Anne31415 commited on
Commit
ea5f78c
·
1 Parent(s): 35fb3be

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +11 -172
app.py CHANGED
@@ -1,178 +1,17 @@
1
  import streamlit as st
2
- from dotenv import load_dotenv
3
- import pickle
4
- from huggingface_hub import Repository
5
- from PyPDF2 import PdfReader
6
- from streamlit_extras.add_vertical_space import add_vertical_space
7
- from langchain.text_splitter import RecursiveCharacterTextSplitter
8
- from langchain.embeddings.openai import OpenAIEmbeddings
9
- from langchain.vectorstores import FAISS
10
- from langchain.llms import OpenAI
11
- from langchain.chains.question_answering import load_qa_chain
12
- from langchain.callbacks import get_openai_callback
13
- import os
14
 
15
- # Step 1: Clone the Dataset Repository
16
- repo = Repository(
17
- local_dir="Private_Book", # Local directory to clone the repository
18
- repo_type="dataset", # Specify that this is a dataset repository
19
-
20
- clone_from="Anne31415/Private_Book", # Replace with your repository URL
21
-
22
- token=os.environ["HUB_TOKEN"] # Use the secret token to authenticate
23
- )
24
- repo.git_pull() # Pull the latest changes (if any)
25
-
26
- # Step 2: Load the PDF File
27
- pdf_file_path = "Private_Book/KOMBI_all2.pdf" # Replace with your PDF file path
28
-
29
-
30
- def cloud_button(label, query, key=None, color=None, overlap=30):
31
- button_id = f"cloud-button-{key or label}".replace(" ", "-")
32
- color_class = f"color-{color}" if color else ""
33
- num_circles = max(3, min(35, len(label) // 4))
34
- circle_size = 60
35
-
36
- circles_html = ''.join([
37
- f'<div class="circle {color_class}" style="margin-right: -{overlap}px;"></div>'
38
- for _ in range(num_circles)
39
- ])
40
- circles_html += f'<div class="circle-text">{label}</div>'
41
-
42
- cloud_button_html = f"""
43
- <div class="cloud" id="{button_id}" style="margin-bottom: 20px; cursor: pointer;">
44
- <div class="wrapper {color_class}">
45
- {circles_html}
46
- </div>
47
  </div>
48
- <script>
49
- document.getElementById("{button_id}").onclick = function() {{
50
- const query = "{query}";
51
- const label = "{label}";
52
- const button_id = "{button_id}";
53
- window.parent.postMessage({{
54
- 'isStreamlitMessage': true,
55
- 'type': 'streamlit:setComponentValue',
56
- 'value': {{'label': label, 'query': query, 'button_id': button_id}},
57
- 'key': 'button_clicked'
58
- }}, '*');
59
- }};
60
- </script>
61
  """
62
- st.markdown(cloud_button_html, unsafe_allow_html=True)
63
-
64
-
65
- def display_chat_history(chat_history):
66
- for sender, msg, _ in chat_history:
67
- background_color = "#FFA07A" if sender == "User" else "#caf"
68
- st.markdown(f"<div style='background-color: {background_color}; padding: 10px; border-radius: 10px; margin: 10px;'>{sender}: {msg}</div>", unsafe_allow_html=True)
69
-
70
-
71
- def load_pdf(file_path):
72
- pdf_reader = PdfReader(file_path)
73
- text = ""
74
- for page in pdf_reader.pages:
75
- text += page.extract_text()
76
-
77
- text_splitter = RecursiveCharacterTextSplitter(
78
- chunk_size=1000,
79
- chunk_overlap=200,
80
- length_function=len
81
- )
82
- chunks = text_splitter.split_text(text=text)
83
-
84
- store_name, _ = os.path.splitext(os.path.basename(file_path))
85
-
86
- if os.path.exists(f"{store_name}.pkl"):
87
- with open(f"{store_name}.pkl", "rb") as f:
88
- VectorStore = pickle.load(f)
89
- else:
90
- embeddings = OpenAIEmbeddings()
91
- VectorStore = FAISS.from_texts(chunks, embedding=embeddings)
92
- with open(f"{store_name}.pkl", "wb") as f:
93
- pickle.dump(VectorStore, f)
94
-
95
- return VectorStore
96
-
97
-
98
-
99
