Update app.py

app.py

@@ -1,64 +1,129 @@
+import os
+import shutil
+import asyncio
+from pathlib import Path
+
 import gradio as gr
-from huggingface_hub import InferenceClient
-
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
+from PyPDF2 import PdfReader  # pip install PyPDF2
+
+from helper import get_openai_api_key, get_llama_cloud_api_key
+from llama_parse import LlamaParse
+from llama_index.core import (
+    Settings, VectorStoreIndex, StorageContext, load_index_from_storage
 )
+from llama_index.llms.openai import OpenAI
+from llama_index.embeddings.openai import OpenAIEmbedding
+from llama_index.core.tools import QueryEngineTool
+from llama_index.core.query_engine import SubQuestionQueryEngine
+from llama_index.core.agent.workflow import FunctionAgent
+from llama_index.core.workflow import Context
+
+# ---- 1. Global Settings & API Keys ----
+Settings.llm           = OpenAI(model="gpt-4o")
+Settings.embed_model   = OpenAIEmbedding(model_name="text-embedding-3-large")
+Settings.chunk_size    = 512
+Settings.chunk_overlap = 64
+
+os.environ["OPENAI_API_KEY"]      = get_openai_api_key()
+os.environ["LLAMA_CLOUD_API_KEY"] = get_llama_cloud_api_key()
+
+# ---- 2. Parser Setup ----
+parser = LlamaParse(
+    api_key  = os.getenv("LLAMA_CLOUD_API_KEY"),
+    base_url = os.getenv("LLAMA_CLOUD_BASE_URL"),
+    result_type = "markdown",
+    content_guideline_instruction = (
+        "You are processing a PDF slide deck. "
+        "Produce Markdown with slide metadata, cleaned bullets, tables, "
+        "charts summaries, figures captions, metrics, and a 1–2 sentence takeaway."
+    ),
+    verbose=True
+)
+
+# ---- 3. Core “Answer” Logic ----
+async def answer(uploaded_files: list[gr.FileData], question: str) -> str:
+    # Validate uploads
+    if not uploaded_files:
+        return "❗ Please upload at least one PDF."
+    if len(uploaded_files) > 5:
+        return "❗ You can upload up to 5 PDF files."
+
+    # Ensure user_data directory
+    user_dir = Path("./user_data")
+    user_dir.mkdir(exist_ok=True)
+
+    # Prepare list of QueryEngineTools
+    tools = []
+    for file_obj in uploaded_files:
+        # Read page count
+        try:
+            reader = PdfReader(file_obj.name)
+        except Exception as e:
+            return f"❗ Error reading {file_obj.name}: {e}"
+        if len(reader.pages) > 20:
+            return f"❗ {Path(file_obj.name).name} has {len(reader.pages)} pages (>20)."
+
+        # Copy file to persistent location
+        dest = user_dir / Path(file_obj.name).name
+        shutil.copyfile(file_obj.name, dest)  # permanent copy :contentReference[oaicite:3]{index=3}
+
+        # Parse PDF into Documents
+        docs = parser.load_data(dest)
+
+        # Index folder named after file stem
+        stem = dest.stem
+        idx_dir = Path(f"./index_data/{stem}")
+
+        # Load or build index
+        if idx_dir.exists() and any(idx_dir.iterdir()):
+            sc = StorageContext.from_defaults(persist_dir=str(idx_dir))
+            idx = load_index_from_storage(sc)
+        else:
+            sc  = StorageContext.from_defaults()
+            idx = VectorStoreIndex.from_documents(docs, storage_context=sc)
+            sc.persist(persist_dir=str(idx_dir))  # persist per-file index :contentReference[oaicite:4]{index=4}
+
+        # Create a QueryEngineTool for this index
+        qe_tool = QueryEngineTool.from_defaults(
+            query_engine=idx.as_query_engine(),
+            name=f"vector_index_{stem}",
+            description=f"Query engine for slides in {stem}.pdf"
+        )
+        tools.append(qe_tool)
+
+    # Combine into SubQuestionQueryEngine + Agent
+    subq = SubQuestionQueryEngine.from_defaults(query_engine_tools=tools)
+    tools.append(
+        QueryEngineTool.from_defaults(
+            query_engine=subq,
+            name="sub_question_query_engine",
+            description="Multi-file comparative queries"
+        )
+    )
+    agent = FunctionAgent(tools=tools, llm=OpenAI(model="gpt-4o"))
+    ctx   = Context(agent)
+
+    # Run agent
+    response = await agent.run(question, ctx=ctx)
+    return str(response)
+
+# ---- 4. Gradio UI ----
+with gr.Blocks() as demo:
+    gr.Markdown("# 📄 PDF Slide Deck Q&A Bot")
+    with gr.Row():
+        file_input = gr.UploadButton(
+            "Upload up to 5 PDFs",
+            file_types=[".pdf"],
+            file_count="multiple"  # support multiple uploads 
+        )
+        question = gr.Textbox(
+            lines=2,
+            placeholder="Ask your question about the uploaded slide decks..."
+        )
+    output = gr.Textbox(label="Answer")
+    submit = gr.Button("Ask")
 
+    submit.click(fn=answer, inputs=[file_input, question], outputs=output)
 
 if __name__ == "__main__":
     demo.launch()