Spaces:
Sleeping
Sleeping
Update pages/1_π_Document_Analysis.py
Browse files- pages/1_π_Document_Analysis.py +238 -262
pages/1_π_Document_Analysis.py
CHANGED
|
@@ -1,263 +1,239 @@
|
|
| 1 |
-
import streamlit as st
|
| 2 |
-
from PIL import Image
|
| 3 |
-
import datetime
|
| 4 |
-
from theme import apply_dark_theme, show_page_header, show_footer
|
| 5 |
-
from utils import (
|
| 6 |
-
process_image,
|
| 7 |
-
process_pdf,
|
| 8 |
-
initialize_session_state,
|
| 9 |
-
create_chat_engine,
|
| 10 |
-
save_to_history,
|
| 11 |
-
generate_pdf_analysis,
|
| 12 |
-
process_captured_image,
|
| 13 |
-
format_analysis_results
|
| 14 |
-
)
|
| 15 |
-
|
| 16 |
-
# Page config
|
| 17 |
-
st.set_page_config(
|
| 18 |
-
page_title="Document Analysis | CiviDoc AI",
|
| 19 |
-
page_icon="π",
|
| 20 |
-
layout="centered",
|
| 21 |
-
initial_sidebar_state="collapsed",
|
| 22 |
-
)
|
| 23 |
-
|
| 24 |
-
# Apply dark theme
|
| 25 |
-
st.markdown(apply_dark_theme(), unsafe_allow_html=True)
|
| 26 |
-
|
| 27 |
-
def document_analysis_page():
|
| 28 |
-
# Initialize states
|
| 29 |
-
initialize_session_state()
|
| 30 |
-
|
| 31 |
-
# Header
|
| 32 |
-
st.markdown(show_page_header(
|
| 33 |
-
"π Document Analysis",
|
| 34 |
-
"Upload or capture documents for instant analysis"
|
| 35 |
-
), unsafe_allow_html=True)
|
| 36 |
-
|
| 37 |
-
# Main content tabs - Mobile friendly
|
| 38 |
-
tabs = st.tabs([
|
| 39 |
-
"π€ Upload ",
|
| 40 |
-
"πΈ Capture",
|
| 41 |
-
"π Results"
|
| 42 |
-
])
|
| 43 |
-
|
| 44 |
-
with tabs[0]:
|
| 45 |
-
st.markdown(
|
| 46 |
-
"<div class='card'>"
|
| 47 |
-
"<h3>Upload Documents</h3>"
|
| 48 |
-
"<p style='margin-bottom: 2rem;'>Support for PDF, JPG, JPEG, and PNG files.</p>"
|
| 49 |
-
"<div class='status-badge status-success'>Multiple files supported</div>"
|
| 50 |
-
"</div>",
|
| 51 |
-
unsafe_allow_html=True
|
| 52 |
-
)
|
| 53 |
-
|
| 54 |
-
# Mobile-friendly file uploader
|
| 55 |
-
uploaded_files = st.file_uploader(
|
| 56 |
-
"Drop files or tap to browse",
|
| 57 |
-
type=["jpg", "jpeg", "png", "pdf"],
|
| 58 |
-
accept_multiple_files=True,
|
| 59 |
-
key="doc_uploader"
|
| 60 |
-
)
|
| 61 |
-
|
| 62 |
-
if uploaded_files:
|
| 63 |
-
st.markdown(
|
| 64 |
-
f"<div class='status-badge status-success' style='margin: 1rem 0;'>"
|
| 65 |
-
f"π {len(uploaded_files)} file(s) uploaded"
|
| 66 |
-
f"</div>",
|
| 67 |
-
unsafe_allow_html=True
|
| 68 |
-
)
|
| 69 |
-
|
| 70 |
-
# File list - Mobile friendly
|
| 71 |
-
st.markdown(
|
| 72 |
-
"<div class='card'>"
|
| 73 |
-
"<h4>Selected Files:</h4>"
|
| 74 |
-
"<div class='touch-spacing'>",
|
| 75 |
-
unsafe_allow_html=True
|
| 76 |
-
)
|
| 77 |
-
|
| 78 |
-
for file in uploaded_files:
|
| 79 |
-
st.markdown(
|
| 80 |
-
f"<div style='display: flex; align-items: center; padding: 0.5rem 0;'>"
|
| 81 |
-
f"<span style='margin-right: 0.5rem;'>π</span>{file.name}"