- def load_chatbot():
100
- return load_qa_chain(llm=OpenAI(), chain_type="stuff")
101
-
102
-
103
- def main():
104
- hide_streamlit_style = """
105
- <style>
106
- #MainMenu {visibility: hidden;}
107
- footer {visibility: hidden;}
108
- </style>
109
- """
110
- st.markdown(hide_streamlit_style, unsafe_allow_html=True)
111
-
112
- st.title("Welcome to BinDocs ChatBot! 🤖")
113
-
114
- pdf_path = pdf_file_path
115
- if not os.path.exists(pdf_path):
116
- st.error("File not found. Please check the file path.")
117
- return
118
-
119
- if "chat_history" not in st.session_state:
120
- st.session_state['chat_history'] = []
121
-
122
- display_chat_history(st.session_state['chat_history'])
123
-
124
- st.write("<!-- Start Spacer -->", unsafe_allow_html=True)
125
- st.write("<div style='flex: 1;'></div>", unsafe_allow_html=True)
126
- st.write("<!-- End Spacer -->", unsafe_allow_html=True)
127
-
128
- if pdf_path is not None:
129
- query = st.text_input("Ask questions about your PDF file (in any preferred language):", key="user_query")
130
-
131
- cloud_buttons = [
132
- ("Was genau ist ein Belegarzt?", "Was genau ist ein Belegarzt?", "1"),
133
- ("Wofür wird die Alpha-ID verwendet?", "Wofür wird die Alpha-ID verwendet?", "2"),
134
- # Add more buttons as needed
135
- ]
136
-
137
- for label, query, color in cloud_buttons:
138
- cloud_button(label, query, color=color)
139
-
140
- user_input = st.empty()
141
-
142
- if "button_clicked" in st.session_state:
143
- button_info = st.session_state["button_clicked"]
144
- if button_info:
145
- st.write(f"You clicked: {button_info['label']}")
146
- st.write(f"Query: {button_info['query']}")
147
- # Handle the button click as needed
148
- # For example, you can call a function to process the query
149
- # process_query(button_info['query'])
150
- st.session_state["button_clicked"] = None # Reset after handling
151
-
152
-
153
- if st.button("Ask"):
154
- user_input = st.session_state.user_query
155
- handle_query(user_input, pdf_path)
156
-
157
- def handle_query(query, pdf_path):
158
- if not query:
159
- st.warning("Please enter a query.")
160
- return
161
-
162
- st.session_state['chat_history'].append(("User", query, "new"))
163
- loading_message = st.empty()
164
- loading_message.text('Bot is thinking...')
165
-
166
- VectorStore = load_pdf(pdf_path)
167
- chain = load_chatbot()
168
- docs = VectorStore.similarity_search(query=query, k=3)
169
- with get_openai_callback() as cb:
170
- response = chain.run(input_documents=docs, question=query)
171
 
172
- st.session_state['chat_history'].append(("Bot", response, "new"))
173
- display_chat_history(st.session_state['chat_history'][-2:])
174
- loading_message.empty()
175
- st.session_state['chat_history'] = [(sender, msg, "old") for sender, msg, _ in st.session_state['chat_history']]
176
 
177
- if __name__ == "__main__":
178
- main()
 
1
  import streamlit as st
 
 
 
 
 
 
 
 
 
 
 
 
2
 
3
+ def cloud_button(label, key):
4
+ button_html = f"""
5
+ <div style="text-align: center;">
6
+ <div style="display: inline-block; background-color: #FF6347; border-radius: 50%; width: 100px; height: 100px;"></div>
7
+ <div style="margin-top: -100px;">{label}</div>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8
  </div>
 
 
 
 
 
 
 
 
 
 
 
 
 
9
  """
10
+ st.markdown(button_html, unsafe_allow_html=True)
11
+ return st.button("", key=key, use_container_width=True)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12
 
13
+ if cloud_button("Click Me!", "button1"):
14
+ st.success("Button 1 clicked!")
 
 
15
 
16
+ if cloud_button("Another Button", "button2"):
17
+ st.success("Button 2 clicked!")