Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,6 +1,7 @@
|
|
| 1 |
import streamlit as st
|
| 2 |
from PIL import Image
|
| 3 |
import time
|
|
|
|
| 4 |
from dotenv import load_dotenv
|
| 5 |
import pickle
|
| 6 |
from huggingface_hub import Repository
|
|
@@ -13,7 +14,6 @@ from langchain.llms import OpenAI
|
|
| 13 |
from langchain.chains.question_answering import load_qa_chain
|
| 14 |
from langchain.callbacks import get_openai_callback
|
| 15 |
import os
|
| 16 |
-
import traceback
|
| 17 |
|
| 18 |
import pandas as pd
|
| 19 |
import pydeck as pdk
|
|
@@ -26,6 +26,8 @@ if 'chat_history_page1' not in st.session_state:
|
|
| 26 |
if 'chat_history_page2' not in st.session_state:
|
| 27 |
st.session_state['chat_history_page2'] = []
|
| 28 |
|
|
|
|
|
|
|
| 29 |
# Step 1: Clone the Dataset Repository
|
| 30 |
repo = Repository(
|
| 31 |
local_dir="Private_Book", # Local directory to clone the repository
|
|
@@ -36,15 +38,17 @@ repo = Repository(
|
|
| 36 |
repo.git_pull() # Pull the latest changes (if any)
|
| 37 |
|
| 38 |
# Step 2: Load the PDF File
|
| 39 |
-
pdf_path = "Private_Book/
|
| 40 |
|
| 41 |
# Step 2: Load the PDF File
|
| 42 |
pdf_path2 = "Private_Book/Deutsche_Kodierrichtlinien_23.pdf" # Replace with your PDF file path
|
| 43 |
|
|
|
|
| 44 |
api_key = os.getenv("OPENAI_API_KEY")
|
| 45 |
# Retrieve the API key from st.secrets
|
| 46 |
|
| 47 |
|
|
|
|
| 48 |
# Updated caching mechanism using st.cache_data
|
| 49 |
@st.cache_data(persist="disk") # Using persist="disk" to save cache across sessions
|
| 50 |
def load_vector_store(file_path, store_name, force_reload=False):
|
|
@@ -70,12 +74,6 @@ def load_vector_store(file_path, store_name, force_reload=False):
|
|
| 70 |
|
| 71 |
return VectorStore
|
| 72 |
|
| 73 |
-
except Exception as e:
|
| 74 |
-
st.error(f"An error occurred: {e}")
|
| 75 |
-
traceback.print_exc()
|
| 76 |
-
return None
|
| 77 |
-
|
| 78 |
-
|
| 79 |
# Utility function to load text from a PDF
|
| 80 |
def load_pdf_text(file_path):
|
| 81 |
pdf_reader = PdfReader(file_path)
|
|
@@ -95,6 +93,7 @@ def display_chat_history(chat_history):
|
|
| 95 |
|
| 96 |
|
| 97 |
|
|
|
|
| 98 |
def page1():
|
| 99 |
try:
|
| 100 |
hide_streamlit_style = """
|
|
@@ -105,73 +104,102 @@ def page1():
|
|
| 105 |
"""
|
| 106 |
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
|
| 107 |
|
| 108 |
-
|
|
|
|
| 109 |
|
| 110 |
with col1:
|
| 111 |
-
st.title("Welcome to BinDocs
|
| 112 |
|
| 113 |
with col2:
|
|
|
|
| 114 |
image = Image.open('BinDoc Logo (Quadratisch).png')
|
| 115 |
st.image(image, use_column_width='always')
|
| 116 |
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 120 |
|
| 121 |
-
|
| 122 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 123 |
|
| 124 |
-
|
| 125 |
-
|
| 126 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 127 |
|
| 128 |
-
|
| 129 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 130 |
|
| 131 |
-
col1, col2 = st.columns(2)
|
| 132 |
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
query = "Was kann ich mit dem Prognose-Analyse-Tool machen?"
|
| 136 |
-
if st.button("Was sagt mir die Farbe der Balken der Bevölkerungsentwicklung?"):
|
| 137 |
-
query = "Was sagt mir die Farbe der Balken der Bevölkerungsentwicklung?"
|
| 138 |
-
if st.button("Ich habe mein Meta-Password vergessen, wie kann ich es zurücksetzen?"):
|
| 139 |
-
query = "Ich habe mein Meta-Password vergessen, wie kann ich es zurücksetzen?"
