Update src/streamlit_app.py
Browse files- src/streamlit_app.py +64 -76
src/streamlit_app.py
CHANGED
|
@@ -181,24 +181,31 @@ with st.sidebar:
|
|
| 181 |
st.text_area("", doc_text[:1000], height=120)
|
| 182 |
st.caption(f"{len(st.session_state.get('chunks', []))} chunks processed.")
|
| 183 |
|
| 184 |
-
|
| 185 |
# ==========================================================
|
| 186 |
-
#
|
| 187 |
# ==========================================================
|
| 188 |
-
|
| 189 |
-
"
|
| 190 |
-
|
| 191 |
-
|
| 192 |
-
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
| 201 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 202 |
|
| 203 |
# ==========================================================
|
| 204 |
# π MAIN SECTION
|
|
@@ -209,78 +216,50 @@ st.caption("Query SAP documentation and enterprise PDFs β powered by reasoning
|
|
| 209 |
doc_choice = st.radio("Select a document:", ["-- Select --", "Sample PDF", "Upload Custom PDF"], index=0)
|
| 210 |
|
| 211 |
# ==========================================================
|
| 212 |
-
# π DOCUMENT HANDLING β
|
| 213 |
# ==========================================================
|
| 214 |
if doc_choice == "-- Select --":
|
| 215 |
st.info("β¬
οΈ Select or upload a document to begin.")
|
| 216 |
else:
|
| 217 |
-
temp_path = None
|
| 218 |
-
|
| 219 |
-
# --- File selection ---
|
| 220 |
if doc_choice == "Sample PDF":
|
| 221 |
temp_path = os.path.join(os.path.dirname(__file__), "sample.pdf")
|
| 222 |
-
st.
|
| 223 |
else:
|
| 224 |
-
uploaded_file = st.file_uploader(
|
| 225 |
-
"Upload a PDF document (max 200MB):",
|
| 226 |
-
type="pdf",
|
| 227 |
-
label_visibility="collapsed"
|
| 228 |
-
)
|
| 229 |
if uploaded_file:
|
| 230 |
temp_path = os.path.join("/tmp", uploaded_file.name)
|
| 231 |
with open(temp_path, "wb") as f:
|
| 232 |
f.write(uploaded_file.getbuffer())
|
| 233 |
-
|
| 234 |
-
# File badge (nice compact indicator)
|
| 235 |
-
st.markdown(
|
| 236 |
-
f"<div style='background:#0f172a;border-radius:8px;"
|
| 237 |
-
f"padding:6px 10px;margin-top:4px;display:inline-block;'>π "
|
| 238 |
-
f"<b>{uploaded_file.name}</b> β {uploaded_file.size/1024:.1f} KB</div>",
|
| 239 |
-
unsafe_allow_html=True
|
| 240 |
-
)
|
| 241 |
-
|
| 242 |
else:
|
| 243 |
-
|
| 244 |
|
| 245 |
-
# --- Real processing begins here ---
|
| 246 |
if temp_path:
|
| 247 |
-
|
| 248 |
-
|
| 249 |
-
|
| 250 |
-
text, toc, toc_source = extract_text_from_pdf(temp_path)
|
| 251 |
-
status.info("π Parsing and chunking document...")
|
| 252 |
-
chunks = chunk_text(text, chunk_size=chunk_size, overlap=overlap)
|
| 253 |
-
|
| 254 |
-
status.info("π§ Building embeddings and search index...")
