Update app.py

app.py

@@ -5,12 +5,10 @@ import os
 from pypdf import PdfReader
 
 # ── CONFIG ───────────────────────────────────────────────────────────────────
-OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")  # ← fixed: use environment variable
-# If testing locally, you can temporarily do:
-# OPENAI_API_KEY = "sk-......................."
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
 
 if not OPENAI_API_KEY:
-    raise ValueError("OPENAI_API_KEY environment variable not set")
+    raise ValueError("Please set the OPENAI_API_KEY environment variable")
 
 client = openai.OpenAI(api_key=OPENAI_API_KEY)
 
@@ -27,67 +25,90 @@ doc_name          = ""
 
 # ── CORE LOGIC ───────────────────────────────────────────────────────────────
 def split_chunks(text, size=CHUNK_SIZE, overlap=OVERLAP):
-    words, chunks, start = text.split(), [], 0
+    words = text.split()
+    chunks = []
+    start = 0
     while start < len(words):
         chunks.append(" ".join(words[start:start + size]))
         start += size - overlap
     return chunks
 
 def embed(text):
-    return client.embeddings.create(model=EMBED_MODEL, input=text).data[0].embedding
+    return client.embeddings.create(
+        model=EMBED_MODEL,
+        input=text
+    ).data[0].embedding
 
 def cosine_similarity(a, b):
-    dot   = sum(x * y for x, y in zip(a, b))
-    normA = math.sqrt(sum(x * x for x in a))
-    normB = math.sqrt(sum(x * x for x in b))
-    return dot / (normA * normB) if normA * normB != 0 else 0
+    dot = sum(x * y for x, y in zip(a, b))
+    norm_a = math.sqrt(sum(x * x for x in a))
+    norm_b = math.sqrt(sum(x * x for x in b))
+    return dot / (norm_a * norm_b) if norm_a * norm_b != 0 else 0.0
 
 def get_top_chunks(question, k=TOP_K):
-    q_vec  = embed(question)
-    scores = [(cosine_similarity(q_vec, cv), c)
-              for cv, c in zip(stored_embeddings, stored_chunks)]
-    scores.sort(reverse=True)
+    if not stored_embeddings:
+        return []
+    q_vec = embed(question)
+    scores = [
+        (cosine_similarity(q_vec, cv), c)
+        for cv, c in zip(stored_embeddings, stored_chunks)
+    ]
+    scores.sort(reverse=True, key=lambda x: x[0])
     return [c for _, c in scores[:k]]
 
 def upload_pdf(file):
     global stored_chunks, stored_embeddings, doc_name
     if file is None:
-        return ("— No file selected.", gr.update(interactive=False))
+        return "— No file selected.", gr.update(interactive=False)
 
-    reader    = PdfReader(file.name)
-    full_text = "".join(page.extract_text() or "" for page in reader.pages)
-    doc_name  = os.path.basename(file.name)
+    try:
+        reader = PdfReader(file.name)
+        full_text = "".join(page.extract_text() or "" for page in reader.pages)
+        doc_name = os.path.basename(file.name)
 
-    stored_chunks     = split_chunks(full_text)
-    stored_embeddings = [embed(c) for c in stored_chunks]
+        stored_chunks = split_chunks(full_text)
+        stored_embeddings = [embed(c) for c in stored_chunks]
 
-    return (
-        f"✦  {doc_name}\n→  {len(stored_chunks)} chunks indexed and ready.",
-        gr.update(interactive=True)
-    )
+        return (
+            f"✦ {doc_name}\n→ {len(stored_chunks)} chunks indexed and ready.",
+            gr.update(interactive=True)
+        )
+    except Exception as e:
+        return f"Error processing PDF: {str(e)}", gr.update(interactive=False)
 
