Evaluation 1

app-original.py

@@ -0,0 +1,158 @@
+import os, threading
+import gradio as gr
+from crew import run_parallel_crew
+from crew import run_crew
+from utils import get_questions
+
+
+def ask(question, openai_api_key, gemini_api_key, anthropic_api_key, file_name = ""):
+    """
+    Ask General AI Assistant a question to answer.
+
+    Args:
+        question (str): The question to answer
+        openai_api_key (str): OpenAI API key
+        gemini_api_key (str): Gemini API key
+        anthropic_api_key (str): Anthropic API key
+        file_name (str): Optional file name
+
+    Returns:
+        str: The answer to the question
+    """
+    if not question:
+        raise gr.Error("Question is required.")
+
+    if not openai_api_key:
+        raise gr.Error("OpenAI API Key is required.")
+    
+    if not gemini_api_key:
+        raise gr.Error("Gemini API Key is required.")
+
+    if not anthropic_api_key:
+        raise gr.Error("Anthropic API Key is required.")
+        
+    if file_name:
+        file_name = f"data/{file_name}"
+    
+    lock = threading.Lock()
+
+    with lock:
+        answer = ""
+
+        try:
+            os.environ["OPENAI_API_KEY"] = openai_api_key
+            os.environ["GEMINI_API_KEY"] = gemini_api_key
+            os.environ["MODEL_API_KEY"] = anthropic_api_key
+            
+            #answer = run_parallel_crew(question, file_name)
+            answer = run_crew(question, file_name)
+        except Exception as e:
+            raise gr.Error(e)
+        finally:
+            del os.environ["OPENAI_API_KEY"]
+            del os.environ["GEMINI_API_KEY"]
+            del os.environ["MODEL_API_KEY"]
+        
+        return answer
+
+gr.close_all()
+
+with gr.Blocks() as grady:
+    gr.Markdown("## Grady - General AI Assistant")
+
+    with gr.Tab("Solution"):
+        gr.Markdown(os.environ.get("DESCRIPTION"))
+
+        with gr.Row():
+            with gr.Column(scale=3):
+                with gr.Row():
+                    question = gr.Textbox(
+                        label="Question *",
+                        placeholder="In the 2025 Gradio Agents & MCP Hackathon, what percentage of participants submitted a solution during the last 24 hours?",
+                        interactive=True
+                    )
+                with gr.Row():
+                    level = gr.Radio(
+                        choices=[1, 2, 3],
+                        label="GAIA Benchmark Level",
+                        interactive=True,
+                        scale=1
+                    )
+                    ground_truth = gr.Textbox(
+                        label="Ground Truth",
+                        interactive=True,
+                        scale=1
+                    )
+                    file_name = gr.Textbox(
+                        label="File Name",
+                        interactive=True,
+                        scale=2
+                    )
+                with gr.Row():
+                    openai_api_key = gr.Textbox(
+                        label="OpenAI API Key *",
+                        type="password",
+                        placeholder="sk‑...",
+                        interactive=True
+                    )
+                    gemini_api_key = gr.Textbox(
+                        label="Gemini API Key *",
+                        type="password",
+                        interactive=True
+                    )
+                    anthropic_api_key = gr.Textbox(
+                        label="Anthropic API Key *",
+                        type="password",
+                        placeholder="sk-ant-...",
+                        interactive=True
+                    )
+                with gr.Row():
+                    clear_btn = gr.ClearButton(
+                        components=[question, level, ground_truth, file_name]
+                    )
+                    submit_btn = gr.Button("Submit", variant="primary")
+            with gr.Column(scale=1):
+                answer = gr.Textbox(
+                    label="Answer",
+                    lines=1,
+                    interactive=False
+                )
+    
+        submit_btn.click(
+            fn=ask,
+            inputs=[question, openai_api_key, gemini_api_key, anthropic_api_key, file_name],
+            outputs=answer
+        )
+    
+        QUESTION_FILE_PATH = "data/gaia_validation.jsonl"
+        
+        gr.Examples(
+            label="GAIA Benchmark Level 1 Problems",
+            examples=get_questions(QUESTION_FILE_PATH, 1),
+            inputs=[question, level, ground_truth, file_name, openai_api_key, gemini_api_key, anthropic_api_key],
+            outputs=answer,
+            cache_examples=False
+        )
+    
+        gr.Examples(
+            label="GAIA Benchmark Level 2 Problems",
+            examples=get_questions(QUESTION_FILE_PATH, 2),
+            inputs=[question, level, ground_truth, file_name, openai_api_key, gemini_api_key, anthropic_api_key],
+            outputs=answer,
+            cache_examples=False
+        )
+    
+        gr.Examples(
+            label="GAIA Benchmark Level 3 Problems",
+            examples=get_questions(QUESTION_FILE_PATH, 3),
+            inputs=[question, level, ground_truth, file_name, openai_api_key, gemini_api_key, anthropic_api_key],
+            outputs=answer,
+            cache_examples=False
+        )
+    with gr.Tab("Documentation"):
+        gr.Markdown(os.environ.get("DOCUMENTATION"))
+
+grady.launch(mcp_server=True)
+
+
+