|
| 82 |
-
f"</div>",
|
| 83 |
-
unsafe_allow_html=True
|
| 84 |
-
)
|
| 85 |
-
|
| 86 |
-
st.markdown("</div></div>", unsafe_allow_html=True)
|
| 87 |
-
|
| 88 |
-
# Analysis button - Touch friendly
|
| 89 |
-
if st.button("π Analyze Documents", use_container_width=True):
|
| 90 |
-
process_uploaded_files(uploaded_files)
|
| 91 |
-
|
| 92 |
-
with tabs[1]:
|
| 93 |
-
st.markdown(
|
| 94 |
-
"<div class='card'>"
|
| 95 |
-
"<h3>Capture Document</h3>"
|
| 96 |
-
"<p>Take a clear photo of your document using your camera.</p>"
|
| 97 |
-
"</div>",
|
| 98 |
-
unsafe_allow_html=True
|
| 99 |
-
)
|
| 100 |
-
|
| 101 |
-
# Mobile-optimized camera input
|
| 102 |
-
picture = st.camera_input(
|
| 103 |
-
"πΈ Tap to capture",
|
| 104 |
-
help="Make sure the document is well-lit and clearly visible"
|
| 105 |
-
)
|
| 106 |
-
|
| 107 |
-
if picture:
|
| 108 |
-
st.markdown(
|
| 109 |
-
"<div class='status-badge status-success' style='margin: 1rem 0;'>"
|
| 110 |
-
"πΈ Image captured successfully"
|
| 111 |
-
"</div>",
|
| 112 |
-
unsafe_allow_html=True
|
| 113 |
-
)
|
| 114 |
-
|
| 115 |
-
# Analysis button - Touch friendly
|
| 116 |
-
if st.button("π Analyze Photo", use_container_width=True):
|
| 117 |
-
process_captured_image(picture)
|
| 118 |
-
|
| 119 |
-
with tabs[2]:
|
| 120 |
-
display_analysis_results()
|
| 121 |
-
|
| 122 |
-
def process_uploaded_files(files):
|
| 123 |
-
"""Process multiple uploaded files with mobile-friendly progress tracking"""
|
| 124 |
-
progress_text = "Processing documents..."
|
| 125 |
-
total_files = len(files)
|
| 126 |
-
|
| 127 |
-
# Progress tracking
|
| 128 |
-
progress_placeholder = st.empty()
|
| 129 |
-
progress_bar = st.progress(0)
|
| 130 |
-
status_text = st.empty()
|
| 131 |
-
|
| 132 |
-
for idx, uploaded_file in enumerate(files, 1):
|
| 133 |
-
try:
|
| 134 |
-
# Update progress
|
| 135 |
-
progress_placeholder.markdown(
|
| 136 |
-
f"<div style='text-align: center; margin: 1rem 0;'>"
|
| 137 |
-
f"Processing file {idx} of {total_files}"
|
| 138 |
-
f"</div>",
|
| 139 |
-
unsafe_allow_html=True
|
| 140 |
-
)
|
| 141 |
-
|
| 142 |
-
status_text.markdown(
|
| 143 |
-
f"<div class='status-badge status-warning'>"
|
| 144 |
-
f"π Processing: {uploaded_file.name}"
|
| 145 |
-
f"</div>",
|
| 146 |
-
unsafe_allow_html=True
|
| 147 |
-
)
|
| 148 |
-
|
| 149 |
-
if uploaded_file.type in ['image/jpeg', 'image/png']:
|
| 150 |
-
# Process image
|
| 151 |
-
image = Image.open(uploaded_file)
|
| 152 |
-
analysis = process_image(image)
|
| 153 |
-
|
| 154 |
-
# Save results
|
| 155 |
-
st.session_state.analyses[uploaded_file.name] = {
|
| 156 |
-
'type': uploaded_file.type,
|
| 157 |
-
'analysis': analysis,
|
| 158 |
-
'timestamp': datetime.datetime.now()
|
| 159 |
-
}
|
| 160 |
-
|
| 161 |
-
# Create chat engine
|
| 162 |
-
