Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,54 +1,97 @@
|
|
1 |
import streamlit as st
|
2 |
-
|
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 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
|
53 |
def load_pdf(file_path):
|
54 |
pdf_reader = PdfReader(file_path)
|
@@ -76,13 +119,36 @@ def load_pdf(file_path):
|
|
76 |
|
77 |
return VectorStore
|
78 |
|
79 |
-
|
80 |
-
|
81 |
def load_chatbot():
|
82 |
return load_qa_chain(llm=OpenAI(), chain_type="stuff")
|
83 |
|
84 |
-
def
|
|
|
|
|
|
|
85 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
86 |
hide_streamlit_style = """
|
87 |
<style>
|
88 |
#MainMenu {visibility: hidden;}
|
@@ -91,11 +157,8 @@ def main():
|
|
91 |
"""
|
92 |
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
|
93 |
|
94 |
-
|
95 |
# Main content
|
96 |
st.title("Welcome to BinDocs ChatBot! 🤖")
|
97 |
-
|
98 |
-
# Directly specifying the path to the PDF file
|
99 |
pdf_path = pdf_file_path
|
100 |
if not os.path.exists(pdf_path):
|
101 |
st.error("File not found. Please check the file path.")
|
@@ -112,74 +175,39 @@ def main():
|
|
112 |
|
113 |
new_messages_placeholder = st.empty()
|
114 |
|
115 |
-
# Add cloud-like button styles
|
116 |
-
cloud_button_style = """
|
117 |
-
<style>
|
118 |
-
.stButton > button {
|
119 |
-
border: none;
|
120 |
-
padding: 10px 20px;
|
121 |
-
border-radius: 25px;
|
122 |
-
font-size: 16px;
|
123 |
-
transition-duration: 0.4s;
|
124 |
-
cursor: pointer;
|
125 |
-
background-color: white;
|
126 |
-
color: black;
|
127 |
-
border: 2px solid #008CBA;
|
128 |
-
}
|
129 |
-
.stButton > button:hover {
|
130 |
-
background-color: #008CBA;
|
131 |
-
color: white;
|
132 |
-
}
|
133 |
-
</style>
|
134 |
-
"""
|
135 |
-
st.markdown(cloud_button_style, unsafe_allow_html=True)
|
136 |
-
|
137 |
if pdf_path is not None:
|
138 |
query = st.text_input("Ask questions about your PDF file (in any preferred language):")
|
139 |
|
140 |
-
if
|
141 |
query = "Was genau ist ein Belegarzt?"
|
142 |
-
if
|
143 |
query = "Wofür wird die Alpha-ID verwendet?"
|
144 |
-
|
145 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
146 |
if st.button("Ask") or (not st.session_state['chat_history'] and query) or (st.session_state['chat_history'] and query != st.session_state['chat_history'][-1][1]):
|
147 |
st.session_state['chat_history'].append(("User", query, "new"))
|
148 |
-
|
149 |
loading_message = st.empty()
|
150 |
loading_message.text('Bot is thinking...')
|
151 |
-
|
152 |
VectorStore = load_pdf(pdf_path)
|
153 |
chain = load_chatbot()
|
154 |
docs = VectorStore.similarity_search(query=query, k=3)
|
155 |
-
|
156 |
-
response = chain.run(input_documents=docs, question=query)
|
157 |
-
|
158 |
st.session_state['chat_history'].append(("Bot", response, "new"))
|
159 |
-
|
160 |
-
# Display new messages at the bottom
|
161 |
new_messages = st.session_state['chat_history'][-2:]
|
162 |
for chat in new_messages:
|
163 |
background_color = "#FFA07A" if chat[2] == "new" else "#acf" if chat[0] == "User" else "#caf"
|
164 |
new_messages_placeholder.markdown(f"<div style='background-color: {background_color}; padding: 10px; border-radius: 10px; margin: 10px;'>{chat[0]}: {chat[1]}</div>", unsafe_allow_html=True)
|
165 |
-
|
166 |
-
# Scroll to the latest response using JavaScript
|
167 |
st.write("<script>document.getElementById('response').scrollIntoView();</script>", unsafe_allow_html=True)
|
168 |
-
|
169 |
loading_message.empty()
|
170 |
-
|
171 |
-
# Clear the input field by setting the query variable to an empty string
|
172 |
query = ""
|
173 |
-
|
174 |
-
# Mark all messages as old after displaying
|
175 |
st.session_state['chat_history'] = [(sender, msg, "old") for sender, msg, _ in st.session_state['chat_history']]
|
176 |
|
177 |
-
|
178 |
-
|
179 |
-
def display_chat_history(chat_history):
|
180 |
-
for chat in chat_history:
|
181 |
-
background_color = "#FFA07A" if chat[2] == "new" else "#acf" if chat[0] == "User" else "#caf"
|
182 |
-
st.markdown(f"<div style='background-color: {background_color}; padding: 10px; border-radius: 10px; margin: 10px;'>{chat[0]}: {chat[1]}</div>", unsafe_allow_html=True)