 def chat(question, history):
     if not stored_chunks:
-        return history + [(question, "Please index a document in the Upload tab first.")]
-
-    context = "\n\n".join([f"[{i+1}] {c}" for i, c in enumerate(get_top_chunks(question))])
-    prompt  = (
-        "Answer strictly from the document text below.\n"
-        "If not found, reply: \"This information is not in the document.\"\n\n"
-        f"TEXT:\n{context}\n\nQUESTION: {question}\n\nANSWER:"
+        return history + [(question, "Please upload and index a document first in the Upload tab.")]
+
+    context_chunks = get_top_chunks(question)
+    if not context_chunks:
+        return history + [(question, "No relevant context found.")]
+
+    context = "\n\n".join([f"[{i+1}] {c}" for i, c in enumerate(context_chunks)])
+    prompt = (
+        "You are a strict document Q&A assistant. "
+        "Answer only using the provided document text below. "
+        "If the information is not present, reply exactly: "
+        "\"This information is not in the document.\"\n\n"
+        f"TEXT:\n{context}\n\n"
+        f"QUESTION: {question}\n\n"
+        "ANSWER:"
     )
+
     response = client.chat.completions.create(
         model=CHAT_MODEL,
         messages=[
-            {"role": "system", "content":
-                "Strict document assistant. Only use provided text. "
-                "If answer absent, say: This information is not in the document."},
+            {"role": "system", "content": "Follow the instructions precisely."},
             {"role": "user", "content": prompt}
-        ]
+        ],
+        temperature=0.1,
+        max_tokens=800
     )
-    return history + [(question, response.choices[0].message.content)]
 
-# ── CSS (cleaner, higher contrast, Apple-like feel) ──────────────────────────
+    answer = response.choices[0].message.content.strip()
+    return history + [(question, answer)]
+
+# ── CSS (clean, high-contrast, Apple-like) ──────────────────────────────────
 css = """
 :root {
     --bg:           #fdfdfd;
@@ -95,39 +116,33 @@ css = """
     --surface-2:    #f5f5f7;
     --border:       #d2d2d7;
     --border-2:     #c7c7cc;
-    --accent:       #0071e3;       /* Apple blue */
+    --accent:       #0071e3;
     --accent-light: #4090ff;
-    --accent-glow:  rgba(0, 113, 227, 0.12);
-    --green:        #34c759;
+    --accent-glow:  rgba(0, 113, 227, 0.14);
     --text-1:       #1d1d1f;
     --text-2:       #3c3c43;
     --text-3:       #6e6e73;
-    --font-sans:    system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
-    --font-mono:    ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, monospace;
-    --r:            12px;
-    --r-lg:         20px;
-    --ease:         cubic-bezier(0.25, 0.1, 0.25, 1);
-    --shadow:       0 1px 3px rgba(0,0,0,0.08), 0 4px 12px rgba(0,0,0,0.05);
+    --radius:       12px;
+    --radius-lg:    20px;
+    --shadow-sm:    0 1px 3px rgba(0,0,0,0.08);
+    --shadow:       0 4px 12px rgba(0,0,0,0.06);
 }
 
-html, body, .gradio-container {
-    font-family: var(--font-sans) !important;
-    background: var(--bg) !important;
-    color: var(--text-1) !important;
+body, .gradio-container {
+    font-family: system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
+    background: var(--bg);
+    color: var(--text-1);
 }
 
 .gradio-container {
-    max-width: 1100px !important;
-    margin: 0 auto !important;
-    padding: 32px 24px 80px !important;
+    max-width: 1100px;
+    margin: 0 auto;
+    padding: 32px 24px 80px;
 }
 
 ::-webkit-scrollbar {
     width: 6px;
 }
-::-webkit-scrollbar-track {
-    background: transparent;
-}
 ::-webkit-scrollbar-thumb {
     background: #c0c0c5;
     border-radius: 3px;
@@ -136,168 +151,125 @@ html, body, .gradio-container {
     background: #a0a0a8;
 }
 