st.session_state.chat_engines[uploaded_file.name] = create_chat_engine(analysis)
|
| 163 |
-
|
| 164 |
-
elif uploaded_file.type == 'application/pdf':
|
| 165 |
-
# Process PDF
|
| 166 |
-
documents = process_pdf(uploaded_file)
|
| 167 |
-
analysis = generate_pdf_analysis(documents)
|
| 168 |
-
|
| 169 |
-
# Save results
|
| 170 |
-
st.session_state.analyses[uploaded_file.name] = {
|
| 171 |
-
'type': uploaded_file.type,
|
| 172 |
-
'analysis': analysis,
|
| 173 |
-
'timestamp': datetime.datetime.now()
|
| 174 |
-
}
|
| 175 |
-
|
| 176 |
-
# Create chat engine
|
| 177 |
-
st.session_state.chat_engines[uploaded_file.name] = create_chat_engine(documents)
|
| 178 |
-
|
| 179 |
-
# Update progress
|
| 180 |
-
progress_bar.progress(idx/total_files)
|
| 181 |
-
|
| 182 |
-
# Save to history
|
| 183 |
-
save_to_history(
|
| 184 |
-
uploaded_file.name,
|
| 185 |
-
uploaded_file.type.split('/')[1].upper(),
|
| 186 |
-
analysis,
|
| 187 |
-
datetime.datetime.now()
|
| 188 |
-
)
|
| 189 |
-
|
| 190 |
-
except Exception as e:
|
| 191 |
-
st.error(
|
| 192 |
-
f"β Error processing {uploaded_file.name}\n"
|
| 193 |
-
f"Details: {str(e)}"
|
| 194 |
-
)
|
| 195 |
-
|
| 196 |
-
# Clear progress indicators
|
| 197 |
-
progress_placeholder.empty()
|
| 198 |
-
progress_bar.empty()
|
| 199 |
-
|
| 200 |
-
# Show completion message
|
| 201 |
-
status_text.markdown(
|
| 202 |
-
"<div class='status-badge status-success' style='margin: 1rem 0;'>"
|
| 203 |
-
"β
All documents processed successfully!"
|
| 204 |
-
"</div>",
|
| 205 |
-
unsafe_allow_html=True
|
| 206 |
-
)
|
| 207 |
-
|
| 208 |
-
|
| 209 |
-
|
| 210 |
-
if st.
|
| 211 |
-
|
| 212 |
-
|
| 213 |
-
st.
|
| 214 |
-
|
| 215 |
-
|
| 216 |
-
|
| 217 |
-
|
| 218 |
-
|
| 219 |
-
|
| 220 |
-
|
| 221 |
-
|
| 222 |
-
|
| 223 |
-
|
| 224 |
-
|
| 225 |
-
|
| 226 |
-
|
| 227 |
-
|
| 228 |
-
|
| 229 |
-
|
| 230 |
-
|
| 231 |
-
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
|
| 235 |
-
|
| 236 |
-
|
| 237 |
-
|
| 238 |
-
|
| 239 |
-
st.markdown(
|
| 240 |
-
f"<div class='card'>"
|
| 241 |
-
f"<div style='display: flex; justify-content: space-between; align-items: center;'>"
|
| 242 |
-
f"<h4>{filename}</h4>"
|
| 243 |
-
f"<div class='status-badge status-success'>{data['type'].split('/')[1].upper()}</div>"
|
| 244 |
-
f"</div>"
|
| 245 |
-
f"<hr style='margin: 0.5rem 0;'>"
|
| 246 |
-
f"{format_analysis_results(data['analysis'])}" # Use the formatting function
|
| 247 |
-
f"<div class='touch-spacing'>"
|
| 248 |
-
f"</div>"
|
| 249 |
-
f"</div>",
|
| 250 |
-
unsafe_allow_html=True
|
| 251 |
-
)
|
| 252 |
-
else:
|
| 253 |
-
st.markdown(
|
| 254 |
-
"<div class='card' style='text-align: center;'>"
|
| 255 |
-
"<h3>No Documents Analyzed</h3>"
|
| 256 |
-
"<p>Upload or capture documents to see analysis results here.</p>"
|
| 257 |
-
"</div>",
|
| 258 |
-
unsafe_allow_html=True
|
