Update app.py

app.py

@@ -1,6 +1,6 @@
 import os
-import json
 import requests
+import json
 import pandas as pd
 import gradio as gr
 
@@ -9,51 +9,71 @@ from langchain.docstore.document import Document
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_community.retrievers import BM25Retriever
 
-from huggingface_hub.inference_api import InferenceApi
 from smolagents import Tool, CodeAgent
 
-# ----- HF Inference Wrapper -----
+from huggingface_hub.inference_api import InferenceApi
+
+# Load HF token from environment
+hf_token = os.getenv("HUGGINGFACE_API_KEY")
+print("Token from env var:", hf_token)
+if hf_token:
+    os.environ["HUGGINGFACE_API_KEY"] = hf_token
+    print("Set HUGGINGFACE_API_KEY in env.")
+else:
+    print("No HUGGINGFACE_API_KEY found in env.")
+
 class HuggingFaceInferenceWrapper:
     def __init__(self, inference_api):
         self.inference_api = inference_api
 
     def generate(self, prompt: str, **kwargs) -> str:
-        raw_response = self.inference_api(inputs=prompt, raw_response=True)
-        json_data = json.loads(raw_response.content)
+        response = self.inference_api(inputs=prompt, raw_response=True)
+
+        # Handle response based on type
+        if hasattr(response, "content"):
+            # requests.Response-like object
+            json_data = json.loads(response.content)
+        else:
+            # Sometimes response might be a string already
+            try:
+                json_data = json.loads(response)
+            except Exception:
+                # Fallback: return raw string response
+                return str(response)
+
+        # Extract generated_text from json
         if isinstance(json_data, dict) and "generated_text" in json_data:
             return json_data["generated_text"].strip()
-        elif isinstance(json_data, list) and len(json_data) > 0 and "generated_text" in json_data[0]:
+        elif (
+            isinstance(json_data, list)
+            and len(json_data) > 0
+            and "generated_text" in json_data[0]
+        ):
             return json_data[0]["generated_text"].strip()
         else:
+            # fallback: return entire json as string
             return str(json_data)
 
-
-# ----- Setup HF API -----
-hf_token = os.getenv("HUGGINGFACE_API_KEY")
-if not hf_token:
-    raise ValueError("HUGGINGFACE_API_KEY environment variable is not set")
-
-inference_api = InferenceApi(repo_id="Qwen/Qwen2.5-VL-7B-Instruct", token=hf_token)
-model = HuggingFaceInferenceWrapper(inference_api)
-
-
 def run_and_submit_all(profile: gr.OAuthProfile | None):
-    space_id = os.getenv("SPACE_ID")
+    space_id = os.getenv("SPACE_ID")  # For linking repo code
 
     if profile:
-        username = profile.username
+        username = f"{profile.username}"
         print(f"User logged in: {username}")
     else:
         print("User not logged in.")
         return "Please Login to Hugging Face with the button.", None
 
-    api_url = "https://agents-course-unit4-scoring.hf.space"
+    api_url = "https://agents-course-unit4-scoring.hf.space"  # Change if needed
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
     try:
+        # Load knowledge base and filter for retriever
         knowledge_base = datasets.load_dataset("m-ric/huggingface_doc", split="train")
-        knowledge_base = knowledge_base.filter(lambda row: row["source"].startswith("huggingface/transformers"))
+        knowledge_base = knowledge_base.filter(
+            lambda row: row["source"].startswith("huggingface/transformers")
+        )
 
         source_docs = [
             Document(page_content=doc["text"], metadata={"source": doc["source"].split("/")[1]})
@@ -72,15 +92,13 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         class RetrieverTool(Tool):
             name = "retriever"
             description = (
-                "Uses lexical search to retrieve the parts of transformers documentation "
-                "that could be most relevant to answer your query."
+                "Uses lexical search to retrieve relevant parts of transformers docs."
             )
             inputs = {
                 "query": {
                     "type": "string",
                     "description": (
-                        "The query to perform. This should be lexically close to your target documents. "
-                        "Use the affirmative form rather than a question."
+                        "The query to perform. Should be lexically close to your target documents."
                     ),
                 }
             }
@@ -99,22 +117,27 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
         retriever_tool = RetrieverTool(docs_processed)
 
-        # Instantiate CodeAgent with our wrapped model
+        # Instantiate HuggingFace Inference API wrapper
+        inference_api = InferenceApi(repo_id="Qwen/Qwen2.5-VL-7B-Instruct", token=hf_token)
+        model_wrapper = HuggingFaceInferenceWrapper(inference_api)
+
+        # Instantiate the agent with our wrapped model
         agent = CodeAgent(
             tools=[retriever_tool],
-            model=model,
+            model=model_wrapper,
             max_steps=4,
             verbosity_level=2,
-            stream_outputs=False,  # Must be False for this wrapper
+            stream_outputs=False,  # must be False for this wrapper
         )
 
     except Exception as e:
         return f"Error initializing agent: {e}", None
 
-    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" if space_id else "Code repo URL not available"
+    agent_code = (
+        f"https://huggingface.co/spaces/{space_id}/tree/main" if space_id else "Code repo URL not available"
+    )
     print(agent_code)
 
-    # Fetch Questions
     try:
         response = requests.get(questions_url, timeout=15)
         response.raise_for_status()
@@ -143,6 +166,8 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
     submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
+    print(f"Submitting {len(answers_payload)} answers...")
+
     try:
         response = requests.post(submit_url, json=submission_data, timeout=60)
         response.raise_for_status()
@@ -157,21 +182,24 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         results_df = pd.DataFrame(results_log)
         return final_status, results_df
     except Exception as e:
+        status_message = f"Submission Failed: {e}"
         results_df = pd.DataFrame(results_log)
-        return f"Submission Failed: {e}", results_df
-
+        return status_message, results_df
 
-# --- Gradio UI ---
 
+# Gradio UI code unchanged from your original snippet
 with gr.Blocks() as demo:
     gr.Markdown("# Basic Agent Evaluation Runner")
     gr.Markdown(
         """
         **Instructions:**
-        1. Clone this space and modify the code to define your agent's logic, tools, packages, etc.
-        2. Log in with the Hugging Face button.
-        3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
+        1.  Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
+        2.  Log in to your Hugging Face account using the button below. This uses your HF username for submission.
+        3.  Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
         ---
+        **Disclaimers:**
+        Once clicking on the "submit" button, it can take quite some time (this is the time for the agent to go through all the questions).
+        This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a separate action or even to answer the questions asynchronously.
         """
     )
 
@@ -182,7 +210,11 @@ with gr.Blocks() as demo:
     status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
     results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
 
-    run_button.click(fn=run_and_submit_all, outputs=[status_output, results_table])
+    run_button.click(
+        fn=run_and_submit_all,
+        outputs=[status_output, results_table]
+    )
+
 
 if __name__ == "__main__":
     print("\n" + "-" * 30 + " App Starting " + "-" * 30)