Update src/streamlit_app.py

src/streamlit_app.py

@@ -1,18 +1,20 @@
 import os
 import streamlit as st
 
-# --- Streamlit safe options (prevents upload 403 / CORS issues) ---
+# --- Streamlit Safe Options (Hugging Face Spaces upload fix) ---
 st.set_option("client.showErrorDetails", True)
 
 # ---------------------------
-# Cache Fix for Hugging Face
+# Hugging Face Cache Fix (/tmp for writable)
 # ---------------------------
 CACHE_DIR = "/tmp/hf_cache"
 os.makedirs(CACHE_DIR, exist_ok=True)
-os.environ["HF_HOME"] = CACHE_DIR
-os.environ["TRANSFORMERS_CACHE"] = CACHE_DIR
-os.environ["HF_DATASETS_CACHE"] = CACHE_DIR
-os.environ["HF_MODULES_CACHE"] = CACHE_DIR
+os.environ.update({
+    "HF_HOME": CACHE_DIR,
+    "TRANSFORMERS_CACHE": CACHE_DIR,
+    "HF_DATASETS_CACHE": CACHE_DIR,
+    "HF_MODULES_CACHE": CACHE_DIR
+})
 
 # ---------------------------
 # Imports AFTER environment setup
@@ -34,7 +36,7 @@ SAMPLE_PATH = os.path.join(BASE_DIR, "sample.pdf")
 # ---------------------------
 st.set_page_config(page_title="Enterprise Knowledge Assistant", layout="wide")
 st.title("📄 Enterprise Knowledge Assistant")
-st.caption("Select a document from the library or upload your own, then ask questions.")
+st.caption("Upload a PDF or use the sample file to explore intelligent document Q&A.")
 
 # ---------------------------
 # Sidebar (Library + Settings + Credits)
@@ -43,7 +45,7 @@ with st.sidebar:
     if os.path.exists(LOGO_PATH):
         st.image(LOGO_PATH, width=150)
 
-    # 1. Document Library
+    # 1️⃣ Document Library
     st.header("📚 Document Library")
     doc_choice = st.radio(
         "Choose a document:",
@@ -53,14 +55,14 @@ with st.sidebar:
 
     st.markdown("---")
 
-    # 2. Settings
+    # 2️⃣ Settings
     st.header("⚙️ Settings")
-    chunk_size = st.slider("Chunk Size", 200, 1000, 500, step=100)
-    top_k = st.slider("Top K Results", 1, 5, 3)
+    chunk_size = st.slider("Chunk Size (characters)", 300, 1200, 800, step=100)
+    top_k = st.slider("Top K Results (retrieved chunks)", 1, 10, 5)
 
     st.markdown("---")
 
-    # 3. Branding
+    # 3️⃣ Branding
     st.caption("👨‍💻 Built by Shubham Sharma")
     st.markdown("[📂 GitHub Repo](https://github.com/shubhamsharma170793-cpu/enterprise-knowledge-assistant)")
 
@@ -70,37 +72,39 @@ with st.sidebar:
 text, chunks, index = None, None, None
 
 if doc_choice == "-- Select --":
-    st.info("⬅️ Please choose **Sample PDF** or **Upload Custom PDF** from the sidebar.")
+    st.info("⬅️ Please choose **Sample PDF** or **Upload Custom PDF** from the sidebar to get started.")
 
 elif doc_choice == "Sample PDF":
     temp_path = SAMPLE_PATH
-    st.success("📘 Sample PDF selected")
-    text = extract_text_from_pdf(temp_path)
-    chunks = chunk_text(text, chunk_size=chunk_size)
-    embeddings = generate_embeddings(chunks)
-    index = build_faiss_index(embeddings)
+    st.success("📘 Using built-in Sample PDF")
+    with st.spinner("🔍 Extracting and processing document..."):
+        text = extract_text_from_pdf(temp_path)
+        chunks = chunk_text(text, chunk_size=chunk_size)
+        embeddings = generate_embeddings(chunks)
+        index = build_faiss_index(embeddings)
 
 elif doc_choice == "Upload Custom PDF":
     uploaded_file = st.file_uploader("📂 Upload your PDF", type="pdf")
     if uploaded_file:
-        # Always write to /tmp (the only guaranteed writable folder)
         temp_path = os.path.join("/tmp", uploaded_file.name)
         with open(temp_path, "wb") as f:
             f.write(uploaded_file.getbuffer())
-        st.success("✅ Document uploaded and processed!")
+        st.success(f"✅ File '{uploaded_file.name}' uploaded and saved to /tmp")
 
-        text = extract_text_from_pdf(temp_path)
-        chunks = chunk_text(text, chunk_size=chunk_size)
-        embeddings = generate_embeddings(chunks)
-        index = build_faiss_index(embeddings)
+        with st.spinner("⚙️ Extracting and processing your document..."):
+            text = extract_text_from_pdf(temp_path)
+            chunks = chunk_text(text, chunk_size=chunk_size)
+            embeddings = generate_embeddings(chunks)
+            index = build_faiss_index(embeddings)
+        st.success("🚀 Document processed successfully!")
 
 # ---------------------------
 # Document Preview
 # ---------------------------
 if chunks:
     st.subheader("📑 Document Preview")
-    st.text_area("Extracted text (first 1000 chars)", text[:1000], height=150)
-    st.caption(f"📦 {len(chunks)} chunks created")
+    st.text_area("Extracted text (first 1000 chars)", text[:1000], height=200)
+    st.caption(f"📦 {len(chunks)} chunks created | Avg chunk length: {int(sum(len(c) for c in chunks) / len(chunks))} chars")
 
 # ---------------------------
 # Query Section
@@ -111,12 +115,24 @@ if index and chunks:
 
     user_query = st.text_input("🔍 Your question about the document:")
     if user_query:
-        retrieved = retrieve_chunks(user_query, index, chunks, top_k=top_k)
-        answer = generate_answer(user_query, retrieved)
+        with st.spinner("🧠 Thinking... retrieving context and generating answer..."):
+            retrieved = retrieve_chunks(user_query, index, chunks, top_k=top_k)
+            answer = generate_answer(user_query, retrieved)
 
+        # Answer Section
         st.markdown("### ✅ Assistant’s Answer")
-        st.write(answer)
+        st.markdown(f"<div style='background-color:#0E1117;padding:12px;border-radius:10px;'>{answer}</div>", unsafe_allow_html=True)
 
-        with st.expander("📄 Supporting Chunks"):
+        # Supporting Chunks Section
+        with st.expander("📄 Supporting Chunks (Context Used)"):
             for i, r in enumerate(retrieved, start=1):
-                st.markdown(f"**Chunk {i}:** {r}")
+                st.markdown(
+                    f"""
+                    <div style='background-color:#111827;padding:10px;border-radius:8px;margin-bottom:6px;'>
+                    <b>Chunk {i}:</b><br>{r}
+                    </div>
+                    """,
+                    unsafe_allow_html=True,
+                )
+else:
+    st.info("📥 Upload or select a document to start exploring.")