app.py

@@ -1,158 +1,90 @@
-import os, threading
+import os
+import requests
+import pandas as pd
 import gradio as gr
-from crew import run_parallel_crew
-from crew import run_crew
-from utils import get_questions
 
+from crew import run_crew
 
-def ask(question, openai_api_key, gemini_api_key, anthropic_api_key, file_name = ""):
-    """
-    Ask General AI Assistant a question to answer.
+# Configuration: endpoint for GAIA evaluation API
+API_URL = os.getenv("GAIA_API_URL", "https://huggingface.co/spaces/Psiska/General_AI_Assistant")
+# Your Space identifier for generating the agent_code URL
+SPACE_ID = os.getenv("SPACE_ID", "Psiska/General_AI_Assistant")
 
-    Args:
-        question (str): The question to answer
-        openai_api_key (str): OpenAI API key
-        gemini_api_key (str): Gemini API key
-        anthropic_api_key (str): Anthropic API key
-        file_name (str): Optional file name
 
-    Returns:
-        str: The answer to the question
+def run_and_submit_all(username: str):
     """
-    if not question:
-        raise gr.Error("Question is required.")
-
-    if not openai_api_key:
-        raise gr.Error("OpenAI API Key is required.")
-    
-    if not gemini_api_key:
-        raise gr.Error("Gemini API Key is required.")
-
-    if not anthropic_api_key:
-        raise gr.Error("Anthropic API Key is required.")
-        
-    if file_name:
-        file_name = f"data/{file_name}"
-    
-    lock = threading.Lock()
-
-    with lock:
-        answer = ""
-
-        try:
-            os.environ["OPENAI_API_KEY"] = openai_api_key
-            os.environ["GEMINI_API_KEY"] = gemini_api_key
-            os.environ["MODEL_API_KEY"] = anthropic_api_key
-            
-            #answer = run_parallel_crew(question, file_name)
+    Fetches all evaluation questions, runs your agent on each,
+    and submits the batch to the /submit endpoint.
+    Returns a status message and a DataFrame of logs.
+    """
+    if not username:
+        return "🔒 Please enter your Hugging Face username.", None
+
+    try:
+        # 1) Fetch questions
+        resp = requests.get(f"{API_URL}/questions", timeout=15)
+        resp.raise_for_status()
+        questions = resp.json()
+
+        # 2) Run agent on each question
+        logs = []
+        answers = []
+        for item in questions:
+            task_id   = item.get("task_id") or item.get("id")
+            question  = item.get("question", "")
+            file_name = item.get("file_name", "")
+
+            # Optional: download attached file
+            if file_name:
+                file_resp = requests.get(f"{API_URL}/files/{task_id}", timeout=15)
+                file_resp.raise_for_status()
+                local_path = os.path.join("data", file_name)
+                os.makedirs(os.path.dirname(local_path), exist_ok=True)
+                with open(local_path, "wb") as f:
+                    f.write(file_resp.content)
+
+            # Get agent's answer
             answer = run_crew(question, file_name)
-        except Exception as e:
-            raise gr.Error(e)
-        finally:
-            del os.environ["OPENAI_API_KEY"]
-            del os.environ["GEMINI_API_KEY"]
-            del os.environ["MODEL_API_KEY"]
-        
-        return answer
-
-gr.close_all()
-
-with gr.Blocks() as grady:
-    gr.Markdown("## Grady - General AI Assistant")
-
-    with gr.Tab("Solution"):
-        gr.Markdown(os.environ.get("DESCRIPTION"))
-
-        with gr.Row():
-            with gr.Column(scale=3):
-                with gr.Row():
-                    question = gr.Textbox(
-                        label="Question *",
-                        placeholder="In the 2025 Gradio Agents & MCP Hackathon, what percentage of participants submitted a solution during the last 24 hours?",
-                        interactive=True
-                    )
-                with gr.Row():
-                    level = gr.Radio(
-                        choices=[1, 2, 3],
