Update app.py

app.py

@@ -1,5 +1,3 @@
-# TOOLS
-
 import os
 import gradio as gr
 import logging
@@ -10,28 +8,36 @@ from llama_index.core.tools import FunctionTool
 from duckduckgo_search import DDGS
 from sentence_transformers import SentenceTransformer
 from sklearn.metrics.pairwise import cosine_similarity
-from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI
-
+from llama_index.llms.huggingface import HuggingFaceLLM
+from llama_index.core.agent import ReActAgent
 import numpy as np
+import ast
+import operator as op
 
-# Load HF_TOKEN from environment
-HF_TOKEN = os.environ.get("HF_TOKEN")
-
+# === Setup ===
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
 
-# === Core Components ===
-class QuestionValidation:
-    def __init__(self, hf_token: str):
-        self.client = InferenceClient(model="HuggingFaceH4/zephyr-7b-beta", token=HF_TOKEN)
+HF_TOKEN = os.environ.get("HF_TOKEN")
 
-        self.llm2_model = "HuggingFaceH4/zephyr-7b-beta"
-        self.embedding_model = SentenceTransformer("sentence-transformers/paraphrase-MiniLM-L3-v2")
+# === LLM Setup (Shared) ===
+llm = HuggingFaceLLM(
+    context_window=4096,
+    max_new_tokens=512,
+    generate_kwargs={"temperature": 0.7, "top_p": 0.95},
+    tokenizer_name="HuggingFaceH4/zephyr-7b-beta",
+    model_name="HuggingFaceH4/zephyr-7b-beta",
+)
+
+# === Question Validation Component ===
+class QuestionValidation:
+    def __init__(self, llm_client):
+        self.client = llm_client
+        self.embedding_model = SentenceTransformer("all-MiniLM-L6-v2")
 
     def guess_question(self, answer: str) -> str:
         prompt = f"This was the answer: {answer}\nWhat question would likely have led to it?"
-        response = self.client.text_generation(prompt=prompt, model=self.llm2_model, max_new_tokens=100)
-        return response
+        return self.client.complete(prompt).text.strip()
 
     def compute_similarity(self, q1: str, q2: str) -> float:
         embeddings = self.embedding_model.encode([q1, q2])
@@ -45,17 +51,14 @@ class QuestionValidation:
             "similarity": round(float(similarity), 4)
         }
 
+# === Tools ===
 def search_web(query: str, max_results: int = 5) -> List[Dict[str, str]]:
     try:
         with DDGS() as ddgs:
-            results = [r for r in ddgs.text(query, max_results=max_results)]
-        return results
+            return [r for r in ddgs.text(query, max_results=max_results)]
     except Exception as e:
         return [{"error": str(e)}]
 
-import ast
-import operator as op
-
 OPERATORS = {
     ast.Add: op.add,
     ast.Sub: op.sub,
@@ -71,29 +74,23 @@ OPERATORS = {
 def evaluate_math_expression(expr: str) -> str:
     try:
         node = ast.parse(expr, mode="eval")
-
         def _eval(node):
             if isinstance(node, ast.Expression):
                 return _eval(node.body)
-            elif isinstance(node, ast.Num):
-                return node.n
-            elif isinstance(node, ast.Constant): 
-                if isinstance(node.value, (int, float)):
-                    return node.value
+            elif isinstance(node, ast.Constant):
+                return node.value
             elif isinstance(node, ast.BinOp):
                 return OPERATORS[type(node.op)](_eval(node.left), _eval(node.right))
             elif isinstance(node, ast.UnaryOp):
                 return OPERATORS[type(node.op)](_eval(node.operand))
             else:
                 raise ValueError(f"Unsupported expression: {ast.dump(node)}")
-
-        result = _eval(node)
-        return str(result)
+        return str(_eval(node))
     except Exception as e:
         return f"Error evaluating expression: {e}"
 
-
-validator = QuestionValidation(hf_token=HF_TOKEN)
+# Instantiate validator using shared LLM
+validator = QuestionValidation(llm)
 
 validate_tool = FunctionTool.from_defaults(
     fn=validator.validate_question_only,
@@ -115,20 +112,6 @@ math_tool = FunctionTool.from_defaults(
 
 TOOLS = [validate_tool, search_tool, math_tool]
 
-from llama_index.core.agent import ReActAgent
-
-
-from llama_index.llms.huggingface import HuggingFaceLLM
-
-llm = HuggingFaceLLM(
-    context_window=4096,
-    max_new_tokens=512,
-    generate_kwargs={"temperature": 0.7, "top_p": 0.95},
-    tokenizer_name="HuggingFaceH4/zephyr-7b-beta",
-    model_name="HuggingFaceH4/zephyr-7b-beta",
-    
-)
-
 agent = ReActAgent.from_tools(
     tools=TOOLS,
     llm=llm,
@@ -136,6 +119,7 @@ agent = ReActAgent.from_tools(
     max_iterations=3
 )
 
+# === Question Loop Logic ===
 def question_loop_agent(user_question: str):
     llm_answer = llm.complete(user_question).text.strip()
     similarity_score = 0.0
@@ -156,57 +140,42 @@ def question_loop_agent(user_question: str):
         f"Similarity Score: {similarity_score:.4f}"
     )
 
-iface_loop = gr.Interface(
-    fn=question_loop_agent,
-    inputs=gr.Textbox(lines=2, placeholder="Ask me a question..."),
-    outputs="text",
-    title="Question Similarity Loop Agent",
-    description="Loops until the guessed question has a similarity score > 0.6."
-)
-
-
-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
-
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a 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)"),
-    ],
-)
+# === Gradio Interfaces ===
+with gr.Blocks() as app:
+    with gr.Tab("Validation Loop"):
+        gr.Interface(
+            fn=question_loop_agent,
+            inputs=gr.Textbox(lines=2, placeholder="Ask me a question..."),
+            outputs="text",
+            title="Question Similarity Loop Agent",
+            description="Loops until the guessed question has a similarity score > 0.6."
+        ).render()
+
+    with gr.Tab("Chat"):
+        def respond(message, history, system_message, max_tokens, temperature, top_p):
+            messages = [{"role": "system", "content": system_message}]
+            for user, bot in history:
+                if user:
+                    messages.append({"role": "user", "content": user})
+                if bot:
+                    messages.append({"role": "assistant", "content": bot})
+            messages.append({"role": "user", "content": message})
+
+            response = ""
+            for msg in llm.chat(messages, stream=True):
+                token = msg.delta or ""
+                response += token
+                yield response
+
+        gr.ChatInterface(
+            respond,
+            additional_inputs=[
+                gr.Textbox(value="You are a 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")
+            ],
+        ).render()
 
 if __name__ == "__main__":
-    iface_loop.launch()
-    demo.launch()  # Uncomment to run the chat template toon
+    app.launch()