|
| 255 |
-
embeddings = cache_embeddings(os.path.basename(temp_path), chunks, embed_chunks)
|
| 256 |
-
index = build_faiss_index(embeddings)
|
| 257 |
|
| 258 |
-
|
|
|
|
|
|
|
| 259 |
|
| 260 |
-
# Save for dev sidebar
|
| 261 |
-
st.session_state.update({
|
| 262 |
-
"text": text,
|
| 263 |
-
"toc": toc,
|
| 264 |
-
"chunks": chunks
|
| 265 |
-
})
|
| 266 |
-
|
| 267 |
-
# --- Suggestion setup ---
|
| 268 |
doc_name = os.path.basename(temp_path)
|
| 269 |
if st.session_state["last_doc"] != doc_name:
|
| 270 |
query_suggestions = generate_dynamic_suggestions_from_toc(toc, chunks, doc_name)
|
| 271 |
-
st.session_state
|
| 272 |
-
|
| 273 |
-
|
| 274 |
-
|
| 275 |
-
|
| 276 |
-
"show_more": False
|
| 277 |
-
})
|
| 278 |
st.rerun()
|
| 279 |
else:
|
| 280 |
query_suggestions = st.session_state["query_suggestions_fixed"]
|
| 281 |
|
| 282 |
-
#
|
|
|
|
|
|
|
| 283 |
st.markdown("### π¬ Ask the Assistant")
|
|
|
|
| 284 |
if query_suggestions:
|
| 285 |
visible = query_suggestions if st.session_state["show_more"] else query_suggestions[:3]
|
| 286 |
cols = st.columns(min(3, len(visible)))
|
|
@@ -295,12 +274,14 @@ else:
|
|
| 295 |
|
| 296 |
user_query = st.text_input("Type your question or click one above:", key="user_query_input")
|
| 297 |
|
|
|
|
|
|
|
|
|
|
| 298 |
if user_query.strip():
|
| 299 |
reasoning_mode = mode == "Extended (Document + General)"
|
| 300 |
with st.spinner("π Generating your answer..."):
|
| 301 |
retrieved = retrieve_chunks(user_query, index, chunks, top_k=top_k, embeddings=embeddings)
|
| 302 |
answer = generate_answer(user_query, retrieved, reasoning_mode=reasoning_mode)
|
| 303 |
-
st.session_state["retrieved"] = retrieved
|
| 304 |
|
| 305 |
st.markdown("### π€ Assistantβs Answer")
|
| 306 |
|
|
@@ -309,16 +290,23 @@ else:
|
|
| 309 |
answer = re.sub(r"(^|\n)-\s*", r"\1<br>β’ ", answer)
|
| 310 |
st.markdown(f"<div class='answer-box'>{answer}</div>", unsafe_allow_html=True)
|
| 311 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 312 |
|
| 313 |
|
| 314 |
-
# ==========================================================
|
| 315 |
-
# π¨ Optional Sidebar Scroll Styling (keeps it clean)
|
| 316 |
-
# ==========================================================
|
| 317 |
-
st.markdown("""
|
| 318 |
-
<style>
|
| 319 |
-
section[data-testid="stSidebar"] div.stExpander {
|
| 320 |
-
max-height: 480px;
|
| 321 |
-
overflow-y: auto;
|
| 322 |
-
}
|
| 323 |
-
</style>
|
| 324 |
-
""", unsafe_allow_html=True)
|
|
|
|
| 181 |
st.text_area("", doc_text[:1000], height=120)
|
| 182 |
st.caption(f"{len(st.session_state.get('chunks', []))} chunks processed.")
|
| 183 |
|
|
|
|
| 184 |
# ==========================================================
|
| 185 |
+
# π§ SIDEBAR (Old Clean Version)
|
| 186 |
# ==========================================================
|
| 187 |
+
with st.sidebar:
|
| 188 |
+
st.markdown("### π§ Response Style")
|
| 189 |
+
mode = st.radio(
|
| 190 |
+
"",
|
| 191 |
+
("Strict (Document-only)", "Extended (Document + General)"),
|
| 192 |
+
index=0,
|
| 193 |
+
help="Strict = answers only from the uploaded document. Extended = may include related general info.",
|
| 194 |
+
)
|
| 195 |
+
|
| 196 |
+
st.markdown("---")
|
| 197 |
+
show_dev = st.checkbox("Show advanced settings (for developers)", value=False)
|
| 198 |
+
if show_dev:
|
| 199 |
+
st.markdown("### βοΈ Developer Options")
|
| 200 |
+
chunk_size = st.slider("Chunk Size", 200, 1500, 1000, step=50)
|
| 201 |
+
overlap = st.slider("Chunk Overlap", 50, 200, 120, step=10)
|
| 202 |
+
top_k = st.slider("Top K Results", 1, 10, 7)
|
| 203 |
+
else:
|
| 204 |
+
chunk_size, overlap, top_k = 1000, 120, 5
|
| 205 |
+
|
| 206 |
+
st.markdown("---")
|
| 207 |
+
st.caption("β¨ Built by Shubham Sharma")
|
| 208 |
+
|
| 209 |
|
| 210 |
# ==========================================================
|
| 211 |
# π MAIN SECTION
|
|
|
|
| 216 |
doc_choice = st.radio("Select a document:", ["-- Select --", "Sample PDF", "Upload Custom PDF"], index=0)
|
| 217 |
|
| 218 |
# ==========================================================
|
| 219 |
+
# π DOCUMENT HANDLING (Old Layout β Dev Logs Below Answer)
|
| 220 |
# ==========================================================
|
| 221 |
if doc_choice == "-- Select --":
|
| 222 |
st.info("β¬
οΈ Select or upload a document to begin.")
|
| 223 |
else:
|
|
|
|
|
|
|
|
|
|
| 224 |
if doc_choice == "Sample PDF":
|
| 225 |
temp_path = os.path.join(os.path.dirname(__file__), "sample.pdf")
|
| 226 |
+
st.success("π Sample document loaded successfully β you can start asking your questions below.")