-                        label="GAIA Benchmark Level",
-                        interactive=True,
-                        scale=1
-                    )
-                    ground_truth = gr.Textbox(
-                        label="Ground Truth",
-                        interactive=True,
-                        scale=1
-                    )
-                    file_name = gr.Textbox(
-                        label="File Name",
-                        interactive=True,
-                        scale=2
-                    )
-                with gr.Row():
-                    openai_api_key = gr.Textbox(
-                        label="OpenAI API Key *",
-                        type="password",
-                        placeholder="sk‑...",
-                        interactive=True
-                    )
-                    gemini_api_key = gr.Textbox(
-                        label="Gemini API Key *",
-                        type="password",
-                        interactive=True
-                    )
-                    anthropic_api_key = gr.Textbox(
-                        label="Anthropic API Key *",
-                        type="password",
-                        placeholder="sk-ant-...",
-                        interactive=True
-                    )
-                with gr.Row():
-                    clear_btn = gr.ClearButton(
-                        components=[question, level, ground_truth, file_name]
-                    )
-                    submit_btn = gr.Button("Submit", variant="primary")
-            with gr.Column(scale=1):
-                answer = gr.Textbox(
-                    label="Answer",
-                    lines=1,
-                    interactive=False
-                )
-    
-        submit_btn.click(
-            fn=ask,
-            inputs=[question, openai_api_key, gemini_api_key, anthropic_api_key, file_name],
-            outputs=answer
-        )
-    
-        QUESTION_FILE_PATH = "data/gaia_validation.jsonl"
-        
-        gr.Examples(
-            label="GAIA Benchmark Level 1 Problems",
-            examples=get_questions(QUESTION_FILE_PATH, 1),
-            inputs=[question, level, ground_truth, file_name, openai_api_key, gemini_api_key, anthropic_api_key],
-            outputs=answer,
-            cache_examples=False
+            answers.append({"task_id": task_id, "submitted_answer": answer})
+            logs.append({"Task ID": task_id, "Question": question, "Answer": answer})
+
+        # 3) Prepare payload
+        payload = {
+            "username":   username,
+            "agent_code": f"https://huggingface.co/spaces/{SPACE_ID}/tree/main",
+            "answers":    answers
+        }
+
+        # 4) Submit answers
+        submit_resp = requests.post(f"{API_URL}/submit", json=payload, timeout=60)
+        submit_resp.raise_for_status()
+        result = submit_resp.json()
+
+        # Format status
+        status = (
+            f"✅ {result['username']} scored {result['score']}% "
+            f"({result['correct_count']}/{result['total_attempted']} correct)"
         )
-    
-        gr.Examples(
-            label="GAIA Benchmark Level 2 Problems",
-            examples=get_questions(QUESTION_FILE_PATH, 2),
-            inputs=[question, level, ground_truth, file_name, openai_api_key, gemini_api_key, anthropic_api_key],
-            outputs=answer,
-            cache_examples=False
-        )
-    
-        gr.Examples(
-            label="GAIA Benchmark Level 3 Problems",
-            examples=get_questions(QUESTION_FILE_PATH, 3),
-            inputs=[question, level, ground_truth, file_name, openai_api_key, gemini_api_key, anthropic_api_key],
-            outputs=answer,
-            cache_examples=False
-        )
-    with gr.Tab("Documentation"):
-        gr.Markdown(os.environ.get("DOCUMENTATION"))
+        return status, pd.DataFrame(logs)
+
+    except Exception as e:
+        return f"❌ Error: {str(e)}", None
+
 