-/* Masthead */
-.mast-title {
-    font-size: 2.4rem;
-    font-weight: 600;
-    letter-spacing: -0.02em;
-    color: var(--text-1);
-}
-.mast-title em {
-    color: var(--accent);
-    font-style: normal;
-    font-weight: 700;
-}
-.mast-sub {
-    font-family: var(--font-mono);
-    font-size: 0.68rem;
-    color: var(--text-3);
-    letter-spacing: 0.06em;
-    text-transform: uppercase;
-    margin-top: 6px;
-}
-
-/* Chatbot fixes – most important for readability */
-.chatbot-wrap > div {
-    background: var(--bg) !important;
+/* Chat */
+.chatbot {
     border: 1px solid var(--border) !important;
-    border-radius: var(--r-lg) !important;
+    border-radius: var(--radius-lg) !important;
+    background: var(--bg) !important;
+    overflow: hidden;
 }
-.message-wrap .user-row .message {
+
+.message.user {
     background: var(--accent) !important;
     color: white !important;
     border-radius: 20px 20px 4px 20px !important;
+    margin: 8px 12px 8px auto !important;
+    max-width: 78% !important;
     padding: 12px 16px !important;
-    font-size: 0.94rem !important;
-    line-height: 1.5 !important;
+    line-height: 1.48;
 }
-.message-wrap .bot-row .message {
+
+.message.bot, .message {
     background: var(--surface) !important;
     color: var(--text-1) !important;
     border: 1px solid var(--border-2) !important;
     border-radius: 20px 20px 20px 4px !important;
+    margin: 8px auto 8px 12px !important;
+    max-width: 78% !important;
     padding: 12px 16px !important;
-    font-size: 0.94rem !important;
-    line-height: 1.5 !important;
+    line-height: 1.48;
 }
 
-/* Inputs */
-textarea, input[type="text"] {
-    background: white !important;
+/* Input & Button */
+.gr-text-input, textarea {
     border: 1px solid var(--border-2) !important;
-    border-radius: var(--r) !important;
-    color: var(--text-1) !important;
-    font-size: 0.96rem !important;
-    padding: 12px 16px !important;
-}
-textarea:focus, input[type="text"]:focus {
-    border-color: var(--accent) !important;
-    box-shadow: 0 0 0 3px var(--accent-glow) !important;
-}
-textarea::placeholder {
-    color: var(--text-3) !important;
+    border-radius: var(--radius) !important;
+    background: white !important;
 }
-
-/* Buttons */
-button.primary {
+.gr-button-primary {
     background: var(--accent) !important;
     color: white !important;
     border: none !important;
-    border-radius: 10px !important;
-    font-weight: 500 !important;
 }
-button.primary:hover {
+.gr-button-primary:hover {
     background: var(--accent-light) !important;
 }
-
-/* Tabs & labels */
-label > span, .gr-form label span {
-    font-family: var(--font-mono) !important;
-    font-size: 0.64rem !important;
-    color: var(--text-3) !important;
-    letter-spacing: 0.04em;
-    text-transform: uppercase;
-}
 """
 
-# ── UI ────────────────────────────────────────────────────────────────────────
-with gr.Blocks(css=css, title="Document Intelligence") as demo:
-
-    gr.HTML("""
-    <div style="display:flex; justify-content:space-between; align-items:flex-start; margin-bottom:32px; padding-bottom:20px; border-bottom:1px solid #d2d2d7;">
-        <div>
-            <div style="font-size:2.4rem; font-weight:600; letter-spacing:-0.02em;">Document <span style="color:#0071e3;">Intelligence</span></div>
-            <div style="font-family:ui-monospace,monospace; font-size:0.68rem; color:#6e6e73; letter-spacing:0.06em; text-transform:uppercase; margin-top:6px;">
-                Retrieval-Augmented Generation · Semantic Search · GPT-4o mini
-            </div>
-        </div>
-        <div style="text-align:right;">
-            <div style="font-size:0.9rem; font-weight:500; color:#1d1d1f;">Your Name / Portfolio</div>
-            <div style="font-size:0.7rem; color:#6e6e73;">Gradio + OpenAI</div>
-        </div>
-    </div>
-    """)
+# ── INTERFACE ────────────────────────────────────────────────────────────────
+with gr.Blocks(title="Document Q&A") as demo:
+
+    gr.Markdown(
+        """
+        # Document Intelligence
+        
+        Upload a PDF → ask questions about its content.
+        Answers are grounded **only** in the uploaded document.
+        """
+    )
 
