Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -3,7 +3,7 @@ import gradio as gr
|
|
3 |
import requests
|
4 |
import inspect
|
5 |
import pandas as pd
|
6 |
-
|
7 |
# (Keep Constants as is)
|
8 |
# --- Constants ---
|
9 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
@@ -15,9 +15,10 @@ class BasicAgent:
|
|
15 |
print("BasicAgent initialized.")
|
16 |
def __call__(self, question: str) -> str:
|
17 |
print(f"Agent received question (first 50 chars): {question[:50]}...")
|
18 |
-
|
19 |
-
|
20 |
-
|
|
|
21 |
|
22 |
def run_and_submit_all( profile: gr.OAuthProfile | None):
|
23 |
"""
|
@@ -73,13 +74,14 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
|
|
73 |
results_log = []
|
74 |
answers_payload = []
|
75 |
print(f"Running agent on {len(questions_data)} questions...")
|
76 |
-
for item in questions_data:
|
77 |
task_id = item.get("task_id")
|
78 |
question_text = item.get("question")
|
79 |
if not task_id or question_text is None:
|
80 |
print(f"Skipping item with missing task_id or question: {item}")
|
81 |
continue
|
82 |
try:
|
|
|
83 |
submitted_answer = agent(question_text)
|
84 |
answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
|
85 |
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
|
|
|
3 |
import requests
|
4 |
import inspect
|
5 |
import pandas as pd
|
6 |
+
from agents import run_orchestrator
|
7 |
# (Keep Constants as is)
|
8 |
# --- Constants ---
|
9 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
|
|
15 |
print("BasicAgent initialized.")
|
16 |
def __call__(self, question: str) -> str:
|
17 |
print(f"Agent received question (first 50 chars): {question[:50]}...")
|
18 |
+
prompt = f"""You are a general AI assistant. I will ask you a question. Each question calls for an answer that is either a string (one or a few words), a number, or a comma separated list of strings or floats, unless specified otherwise. There is only one correct answer. Hence, evaluation is done via quasi exact match between a model’s answer and the ground truth (up to some normalization that is tied to the “type” of the ground truth). Question: {question}"""
|
19 |
+
answer = run_orchestrator(prompt)
|
20 |
+
print(f"Agent returning answer: {answer}")
|
21 |
+
return answer
|
22 |
|
23 |
def run_and_submit_all( profile: gr.OAuthProfile | None):
|
24 |
"""
|
|
|
74 |
results_log = []
|
75 |
answers_payload = []
|
76 |
print(f"Running agent on {len(questions_data)} questions...")
|
77 |
+
for i, item in enumerate(questions_data, 1):
|
78 |
task_id = item.get("task_id")
|
79 |
question_text = item.get("question")
|
80 |
if not task_id or question_text is None:
|
81 |
print(f"Skipping item with missing task_id or question: {item}")
|
82 |
continue
|
83 |
try:
|
84 |
+
print(f"Question {i}:")
|
85 |
submitted_answer = agent(question_text)
|
86 |
answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
|
87 |
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
|