Spaces:
Sleeping
Sleeping
File size: 1,698 Bytes
324d83a f02aeda 324d83a f02aeda 324d83a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 |
import json
from openai import OpenAI
from tests.testing_prompts import grader_prompt
def grade(json_file_path, model="gpt-4-turbo"):
client = OpenAI(base_url="https://api.openai.com/v1")
with open(json_file_path) as file:
interview_data = json.load(file)
messages = [
{"role": "system", "content": grader_prompt},
{"role": "user", "content": f"Interview data: {interview_data}"},
{"role": "user", "content": "Please evaluate the interview."},
]
response = client.chat.completions.create(model=model, messages=messages, temperature=1, response_format={"type": "json_object"})
feedback = json.loads(response.choices[0].message.content)
feedback["file_name"] = json_file_path
feedback["agent_llm"] = interview_data["interviewer_llm"]
feedback["candidate_llm"] = interview_data["candidate_llm"]
feedback["type"] = interview_data["inputs"]["interview_type"]
feedback["difficulty"] = interview_data["inputs"]["difficulty"]
feedback["topic"] = interview_data["inputs"]["topic"]
feedback["average_response_time_seconds"] = interview_data["average_response_time_seconds"]
feedback["number_of_messages"] = len(interview_data["transcript"])
scores = [
feedback[x]
for x in feedback
if (x.startswith("interviewer_") or x.startswith("feedback_") or x.startswith("problem_")) and feedback[x] is not None
]
feedback["overall_score"] = sum(scores) / len(scores)
# save results to json file in the same folder as the interview data
with open(json_file_path.replace(".json", "_feedback.json"), "w") as file:
json.dump(feedback, file, indent=4)
return feedback
|