File upload added

app.py

@@ -1,10 +1,28 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+import pymupdf
+from duckduckgo_search import DDGS
 
 """
 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")
+client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
+
+
+# PDF Parsing
+def extract_text_from_pdf(pdf_file):
+    doc = pymupdf.open(pdf_file)
+    text = " ".join([page.get_textpage().extractTEXT() for page in doc])
+    return text
+
+
+# Web search fallback
+def search_web(query):
+    with DDGS() as ddgs:
+        results = ddgs.text(query)
+        if results:
+            return results[0]["body"]
+    return "No relevant results found on the web."
 
 
 def respond(
@@ -14,19 +32,51 @@ def respond(
     max_tokens,
     temperature,
     top_p,
+    pdf_file=None,
 ):
-    messages = [{"role": "system", "content": system_message}]
+    recent_keywords = [
+        "latest",
+        "today",
+        "current",
+        "now",
+        "recent",
+        "news",
+        "update",
+        "price",
+        "who won",
+        "what happened",
+    ]
+    message_lower = message.lower()
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    # Check if this is a recent/live query
+    if any(kw in message_lower for kw in recent_keywords):
+        web_result = search_web(message)
+        if web_result:
+            yield f"[Answer from Web Search]\n{web_result}"
+            return
 
+    # Check if PDF was uploaded
+    context_text = ""
+    if pdf_file is not None:
+        try:
+            context_text = extract_text_from_pdf(pdf_file.name)
+        except Exception as e:
+            yield f"[Error reading PDF: {str(e)}]"
+
+    # Try answering from PDF if context exists
+    if context_text.strip():
+        message = f"Given the following document context:\n\n{context_text[:2000]}\n\nAnswer the question:\n{message}"
+
+    # Fallback to LLM (Zephyr)
+    messages = [{"role": "system", "content": system_message}]
+    for user, assistant in history:
+        if user:
+            messages.append({"role": "user", "content": user})
+        if assistant:
+            messages.append({"role": "assistant", "content": assistant})
     messages.append({"role": "user", "content": message})
 
     response = ""
-
     for message in client.chat_completion(
         messages,
         max_tokens=max_tokens,
@@ -35,8 +85,7 @@ def respond(
         top_p=top_p,
     ):
         token = message.choices[0].delta.content
-
-        response += token
+        response += token if token is not None else ""
         yield response
 
 
@@ -56,6 +105,7 @@ demo = gr.ChatInterface(
             step=0.05,
             label="Top-p (nucleus sampling)",
         ),
+        gr.File(label="Upload a PDF", file_types=[".pdf"]),
     ],
 )
 

requirements.txt

@@ -1 +1,4 @@
-huggingface_hub==0.25.2
+huggingface_hub==0.25.2
+duckduckgo_search
+pymupdf
+gradio