
88 |
+
return vector_db, "Database created!"
|
89 |
+
|
90 |
+
# Initialize LLM
|
91 |
+
def initialize_LLM(llm_option, llm_temperature, max_tokens, top_k, vector_db, progress=gr.Progress()):
|
92 |
+
# print("llm_option",llm_option)
|
93 |
+
llm_name = list_llm[llm_option]
|
94 |
+
print("llm_name: ",llm_name)
|
95 |
+
qa_chain = initialize_llmchain(llm_name, llm_temperature, max_tokens, top_k, vector_db, progress)
|
96 |
+
return qa_chain, "QA chain initialized. Chatbot is ready!"
|
97 |
+
|
98 |
+
|
99 |
+
def format_chat_history(message, chat_history):
|
100 |
+
formatted_chat_history = []
|
101 |
+
for user_message, bot_message in chat_history:
|
102 |
+
formatted_chat_history.append(f"User: {user_message}")
|
103 |
+
formatted_chat_history.append(f"Assistant: {bot_message}")
|
104 |
+
return formatted_chat_history
|
105 |
+
|
106 |
+
|
107 |
+
def conversation(qa_chain, message, history):
|
108 |
+
formatted_chat_history = format_chat_history(message, history)
|
109 |
+
# Generate response using QA chain
|
110 |
+
response = qa_chain.invoke({"question": message, "chat_history": formatted_chat_history})
|
111 |
+
response_answer = response["answer"]
|
112 |
+
if response_answer.find("Helpful Answer:") != -1:
|
113 |
+
response_answer = response_answer.split("Helpful Answer:")[-1]
|
114 |
+
response_sources = response["source_documents"]
|
115 |
+
response_source1 = response_sources[0].page_content.strip()
|
116 |
+
response_source2 = response_sources[1].page_content.strip()
|
117 |
+
response_source3 = response_sources[2].page_content.strip()
|
118 |
+
# Langchain sources are zero-based
|
119 |
+
response_source1_page = response_sources[0].metadata["page"] + 1
|
120 |
+
response_source2_page = response_sources[1].metadata["page"] + 1
|
121 |
+
response_source3_page = response_sources[2].metadata["page"] + 1
|
122 |
+
# Append user message and response to chat history
|
123 |
+
new_history = history + [(message, response_answer)]
|
124 |
+
return qa_chain, gr.update(value=""), new_history, response_source1, response_source1_page, response_source2, response_source2_page, response_source3, response_source3_page
|
125 |
+
|
126 |
+
|
127 |
+
def upload_file(file_obj):
|
128 |
+
list_file_path = []
|
129 |
+
for idx, file in enumerate(file_obj):
|
130 |
+
file_path = file_obj.name
|
131 |
+
list_file_path.append(file_path)
|
132 |
+
return list_file_path
|
133 |
+
|
134 |
+
|
135 |
+
def demo():
|
136 |
+
with gr.Blocks(theme=gr.themes.Default(primary_hue="sky")) as demo:
|
137 |
+
vector_db = gr.State()
|
138 |
+
qa_chain = gr.State()
|
139 |
+
gr.HTML("<center><h1>RAG PDF chatbot</h1><center>")
|
140 |
+
gr.Markdown("""<b>Query your PDF documents!</b> This AI agent is designed to perform retrieval augmented generation (RAG) on PDF documents. The app is hosted on Hugging Face Hub for the sole purpose of demonstration. \
|
141 |
+
<b>Please do not upload confidential documents.</b>
|
142 |
+
""")
|
143 |
+
with gr.Row():
|
144 |
+
with gr.Column(scale = 86):
|
145 |
+
gr.Markdown("<b>Step 1 - Upload PDF documents and Initialize RAG pipeline</b>")
|
146 |
+
with gr.Row():
|
147 |
+
document = gr.Files(height=300, file_count="multiple", file_types=["pdf"], interactive=True, label="Upload PDF documents")
|
148 |
+
with gr.Row():
|
149 |
+
db_btn = gr.Button("Create vector database")
|
150 |
+
with gr.Row():
|
151 |
+
db_progress = gr.Textbox(value="Not initialized", show_label=False) # label="Vector database status",
|
152 |
+