|
| 140 |
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
| 151 |
-
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
end_time = time.time()
|
| 159 |
-
duration = end_time - start_time
|
| 160 |
-
st.text(f"Response time: {duration:.2f} seconds")
|
| 161 |
-
|
| 162 |
-
st.session_state['chat_history_page1'].append(("Bot", response, "new"))
|
| 163 |
-
|
| 164 |
-
new_messages = st.session_state['chat_history_page1'][-2:]
|
| 165 |
-
for chat in new_messages:
|
| 166 |
-
background_color = "#ffeecf"
|
| 167 |
-
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)
|
| 168 |
-
|
| 169 |
-
query = ""
|
| 170 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 171 |
st.session_state['chat_history_page1'] = [(sender, msg, "old") for sender, msg, _ in st.session_state['chat_history_page1']]
|
| 172 |
|
| 173 |
except Exception as e:
|
| 174 |
st.error(f"Upsi, an unexpected error occurred: {e}")
|
|
|
|
|
|
|
| 175 |
|
| 176 |
|
| 177 |
|
|
@@ -284,7 +312,9 @@ def page2():
|
|
| 284 |
st.error(f"Upsi, an unexpected error occurred: {e}")
|
| 285 |
# Optionally log the exception details to a file or error tracking service
|
| 286 |
|
| 287 |
-
|
|
|
|
|
|
|
| 288 |
|
| 289 |
def main():
|
| 290 |
# Sidebar content
|
|
|
|
| 1 |
import streamlit as st
|
| 2 |
from PIL import Image
|
| 3 |
import time
|
| 4 |
+
import streamlit_analytics
|
| 5 |
from dotenv import load_dotenv
|
| 6 |
import pickle
|
| 7 |
from huggingface_hub import Repository
|
|
|
|
| 14 |
from langchain.chains.question_answering import load_qa_chain
|
| 15 |
from langchain.callbacks import get_openai_callback
|
| 16 |
import os
|
|
|
|
| 17 |
|
| 18 |
import pandas as pd
|
| 19 |
import pydeck as pdk
|
|
|
|
| 26 |
if 'chat_history_page2' not in st.session_state:
|
| 27 |
st.session_state['chat_history_page2'] = []
|
| 28 |
|
| 29 |
+
|
| 30 |
+
|
| 31 |
# Step 1: Clone the Dataset Repository
|
| 32 |
repo = Repository(
|
| 33 |
local_dir="Private_Book", # Local directory to clone the repository
|
|
|
|
| 38 |
repo.git_pull() # Pull the latest changes (if any)
|
| 39 |
|
| 40 |
# Step 2: Load the PDF File
|
| 41 |
+
pdf_path = "Private_Book/141123_Kombi.pdf" # Replace with your PDF file path
|
| 42 |
|
| 43 |
# Step 2: Load the PDF File
|
| 44 |
pdf_path2 = "Private_Book/Deutsche_Kodierrichtlinien_23.pdf" # Replace with your PDF file path
|
| 45 |
|
| 46 |
+
|
| 47 |
api_key = os.getenv("OPENAI_API_KEY")
|
| 48 |
# Retrieve the API key from st.secrets
|
| 49 |
|
| 50 |
|
| 51 |
+
|
| 52 |
# Updated caching mechanism using st.cache_data
|
| 53 |
@st.cache_data(persist="disk") # Using persist="disk" to save cache across sessions
|
| 54 |
def load_vector_store(file_path, store_name, force_reload=False):
|
|
|
|
| 74 |
|
| 75 |
return VectorStore
|
| 76 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 77 |
# Utility function to load text from a PDF
|
| 78 |
def load_pdf_text(file_path):
|
| 79 |
pdf_reader = PdfReader(file_path)
|
|
|
|
| 93 |
|
| 94 |
|
| 95 |
|
| 96 |
+
|
| 97 |
def page1():
|
| 98 |
try:
|
| 99 |
hide_streamlit_style = """
|
|
|
|
| 104 |
"""
|
| 105 |
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
|
| 106 |
|
| 107 |
+
# Create columns for layout
|
| 108 |
+
col1, col2 = st.columns([3, 1]) # Adjust the ratio to your liking
|
| 109 |
|
| 110 |
with col1:
|
| 111 |
+
st.title("Welcome to BinDocs AI!")
|
| 112 |
|
| 113 |
with col2:
|
| 114 |
+
# Load and display the image in the right column, which will be the top-right corner of the page
|
| 115 |
image = Image.open('BinDoc Logo (Quadratisch).png')
|
| 116 |
st.image(image, use_column_width='always')
|
| 117 |
|
| 118 |
+
|
| 119 |
+
# Start tracking user interactions
|
| 120 |
+
with streamlit_analytics.track():
|
| 121 |
+
if not os.path.exists(pdf_path):
|
| 122 |
+
st.error("File not found. Please check the file path.")