|
| 227 |
else:
|
| 228 |
+
uploaded_file = st.file_uploader("", type="pdf", label_visibility="collapsed")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 229 |
if uploaded_file:
|
| 230 |
temp_path = os.path.join("/tmp", uploaded_file.name)
|
| 231 |
with open(temp_path, "wb") as f:
|
| 232 |
f.write(uploaded_file.getbuffer())
|
| 233 |
+
st.success("β
Document processed successfully β you can start asking your questions below.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 234 |
else:
|
| 235 |
+
temp_path = None
|
| 236 |
|
|
|
|
| 237 |
if temp_path:
|
| 238 |
+
with st.spinner("π Processing document..."):
|
| 239 |
+
text, toc, toc_source = extract_text_from_pdf(temp_path)
|
| 240 |
+
chunks = chunk_text(text, chunk_size=chunk_size, overlap=overlap)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 241 |
|
| 242 |
+
with st.spinner("βοΈ Building search index..."):
|
| 243 |
+
embeddings = cache_embeddings(os.path.basename(temp_path), chunks, embed_chunks)
|
| 244 |
+
index = build_faiss_index(embeddings)
|
| 245 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 246 |
doc_name = os.path.basename(temp_path)
|
| 247 |
if st.session_state["last_doc"] != doc_name:
|
| 248 |
query_suggestions = generate_dynamic_suggestions_from_toc(toc, chunks, doc_name)
|
| 249 |
+
st.session_state["query_suggestions_fixed"] = query_suggestions
|
| 250 |
+
st.session_state["last_doc"] = doc_name
|
| 251 |
+
st.session_state["user_query_input"] = ""
|
| 252 |
+
st.session_state["selected_suggestion"] = None
|
| 253 |
+
st.session_state["show_more"] = False
|
|
|
|
|
|
|
| 254 |
st.rerun()
|
| 255 |
else:
|
| 256 |
query_suggestions = st.session_state["query_suggestions_fixed"]
|
| 257 |
|
| 258 |
+
# ----------------------------------------------------------
|
| 259 |
+
# π¬ ASK SECTION
|
| 260 |
+
# ----------------------------------------------------------
|
| 261 |
st.markdown("### π¬ Ask the Assistant")
|
| 262 |
+
|
| 263 |
if query_suggestions:
|
| 264 |
visible = query_suggestions if st.session_state["show_more"] else query_suggestions[:3]
|
| 265 |
cols = st.columns(min(3, len(visible)))
|
|
|
|
| 274 |
|
| 275 |
user_query = st.text_input("Type your question or click one above:", key="user_query_input")
|
| 276 |
|
| 277 |
+
# ----------------------------------------------------------
|
| 278 |
+
# π‘ RESPONSE SECTION
|
| 279 |
+
# ----------------------------------------------------------
|
| 280 |
if user_query.strip():
|
| 281 |
reasoning_mode = mode == "Extended (Document + General)"
|
| 282 |
with st.spinner("π Generating your answer..."):
|
| 283 |
retrieved = retrieve_chunks(user_query, index, chunks, top_k=top_k, embeddings=embeddings)
|
| 284 |
answer = generate_answer(user_query, retrieved, reasoning_mode=reasoning_mode)
|
|
|
|
| 285 |
|
| 286 |
st.markdown("### π€ Assistantβs Answer")
|
| 287 |
|
|
|
|
| 290 |
answer = re.sub(r"(^|\n)-\s*", r"\1<br>β’ ", answer)
|
| 291 |
st.markdown(f"<div class='answer-box'>{answer}</div>", unsafe_allow_html=True)
|
| 292 |
|
| 293 |
+
# ----------------------------------------------------------
|
| 294 |
+
# π§© DEV LOGS BELOW ANSWER (Old placement)
|
| 295 |
+
# ----------------------------------------------------------
|
| 296 |
+
if show_dev:
|
| 297 |
+
st.markdown("---")
|
| 298 |
+
st.markdown("### π§© Developer Insights")
|
| 299 |
+
with st.expander("π Supporting Context", expanded=False):
|
| 300 |
+
for i, r in enumerate(retrieved, start=1):
|
| 301 |
+
st.markdown(f"**Chunk {i}:** {r}")
|
| 302 |
+
|
| 303 |
+
if toc:
|
| 304 |
+
with st.expander("π Explore Document Sections", expanded=False):
|
| 305 |
+
toc_text = "\n".join([f"{sec}. {title}" for sec, title in toc])
|
| 306 |
+
st.text_area("", toc_text, height=140)
|
| 307 |
+
|
| 308 |
+
with st.expander("π Document Preview", expanded=False):
|
| 309 |
+
st.text_area("", text[:1000], height=140)
|
| 310 |
+
st.caption(f"{len(chunks)} chunks processed.")
|
| 311 |
|
| 312 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|