     with gr.Tabs():
-        with gr.TabItem("Upload & Index"):
-            with gr.Row():
-                with gr.Column(scale=1):
-                    pdf_input = gr.File(label="Upload PDF document", file_types=[".pdf"])
-                    upload_btn = gr.Button("Index Document", variant="primary")
-                    upload_status = gr.Textbox(
-                        label="Status",
-                        lines=3,
-                        interactive=False,
-                        value="— Waiting for PDF…",
-                    )
-
-                with gr.Column(scale=1):
-                    gr.Markdown("### Settings & Pipeline")
-                    gr.Markdown("""
-                    - Embedding model: **text-embedding-3-small**  
-                    - Chat model: **gpt-4o-mini**  
-                    - Chunk size: **100** words  
-                    - Overlap: **40** words  
-                    - Retrieval: **top-2** chunks
-                    """)
-
-        with gr.TabItem("Chat"):
+        with gr.Tab("Upload & Index"):
+            pdf_input = gr.File(
+                label="Upload PDF",
+                file_types=[".pdf"],
+                type="filepath"
+            )
+            upload_btn = gr.Button("Index Document", variant="primary")
+            status = gr.Textbox(
+                label="Status",
+                interactive=False,
+                lines=3,
+                value="— Waiting for PDF…"
+            )
+
+        with gr.Tab("Ask Questions"):
             chatbot = gr.Chatbot(
                 height=520,
                 show_label=False,
-                bubble_full_width=False,
-                elem_classes=["chatbot-wrap"],
+                elem_classes=["chatbot"],
+                avatar_images=(None, None)
             )
             with gr.Row():
-                msg_box = gr.Textbox(
-                    placeholder="Ask anything about the document…",
+                msg = gr.Textbox(
+                    placeholder="Ask a question about the document…",
                     container=False,
                     scale=5,
+                    autofocus=True
                 )
-                send_btn = gr.Button("Send", variant="primary", scale=1, min_width=80)
+                send = gr.Button("Send", variant="primary", scale=1, min_width=80)
 
             gr.Markdown(
-                "*Only answers grounded in the uploaded document. Clear chat to start over.*",
-                elem_classes=["small-footer"]
+                "*Answers are strictly based on the indexed document content.*",
+                elem_classes=["small muted"]
             )
 
     # Events
     upload_btn.click(
         fn=upload_pdf,
         inputs=pdf_input,
-        outputs=[upload_status, send_btn]
+        outputs=[status, send]
     )
 
-    send_btn.click(
-        fn=chat,
-        inputs=[msg_box, chatbot],
-        outputs=chatbot
-    ).then(lambda: "", None, msg_box)
+    def respond(message, chat_history):
+        new_history = chat(message, chat_history)
+        return "", new_history
 
-    msg_box.submit(
-        fn=chat,
-        inputs=[msg_box, chatbot],
-        outputs=chatbot
-    ).then(lambda: "", None, msg_box)
+    send.click(
+        fn=respond,
+        inputs=[msg, chatbot],
+        outputs=[msg, chatbot]
+    )
+    msg.submit(
+        fn=respond,
+        inputs=[msg, chatbot],
+        outputs=[msg, chatbot]
+    )
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch(
+        css=css,
+        share=False,          # change to True if you want public link
+        server_name="0.0.0.0",
+        server_port=7860
+    )