| 259 |
-
)
|
| 260 |
-
|
| 261 |
-
|
| 262 |
-
if __name__ == "__main__":
|
| 263 |
document_analysis_page()
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
from PIL import Image
|
| 3 |
+
import datetime
|
| 4 |
+
from theme import apply_dark_theme, show_page_header, show_footer
|
| 5 |
+
from utils import (
|
| 6 |
+
process_image,
|
| 7 |
+
process_pdf,
|
| 8 |
+
initialize_session_state,
|
| 9 |
+
create_chat_engine,
|
| 10 |
+
save_to_history,
|
| 11 |
+
generate_pdf_analysis,
|
| 12 |
+
process_captured_image,
|
| 13 |
+
format_analysis_results
|
| 14 |
+
)
|
| 15 |
+
|
| 16 |
+
# Page config
|
| 17 |
+
st.set_page_config(
|
| 18 |
+
page_title="Document Analysis | CiviDoc AI",
|
| 19 |
+
page_icon="π",
|
| 20 |
+
layout="centered",
|
| 21 |
+
initial_sidebar_state="collapsed",
|
| 22 |
+
)
|
| 23 |
+
|
| 24 |
+
# Apply dark theme
|
| 25 |
+
st.markdown(apply_dark_theme(), unsafe_allow_html=True)
|
| 26 |
+
|
| 27 |
+
def document_analysis_page():
|
| 28 |
+
# Initialize states
|
| 29 |
+
initialize_session_state()
|
| 30 |
+
|
| 31 |
+
# Header
|
| 32 |
+
st.markdown(show_page_header(
|
| 33 |
+
"π Document Analysis",
|
| 34 |
+
"Upload or capture documents for instant analysis"
|
| 35 |
+
), unsafe_allow_html=True)
|
| 36 |
+
|
| 37 |
+
# Main content tabs - Mobile friendly
|
| 38 |
+
tabs = st.tabs([
|
| 39 |
+
"π€ Upload ",
|
| 40 |
+
"πΈ Capture",
|
| 41 |
+
"π Results"
|
| 42 |
+
])
|
| 43 |
+
|
| 44 |
+
with tabs[0]:
|
| 45 |
+
st.markdown(
|
| 46 |
+
"<div class='card'>"
|
| 47 |
+
"<h3>Upload Documents</h3>"
|
| 48 |
+
"<p style='margin-bottom: 2rem;'>Support for PDF, JPG, JPEG, and PNG files.</p>"
|
| 49 |
+
"<div class='status-badge status-success'>Multiple files supported</div>"
|
| 50 |
+
"</div>",
|
| 51 |
+
unsafe_allow_html=True
|
| 52 |
+
)
|
| 53 |
+
|
| 54 |
+
# Mobile-friendly file uploader
|
| 55 |
+
uploaded_files = st.file_uploader(
|
| 56 |
+
"Drop files or tap to browse",
|
| 57 |
+
type=["jpg", "jpeg", "png", "pdf"],
|
| 58 |
+
accept_multiple_files=True,
|
| 59 |
+
key="doc_uploader"
|
| 60 |
+
)
|
| 61 |
+
|
| 62 |
+
if uploaded_files:
|
| 63 |
+
st.markdown(
|
| 64 |
+
f"<div class='status-badge status-success' style='margin: 1rem 0;'>"
|
| 65 |
+
f"π {len(uploaded_files)} file(s) uploaded"
|
| 66 |
+
f"</div>",
|
| 67 |
+
unsafe_allow_html=True
|
| 68 |
+
)
|
| 69 |
+
|
| 70 |
+
# File list - Mobile friendly
|
| 71 |
+
st.markdown(
|
| 72 |
+
"<div class='card'>"
|
| 73 |
+
"<h4>Selected Files:</h4>"
|
| 74 |
+
"<div class='touch-spacing'>",
|
| 75 |
+
unsafe_allow_html=True
|
| 76 |
+
)
|
| 77 |
+
|
| 78 |
+
for file in uploaded_files:
|
| 79 |
+
st.markdown(
|
| 80 |
+
f"<div style='display: flex; align-items: center; padding: 0.5rem 0;'>"
|
| 81 |
+
f"<span style='margin-right: 0.5rem;'>π</span>{file.name}"
|
| 82 |
+
f"</div>",
|
| 83 |
+