|
183 |
-
|
184 |
if __name__ == "__main__":
|
185 |
-
main()
|
|
|
1 |
import streamlit as st
|
2 |
+
import os
|
3 |
import pickle
|
4 |
from huggingface_hub import Repository
|
5 |
from PyPDF2 import PdfReader
|
|
|
6 |
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
7 |
from langchain.embeddings.openai import OpenAIEmbeddings
|
8 |
from langchain.vectorstores import FAISS
|
9 |
from langchain.llms import OpenAI
|
10 |
from langchain.chains.question_answering import load_qa_chain
|
11 |
from langchain.callbacks import get_openai_callback
|
|
|
12 |
|
13 |
# Step 1: Clone the Dataset Repository
|
14 |
repo = Repository(
|
15 |
local_dir="Private_Book", # Local directory to clone the repository
|
16 |
repo_type="dataset", # Specify that this is a dataset repository
|
|
|
17 |
clone_from="Anne31415/Private_Book", # Replace with your repository URL
|
18 |
+
token=os.getenv("HUB_TOKEN") # Use the secret token to authenticate
|
|
|
19 |
)
|
20 |
repo.git_pull() # Pull the latest changes (if any)
|
21 |
|
22 |
# Step 2: Load the PDF File
|
23 |
pdf_file_path = "Private_Book/KOMBI_all2.pdf" # Replace with your PDF file path
|
24 |
|
25 |
+
def cloud_button(label, key=None):
|
26 |
+
button_id = f"cloud-button-{key or label}"
|
27 |
+
cloud_button_html = f"""
|
28 |
+
<div class="cloud" id="{button_id}">
|
29 |
+
<div class="circle small"></div>
|
30 |
+
<div class="circle medium"></div>
|
31 |
+
<div class="circle large"></div>
|
32 |
+
<div class="rectangle">{label}</div>
|
33 |
+
<div class="circle medium"></div>
|
34 |
+
<div class="circle small"></div>
|
35 |
+
</div>
|
36 |
+
<style>
|
37 |
+
.cloud {{
|
38 |
+
position: relative;
|
39 |
+
display: inline-block;
|
40 |
+
cursor: pointer;
|
41 |
+
user-select: none;
|
42 |
+
}}
|
43 |
+
.rectangle {{
|
44 |
+
width: 120px;
|
45 |
+
height: 60px;
|
46 |
+
background-color: white;
|
47 |
+
border-radius: 30px;
|
48 |
+
position: absolute;
|
49 |
+
top: 20px;
|
50 |
+
left: 10px;
|
51 |
+
display: flex;
|
52 |
+
align-items: center;
|
53 |
+
justify-content: center;
|
54 |
+
font-size: 16px;
|
55 |
+
font-weight: bold;
|
56 |
+
transition: background-color 0.4s;
|
57 |
+
}}
|
58 |
+
.circle {{
|
59 |
+
position: absolute;
|
60 |
+
background-color: white;
|
61 |
+
}}
|
62 |
+
.circle.small {{
|
63 |
+
width: 40px;
|
64 |
+
height: 40px;
|
65 |
+
border-radius: 20px;
|
66 |
+
top: 10px;
|
67 |
+
left: 50px;
|
68 |
+
}}
|
69 |
+
.circle.medium {{
|
70 |
+
width: 60px;
|
71 |
+
height: 60px;
|
72 |
+
border-radius: 30px;
|
73 |
+
}}
|
74 |
+
.circle.large {{
|
75 |
+
width: 80px;
|
76 |
+
height: 80px;
|
77 |
+
border-radius: 40px;
|
78 |
+
top: -10px;
|
79 |
+
left: 40px;
|
80 |
+
}}
|
81 |
+
.cloud:hover .rectangle {{
|
82 |
+
background-color: #008CBA;
|
83 |
+
color: white;
|
84 |
+
}}
|
85 |
+
</style>
|
86 |
+
<script>
|
87 |
+
document.getElementById("{button_id}").onclick = function() {{
|
88 |
+
google.script.run.withSuccessHandler(function(e) {{
|
89 |
+
document.getElementById("{button_id}").innerText = e;
|
90 |
+
}}).getCloudButtonValue("{label}");
|
91 |
+
}};
|
92 |
+
</script>
|
93 |
+
"""
|
94 |
+
st.markdown(cloud_button_html, unsafe_allow_html=True)
|
95 |
|
96 |
def load_pdf(file_path):
|
97 |
pdf_reader = PdfReader(file_path)
|
|
|
119 |
|
120 |
return VectorStore
|
121 |
|
|
|
|
|
122 |
def load_chatbot():
|
123 |
return load_qa_chain(llm=OpenAI(), chain_type="stuff")
|
124 |
|
125 |
+
def display_chat_history(chat_history):
|
126 |
+
for chat in chat_history:
|
127 |
+
background_color = "#FFA07A" if chat[2] == "new" else "#acf" if chat[0] == "User" else "#caf"
|
128 |
+
st.markdown(f"<div style='background-color: {background_color}; padding: 10px; border-radius: 10px; margin: 10px;'>{chat[0]}: {chat[1]}</div>", unsafe_allow_html=True)
|
129 |
|
130 |
+
def main():
|
131 |
+
with st.sidebar:
|
132 |
+
st.title('BinDoc GmbH')
|
133 |
+
st.markdown("Experience revolutionary interaction with BinDocs Chat App, leveraging state-of-the-art AI technology.")