-grady.launch(mcp_server=True)
+# Build Gradio interface
+with gr.Blocks(title="GAIA Evaluation Runner") as demo:
+    gr.Markdown("# GAIA Evaluation Runner")
+    username_input = gr.Textbox(label="Hugging Face Username")
 
+    run_btn = gr.Button("Run & Submit All Answers")
+    status  = gr.Textbox(label="Status", interactive=False)
+    table   = gr.DataFrame(headers=["Task ID", "Question", "Answer"], label="Log of Q&A")
 
+    run_btn.click(
+        fn=run_and_submit_all,
+        inputs=[username_input],
+        outputs=[status, table]
+    )
 
+if __name__ == "__main__":
+    demo.launch()

evaluation.py

@@ -1,92 +0,0 @@
-import os
-import requests
-import pandas as pd
-import gradio as gr
-
-from crew import run_crew
-
-# Configuration: endpoint for GAIA evaluation API
-API_URL = os.getenv("GAIA_API_URL", "https://huggingface.co/spaces/Psiska/General_AI_Assistant")
-# Your Space identifier for generating the agent_code URL
-SPACE_ID = os.getenv("SPACE_ID", "Psiska/General_AI_Assistant")
-
-
-def run_and_submit_all(profile: gr.OAuthProfile | None):
-    """
-    Fetches all evaluation questions, runs your agent on each,
-    and submits the batch to the /submit endpoint.
-    Returns a status message and a DataFrame of logs.
-    """
-    if profile is None:
-        return "🔒 Please log in with your Hugging Face account.", None
-
-    username = profile.username
-    try:
-        # 1) Fetch questions
-        resp = requests.get(f"{API_URL}/questions", timeout=15)
-        resp.raise_for_status()
-        questions = resp.json()
-
-        # 2) Run agent on each question
-        logs = []
-        answers = []
-        for item in questions:
-            task_id   = item.get("task_id") or item.get("id")
-            question  = item.get("question", "")
-            file_name = item.get("file_name", "")
-
-            # Optional: download attached file
-            if file_name:
-                file_resp = requests.get(f"{API_URL}/files/{task_id}", timeout=15)
-                file_resp.raise_for_status()
-                local_path = os.path.join("data", file_name)
-                os.makedirs(os.path.dirname(local_path), exist_ok=True)
-                with open(local_path, "wb") as f:
-                    f.write(file_resp.content)
-                # pass file_name or path to your agent if needed
-
-            # Get agent's answer
-            answer = run_crew(question, file_name)
-            answers.append({"task_id": task_id, "submitted_answer": answer})
-            logs.append({"Task ID": task_id, "Question": question, "Answer": answer})
-
-        # 3) Prepare payload
-        payload = {
-            "username":   username,
-            "agent_code": f"https://huggingface.co/spaces/{SPACE_ID}/tree/main",
-            "answers":    answers
-        }
-
-        # 4) Submit answers
-        submit_resp = requests.post(f"{API_URL}/submit", json=payload, timeout=60)
-        submit_resp.raise_for_status()
-        result = submit_resp.json()
-
-        # Format status
-        status = (
-            f"✅ {result['username']} scored {result['score']}% "
-            f"({result['correct_count']}/{result['total_attempted']} correct)"
-        )
-        return status, pd.DataFrame(logs)
-
-    except Exception as e:
-        return f"❌ Error: {str(e)}", None
-
-
-# Build Gradio interface
-with gr.Blocks(title="GAIA Evaluation Runner") as demo:
-    gr.Markdown("# GAIA Evaluation Runner")
-    login = gr.LoginButton()
-
-    run_btn = gr.Button("Run & Submit All Answers")
-    status  = gr.Textbox(label="Status", interactive=False)
-    table   = gr.DataFrame(headers=["Task ID", "Question", "Answer"], label="Log of Q&A")
-
-    run_btn.click(
-        fn=run_and_submit_all,
-        inputs=[login],
-        outputs=[status, table]
-    )
-
-if __name__ == "__main__":
-    demo.launch()