unsafe_allow_html=True
|
| 84 |
+
)
|
| 85 |
+
|
| 86 |
+
st.markdown("</div></div>", unsafe_allow_html=True)
|
| 87 |
+
|
| 88 |
+
# Analysis button - Touch friendly
|
| 89 |
+
if st.button("π Analyze Documents", use_container_width=True):
|
| 90 |
+
process_uploaded_files(uploaded_files)
|
| 91 |
+
|
| 92 |
+
with tabs[1]:
|
| 93 |
+
st.markdown(
|
| 94 |
+
"<div class='card'>"
|
| 95 |
+
"<h3>Capture Document</h3>"
|
| 96 |
+
"<p>Take a clear photo of your document using your camera.</p>"
|
| 97 |
+
"</div>",
|
| 98 |
+
unsafe_allow_html=True
|
| 99 |
+
)
|
| 100 |
+
|
| 101 |
+
# Mobile-optimized camera input
|
| 102 |
+
picture = st.camera_input(
|
| 103 |
+
"πΈ Tap to capture",
|
| 104 |
+
help="Make sure the document is well-lit and clearly visible"
|
| 105 |
+
)
|
| 106 |
+
|
| 107 |
+
if picture:
|
| 108 |
+
st.markdown(
|
| 109 |
+
"<div class='status-badge status-success' style='margin: 1rem 0;'>"
|
| 110 |
+
"πΈ Image captured successfully"
|
| 111 |
+
"</div>",
|
| 112 |
+
unsafe_allow_html=True
|
| 113 |
+
)
|
| 114 |
+
|
| 115 |
+
# Analysis button - Touch friendly
|
| 116 |
+
if st.button("π Analyze Photo", use_container_width=True):
|
| 117 |
+
process_captured_image(picture)
|
| 118 |
+
|
| 119 |
+
with tabs[2]:
|
| 120 |
+
display_analysis_results()
|
| 121 |
+
|
| 122 |
+
def process_uploaded_files(files):
|
| 123 |
+
"""Process multiple uploaded files with mobile-friendly progress tracking"""
|
| 124 |
+
progress_text = "Processing documents..."
|
| 125 |
+
total_files = len(files)
|
| 126 |
+
|
| 127 |
+
# Progress tracking
|
| 128 |
+
progress_placeholder = st.empty()
|
| 129 |
+
progress_bar = st.progress(0)
|
| 130 |
+
status_text = st.empty()
|
| 131 |
+
|
| 132 |
+
for idx, uploaded_file in enumerate(files, 1):
|
| 133 |
+
try:
|
| 134 |
+
# Update progress
|
| 135 |
+
progress_placeholder.markdown(
|
| 136 |
+
f"<div style='text-align: center; margin: 1rem 0;'>"
|
| 137 |
+
f"Processing file {idx} of {total_files}"
|
| 138 |
+
f"</div>",
|
| 139 |
+
unsafe_allow_html=True
|
| 140 |
+
)
|
| 141 |
+
|
| 142 |
+
status_text.markdown(
|
| 143 |
+
f"<div class='status-badge status-warning'>"
|
| 144 |
+
f"π Processing: {uploaded_file.name}"
|
| 145 |
+
f"</div>",
|
| 146 |
+
unsafe_allow_html=True
|
| 147 |
+
)
|
| 148 |
+
|
| 149 |
+
if uploaded_file.type in ['image/jpeg', 'image/png']:
|
| 150 |
+
# Process image
|
| 151 |
+
image = Image.open(uploaded_file)
|
| 152 |
+
analysis = process_image(image)
|
| 153 |
+
|
| 154 |
+
# Save results
|
| 155 |
+
st.session_state.analyses[uploaded_file.name] = {
|
| 156 |
+
'type': uploaded_file.type,
|
| 157 |
+
'analysis': analysis,
|
| 158 |
+
'timestamp': datetime.datetime.now()
|
| 159 |
+
}
|
| 160 |
+
|
| 161 |
+
# Create chat engine
|
| 162 |
+
st.session_state.chat_engines[uploaded_file.name] = create_chat_engine(analysis)
|