|
134 |
+
|
135 |
+
add_vertical_space(1) # Adjust as per the desired spacing
|
136 |
+
|
137 |
+
st.markdown("""
|
138 |
+
Hello! I’m here to assist you with:<br><br>
|
139 |
+
📘 **Glossary Inquiries:**<br>
|
140 |
+
I can clarify terms like "DiGA", "AOP", or "BfArM", providing clear and concise explanations to help you understand our content better.<br><br>
|
141 |
+
🆘 **Help Page Navigation:**<br>
|
142 |
+
Ask me if you forgot your password or want to know more about topics related to the platform.<br><br>
|
143 |
+
📰 **Latest Whitepapers Insights:**<br>
|
144 |
+
Curious about our recent publications? Feel free to ask about our latest whitepapers!<br><br>
|
145 |
+
""", unsafe_allow_html=True)
|
146 |
+
|
147 |
+
add_vertical_space(1) # Adjust as per the desired spacing
|
148 |
+
st.write('Made with ❤️ by BinDoc GmbH')
|
149 |
+
api_key = os.getenv("OPENAI_API_KEY")
|
150 |
+
|
151 |
+
|
152 |
hide_streamlit_style = """
|
153 |
<style>
|
154 |
#MainMenu {visibility: hidden;}
|
|
|
157 |
"""
|
158 |
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
|
159 |
|
|
|
160 |
# Main content
|
161 |
st.title("Welcome to BinDocs ChatBot! 🤖")
|
|
|
|
|
162 |
pdf_path = pdf_file_path
|
163 |
if not os.path.exists(pdf_path):
|
164 |
st.error("File not found. Please check the file path.")
|
|
|
175 |
|
176 |
new_messages_placeholder = st.empty()
|
177 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
178 |
if pdf_path is not None:
|
179 |
query = st.text_input("Ask questions about your PDF file (in any preferred language):")
|
180 |
|
181 |
+
if cloud_button("Was genau ist ein Belegarzt?"):
|
182 |
query = "Was genau ist ein Belegarzt?"
|
183 |
+
if cloud_button("Wofür wird die Alpha-ID verwendet?"):
|
184 |
query = "Wofür wird die Alpha-ID verwendet?"
|
185 |
+
if cloud_button("Was sind die Vorteile des ambulanten operierens?"):
|
186 |
+
query = "Was sind die Vorteile des ambulanten operierens?"
|
187 |
+
if cloud_button("Was kann ich mit dem Prognose-Analyse Toll machen?"):
|
188 |
+
query = "Was kann ich mit dem Prognose-Analyse Toll machen?"
|
189 |
+
if cloud_button("Was sagt mir die Farbe der Balken der Bevölkerungsentwicklung?"):
|
190 |
+
query = "Was sagt mir die Farbe der Balken der Bevölkerungsentwicklung?"
|
191 |
+
if cloud_button("Ich habe mein Meta Password vergessen, wie kann ich es zurücksetzen?"):
|
192 |
+
query = "Ich habe mein Meta Password vergessen, wie kann ich es zurücksetzen?"
|
193 |
+
|
194 |
if st.button("Ask") or (not st.session_state['chat_history'] and query) or (st.session_state['chat_history'] and query != st.session_state['chat_history'][-1][1]):
|
195 |
st.session_state['chat_history'].append(("User", query, "new"))
|
|
|
196 |
loading_message = st.empty()
|
197 |
loading_message.text('Bot is thinking...')
|
|
|
198 |
VectorStore = load_pdf(pdf_path)
|
199 |
chain = load_chatbot()
|
200 |
docs = VectorStore.similarity_search(query=query, k=3)
|
201 |
+
response = chain.run(input_documents=docs, question=query)
|
|
|
|
|
202 |
st.session_state['chat_history'].append(("Bot", response, "new"))
|
|
|
|
|
203 |
new_messages = st.session_state['chat_history'][-2:]
|
204 |
for chat in new_messages:
|
205 |
background_color = "#FFA07A" if chat[2] == "new" else "#acf" if chat[0] == "User" else "#caf"
|
206 |
new_messages_placeholder.markdown(f"<div style='background-color: {background_color}; padding: 10px; border-radius: 10px; margin: 10px;'>{chat[0]}: {chat[1]}</div>", unsafe_allow_html=True)
|
|
|
|
|
207 |
st.write("<script>document.getElementById('response').scrollIntoView();</script>", unsafe_allow_html=True)
|
|
|
208 |
loading_message.empty()
|
|
|
|
|
209 |
query = ""
|
|
|
|
|
210 |
st.session_state['chat_history'] = [(sender, msg, "old") for sender, msg, _ in st.session_state['chat_history']]
|
211 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
212 |
if __name__ == "__main__":
|
213 |
+
main()
|