|
| 123 |
+
return
|
| 124 |
+
|
| 125 |
+
VectorStore = load_vector_store(pdf_path, "vector_store_page1", force_reload=False)
|
| 126 |
|
| 127 |
+
display_chat_history(st.session_state['chat_history_page1'])
|
| 128 |
+
|
| 129 |
+
st.write("<!-- Start Spacer -->", unsafe_allow_html=True)
|
| 130 |
+
st.write("<div style='flex: 1;'></div>", unsafe_allow_html=True)
|
| 131 |
+
st.write("<!-- End Spacer -->", unsafe_allow_html=True)
|
| 132 |
+
|
| 133 |
+
new_messages_placeholder = st.empty()
|
| 134 |
|
| 135 |
+
query = st.text_input("Geben Sie hier Ihre Frage ein / Enter your question here:")
|
| 136 |
+
|
| 137 |
+
add_vertical_space(2) # Adjust as per the desired spacing
|
| 138 |
+
|
| 139 |
+
# Create two columns for the buttons
|
| 140 |
+
col1, col2 = st.columns(2)
|
| 141 |
+
|
| 142 |
+
with col1:
|
| 143 |
+
if st.button("Was kann ich mit dem Prognose-Analyse-Tool machen?"):
|
| 144 |
+
query = "Was kann ich mit dem Prognose-Analyse-Tool machen?"
|
| 145 |
+
if st.button("Was sagt mir die Farbe der Balken der Bevölkerungsentwicklung?"):
|
| 146 |
+
query = "Was sagt mir die Farbe der Balken der Bevölkerungsentwicklung?"
|
| 147 |
+
if st.button("Ich habe mein Meta-Password vergessen, wie kann ich es zurücksetzen?"):
|
| 148 |
+
query = "Ich habe mein Meta-Password vergessen, wie kann ich es zurücksetzen?"
|
| 149 |
|
| 150 |
+
|
| 151 |
+
with col2:
|
| 152 |
+
if st.button("Dies ist eine reine Test Frage, welche aber eine ausreichende Länge hat."):
|
| 153 |
+
query = "Dies ist eine reine Test Frage, welche aber eine ausreichende Länge hat."
|
| 154 |
+
if st.button("Was sagt mir denn generell die wundervolle Bevölkerungsentwicklung?"):
|
| 155 |
+
query = "Was sagt mir denn generell die wundervolle Bevölkerungsentwicklung?"
|
| 156 |
+
if st.button("Ob ich hier wohl viel schreibe, dass die Fragen vom Layout her passen?"):
|
| 157 |
+
query = "Ob ich hier wohl viel schreibe, dass die Fragen vom Layout her passen?"
|
| 158 |
|
|
|
|
| 159 |
|
| 160 |
+
if query:
|
| 161 |
+
st.session_state['chat_history_page1'].append(("User", query, "new"))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 162 |
|
| 163 |
+
# Start timing
|
| 164 |
+
start_time = time.time()
|
| 165 |
+
|
| 166 |
+
with st.spinner('Bot is thinking...'):
|
| 167 |
+
# Use the VectorStore loaded at the start from the session state
|
| 168 |
+
chain = load_chatbot()
|
| 169 |
+
docs = VectorStore.similarity_search(query=query, k=3)
|
| 170 |
+
with get_openai_callback() as cb:
|
| 171 |
+
response = chain.run(input_documents=docs, question=query)
|
| 172 |
+
|
| 173 |
+
|
| 174 |
+
# Stop timing
|
| 175 |
+
end_time = time.time()
|
| 176 |
+
|
| 177 |
+
# Calculate duration
|
| 178 |
+
duration = end_time - start_time
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 179 |
|
| 180 |
+
# You can use Streamlit's text function to display the timing
|
| 181 |
+
st.text(f"Response time: {duration:.2f} seconds")
|
| 182 |
+
|
| 183 |
+
st.session_state['chat_history_page1'].append(("Bot", response, "new"))
|
| 184 |
+
|
| 185 |
+
|
| 186 |
+
# Display new messages at the bottom
|
| 187 |
+
new_messages = st.session_state['chat_history_page1'][-2:]
|
| 188 |
+
for chat in new_messages:
|
| 189 |
+
background_color = "#ffeecf" if chat[2] == "new" else "#ffeecf" if chat[0] == "User" else "#ffeecf"
|
| 190 |
+
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)
|
| 191 |
+
|
| 192 |
+
|
| 193 |
+
# Clear the input field after the query is made
|
| 194 |
+
query = ""
|
| 195 |
+
|
| 196 |
+
# Mark all messages as old after displaying
|
| 197 |
st.session_state['chat_history_page1'] = [(sender, msg, "old") for sender, msg, _ in st.session_state['chat_history_page1']]
|
| 198 |
|
| 199 |
except Exception as e:
|
| 200 |
st.error(f"Upsi, an unexpected error occurred: {e}")
|
| 201 |
+
# Optionally log the exception details to a file or error tracking service
|
| 202 |
+
|
| 203 |
|
| 204 |
|
| 205 |
|
|
|
|
| 312 |
st.error(f"Upsi, an unexpected error occurred: {e}")
|
| 313 |
# Optionally log the exception details to a file or error tracking service
|
| 314 |
|
| 315 |
+
|
| 316 |
+
|
| 317 |
+
|
| 318 |
|
| 319 |
def main():
|
| 320 |
# Sidebar content
|