| 163 |
+
|
| 164 |
+
elif uploaded_file.type == 'application/pdf':
|
| 165 |
+
# Process PDF
|
| 166 |
+
documents = process_pdf(uploaded_file)
|
| 167 |
+
analysis = generate_pdf_analysis(documents)
|
| 168 |
+
|
| 169 |
+
# Save results
|
| 170 |
+
st.session_state.analyses[uploaded_file.name] = {
|
| 171 |
+
'type': uploaded_file.type,
|
| 172 |
+
'analysis': analysis,
|
| 173 |
+
'timestamp': datetime.datetime.now()
|
| 174 |
+
}
|
| 175 |
+
|
| 176 |
+
# Create chat engine
|
| 177 |
+
st.session_state.chat_engines[uploaded_file.name] = create_chat_engine(documents)
|
| 178 |
+
|
| 179 |
+
# Update progress
|
| 180 |
+
progress_bar.progress(idx/total_files)
|
| 181 |
+
|
| 182 |
+
# Save to history
|
| 183 |
+
save_to_history(
|
| 184 |
+
uploaded_file.name,
|
| 185 |
+
uploaded_file.type.split('/')[1].upper(),
|
| 186 |
+
analysis,
|
| 187 |
+
datetime.datetime.now()
|
| 188 |
+
)
|
| 189 |
+
|
| 190 |
+
except Exception as e:
|
| 191 |
+
st.error(
|
| 192 |
+
f"β Error processing {uploaded_file.name}\n"
|
| 193 |
+
f"Details: {str(e)}"
|
| 194 |
+
)
|
| 195 |
+
|
| 196 |
+
# Clear progress indicators
|
| 197 |
+
progress_placeholder.empty()
|
| 198 |
+
progress_bar.empty()
|
| 199 |
+
|
| 200 |
+
# Show completion message
|
| 201 |
+
status_text.markdown(
|
| 202 |
+
"<div class='status-badge status-success' style='margin: 1rem 0;'>"
|
| 203 |
+
"β
All documents processed successfully!"
|
| 204 |
+
"</div>",
|
| 205 |
+
unsafe_allow_html=True
|
| 206 |
+
)
|
| 207 |
+
|
| 208 |
+
def display_analysis_results():
|
| 209 |
+
"""Display analysis results with mobile-friendly layout"""
|
| 210 |
+
if st.session_state.analyses:
|
| 211 |
+
|
| 212 |
+
# Display results
|
| 213 |
+
if st.session_state.analyses:
|
| 214 |
+
for filename, data in st.session_state.analyses.items():
|
| 215 |
+
st.markdown(
|
| 216 |
+
f"<div class='card'>"
|
| 217 |
+
f"<div style='display: flex; justify-content: space-between; align-items: center;'>"
|
| 218 |
+
f"<h4>{filename}</h4>"
|
| 219 |
+
f"<div class='status-badge status-success'>{data['type'].split('/')[1].upper()}</div>"
|
| 220 |
+
f"</div>"
|
| 221 |
+
f"<hr style='margin: 0.5rem 0;'>"
|
| 222 |
+
f"{format_analysis_results(data['analysis'])}" # Use the formatting function
|
| 223 |
+
f"<div class='touch-spacing'>"
|
| 224 |
+
f"</div>"
|
| 225 |
+
f"</div>",
|
| 226 |
+
unsafe_allow_html=True
|
| 227 |
+
)
|
| 228 |
+
else:
|
| 229 |
+
st.markdown(
|
| 230 |
+
"<div class='card' style='text-align: center;'>"
|
| 231 |
+
"<h3>No Documents Analyzed</h3>"
|
| 232 |
+
"<p>Upload or capture documents to see analysis results here.</p>"
|
| 233 |
+
"</div>",
|
| 234 |
+
unsafe_allow_html=True
|
| 235 |
+
)
|
| 236 |
+
|
| 237 |
+
|
| 238 |
+
if __name__ == "__main__":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 239 |
document_analysis_page()
|