gr.Markdown("<style>body { font-size: 16px; }</style><b>Select Large Language Model (LLM) and input parameters</b>")
|
153 |
+
with gr.Row():
|
154 |
+
llm_btn = gr.Radio(list_llm_simple, label="Available LLMs", value = list_llm_simple[0], type="index") # info="Select LLM", show_label=False
|
155 |
+
with gr.Row():
|
156 |
+
with gr.Accordion("LLM input parameters", open=False):
|
157 |
+
with gr.Row():
|
158 |
+
slider_temperature = gr.Slider(minimum = 0.01, maximum = 1.0, value=0.5, step=0.1, label="Temperature", info="Controls randomness in token generation", interactive=True)
|
159 |
+
with gr.Row():
|
160 |
+
slider_maxtokens = gr.Slider(minimum = 128, maximum = 9192, value=4096, step=128, label="Max New Tokens", info="Maximum number of tokens to be generated",interactive=True)
|
161 |
+
with gr.Row():
|
162 |
+
slider_topk = gr.Slider(minimum = 1, maximum = 10, value=3, step=1, label="top-k", info="Number of tokens to select the next token from", interactive=True)
|
163 |
+
with gr.Row():
|
164 |
+
qachain_btn = gr.Button("Initialize Question Answering Chatbot")
|
165 |
+
with gr.Row():
|
166 |
+
llm_progress = gr.Textbox(value="Not initialized", show_label=False) # label="Chatbot status",
|
167 |
+
|
168 |
+
with gr.Column(scale = 200):
|
169 |
+
gr.Markdown("<b>Step 2 - Chat with your Document</b>")
|
170 |
+
chatbot = gr.Chatbot(height=505)
|
171 |
+
with gr.Accordion("Relevent context from the source document", open=False):
|
172 |
+
with gr.Row():
|
173 |
+
doc_source1 = gr.Textbox(label="Reference 1", lines=2, container=True, scale=20)
|
174 |
+
source1_page = gr.Number(label="Page", scale=1)
|
175 |
+
with gr.Row():
|
176 |
+
doc_source2 = gr.Textbox(label="Reference 2", lines=2, container=True, scale=20)
|
177 |
+
source2_page = gr.Number(label="Page", scale=1)
|
178 |
+
with gr.Row():
|
179 |
+
doc_source3 = gr.Textbox(label="Reference 3", lines=2, container=True, scale=20)
|
180 |
+
source3_page = gr.Number(label="Page", scale=1)
|
181 |
+
with gr.Row():
|
182 |
+
msg = gr.Textbox(placeholder="Ask a question", container=True)
|
183 |
+
with gr.Row():
|
184 |
+
submit_btn = gr.Button("Submit")
|
185 |
+
clear_btn = gr.ClearButton([msg, chatbot], value="Clear")
|
186 |
+
|
187 |
+
# Preprocessing events
|
188 |
+
db_btn.click(initialize_database, \
|
189 |
+
inputs=[document], \
|
190 |
+
outputs=[vector_db, db_progress])
|
191 |
+
qachain_btn.click(initialize_LLM, \
|
192 |
+
inputs=[llm_btn, slider_temperature, slider_maxtokens, slider_topk, vector_db], \
|
193 |
+
outputs=[qa_chain, llm_progress]).then(lambda:[None,"",0,"",0,"",0], \
|
194 |
+
inputs=None, \
|
195 |
+
outputs=[chatbot, doc_source1, source1_page, doc_source2, source2_page, doc_source3, source3_page], \
|
196 |
+
queue=False)
|
197 |
|
198 |
+
# Chatbot events
|
199 |
+
msg.submit(conversation, \
|
200 |
+
inputs=[qa_chain, msg, chatbot], \
|
201 |
+
outputs=[qa_chain, msg, chatbot, doc_source1, source1_page, doc_source2, source2_page, doc_source3, source3_page], \
|
202 |
+
queue=False)
|
203 |
+
submit_btn.click(conversation, \
|
204 |
+
inputs=[qa_chain, msg, chatbot], \
|
205 |
+
outputs=[qa_chain, msg, chatbot, doc_source1, source1_page, doc_source2, source2_page, doc_source3, source3_page], \
|
206 |
+
queue=False)
|
207 |
+
clear_btn.click(lambda:[None,"",0,"",0,"",0], \
|
208 |
+
inputs=None, \
|
209 |
+
outputs=[chatbot, doc_source1, source1_page, doc_source2, source2_page, doc_source3, source3_page], \
|
210 |
+
queue=False)
|
211 |
+
demo.queue().launch(debug=True)
|
212 |
|
|
|
|
|
213 |
|
214 |
+
if __name__ == "__main__":
|
215 |
+
demo()
|