Create app.py

app.py

@@ -0,0 +1,117 @@
+import os
+import io
+import pdfplumber
+from huggingface_hub import InferenceClient
+from PIL import Image
+import gradio as gr
+import base64
+
+# ---------------- CONFIG ----------------
+LLAMA_MODEL = "Groq/Llama-3-Groq-8B-Tool-Use"
+TTS_MODEL = "espnet/kan-bayashi_ljspeech_vits"
+SDXL_MODEL = "stabilityai/stable-diffusion-xl-base-1.0"
+
+HF_TOKEN = os.environ.get("HF_TOKEN")
+GROQ_TOKEN = os.environ.get("GROQ_TOKEN")
+
+client = None
+if GROQ_TOKEN:
+    client = InferenceClient(provider="groq", api_key=GROQ_TOKEN)
+elif HF_TOKEN:
+    client = InferenceClient(api_key=HF_TOKEN)
+
+# ---------------- HELPERS ----------------
+def pdf_to_text(file):
+    text_chunks = []
+    pages = 0
+    with pdfplumber.open(file) as pdf:
+        pages = len(pdf.pages)
+        for page in pdf.pages:
+            ptext = page.extract_text() or ""
+            text_chunks.append(ptext)
+    return "\n\n".join(text_chunks), pages
+
+def llama_summarize(text):
+    messages = [
+        {"role": "system", "content": "You are a concise summarizer. Give 6 short bullet points."},
+        {"role": "user", "content": f"Summarize this document in 6 concise bullet points:\n\n{text}"}
+    ]
+    resp = client.chat.completions.create(model=LLAMA_MODEL, messages=messages)
+    return resp.choices[0].message["content"]
+
+def llama_chat(history, question):
+    messages = history + [{"role": "user", "content": question}]
+    resp = client.chat.completions.create(model=LLAMA_MODEL, messages=messages)
+    return resp.choices[0].message["content"]
+
+def tts_synthesize(text):
+    audio_bytes = client.text_to_speech(model=TTS_MODEL, inputs=text)
+    return audio_bytes
+
+def generate_image(prompt_text):
+    img_bytes = client.text_to_image(prompt_text, model=SDXL_MODEL)
+    return Image.open(io.BytesIO(img_bytes))
+
+def ask_question_and_maybe_diagram(chat_text, question, history):
+    if not history:
+        history = [{"role": "system", "content": f"Document context:\n{chat_text[:4000]}"}]
+    ans = llama_chat(history, question)
+    history.append({"role": "user", "content": question})
+    history.append({"role": "assistant", "content": ans})
+
+    diagram_img = None
+    if question.strip().lower().startswith("!diagram"):
+        prompt = question[len("!diagram"):].strip()
+        if prompt:
+            diagram_img = generate_image(prompt)
+    return ans, diagram_img, history
+
+# ---------------- GRADIO INTERFACE ----------------
+with gr.Blocks() as demo:
+    gr.Markdown("## 📄 PDF Buddy — Summarize • Speak • Chat • Draw")
+    
+    with gr.Row():
+        pdf_file = gr.File(label="Upload PDF", type="file")
+        extract_status = gr.Textbox(label="Status")
+    
+    extracted_text = gr.Textbox(label="Document Preview", lines=10)
+    
+    with gr.Row():
+        summarize_btn = gr.Button("📝 Summarize")
+        summary_output = gr.Textbox(label="Summary", lines=6)
+    
+    tts_btn = gr.Button("🔊 Synthesize Summary to Audio")
+    audio_out = gr.Audio(label="Audio", type="filepath")
+    
+    chat_question = gr.Textbox(label="Ask a question (use !diagram for image)")
+    chat_btn = gr.Button("❓ Ask")
+    chat_output = gr.Textbox(label="Answer")
+    diagram_out = gr.Image(label="Diagram (optional)")
+    chat_history_state = gr.State()
+    
+    # ---------------- CALLBACKS ----------------
+    pdf_file.change(
+        lambda f: pdf_to_text(f) if f else ("No file uploaded", "", None),
+        inputs=pdf_file,
+        outputs=[extract_status, extracted_text]
+    )
+    
+    summarize_btn.click(
+        lambda text: llama_summarize(text[:30000]) if text else "No text to summarize",
+        inputs=extracted_text,
+        outputs=summary_output
+    )
+    
+    tts_btn.click(
+        lambda summary: tts_synthesize(summary) if summary else None,
+        inputs=summary_output,
+        outputs=audio_out
+    )
+    
+    chat_btn.click(
+        ask_question_and_maybe_diagram,
+        inputs=[extracted_text, chat_question, chat_history_state],
+        outputs=[chat_output, diagram_out, chat_history_state]
+    )
+
+demo.launch()