metric.py

import os
from openai import OpenAI
if "OPENAI" in os.environ:
    pass
else:
    print('Doesn\'t find OPENAI')
client = OpenAI(api_key = os.environ['OPENAI'])

import pandas as pd
from huggingface_hub import hf_hub_download

def compute(params):
    public_score = 0
    private_score = 0

    solution_file = hf_hub_download(
        repo_id=params.competition_id,
        filename="solution.csv",
        token=params.token,
        repo_type="dataset",
    )

    solution_df = pd.read_csv(solution_file)

    submission_filename = f"submissions/{params.team_id}-{params.submission_id}.csv"
    submission_file = hf_hub_download(
        repo_id=params.competition_id,
        filename=submission_filename,
        token=params.token,
        repo_type="dataset",
    )
    submission_df = pd.read_csv(submission_file)

    public_ids = solution_df[solution_df.split == "public"][params.submission_id_col].values
    private_ids = solution_df[solution_df.split == "private"][params.submission_id_col].values

    public_solution_df = solution_df[solution_df[params.submission_id_col].isin(public_ids)]
    public_submission_df = submission_df[submission_df[params.submission_id_col].isin(public_ids)]

    private_solution_df = solution_df[solution_df[params.submission_id_col].isin(private_ids)]
    private_submission_df = submission_df[submission_df[params.submission_id_col].isin(private_ids)]

    public_solution_df = public_solution_df.sort_values(params.submission_id_col).reset_index(drop=True)
    public_submission_df = public_submission_df.sort_values(params.submission_id_col).reset_index(drop=True)

    private_solution_df = private_solution_df.sort_values(params.submission_id_col).reset_index(drop=True)
    private_submission_df = private_submission_df.sort_values(params.submission_id_col).reset_index(drop=True)



    # # METRICS Calculation Evaluation
    # # _metric = SOME METRIC FUNCTION
    # def _metric(outputs, targets):
    #     # input example: public_solution_df[target_cols], public_submission_df[target_cols]
        
    #     score = 0.5
    #     return score

    
    print('public_solution_df', public_solution_df)
    print('private_solution_df', private_solution_df)
    
    ## LLM Scoring Evaluation
    def _metric(outputs, targets):
        # inputs: public_solution_df[target_cols], public_submission_df[target_cols]
        # output: score
        for row, output in outputs.iterrows():
            print('outputs type', type(outputs), 'targets type', type(outputs))
            print('output', output)
            answer = output['pred']
            label = str(targets.iloc[row]['pred'])
            print('answer', answer)
            
            prompt=f"Give me a score from 1 to 10 (higher is better) judging how similar these two captions are. Caption one: {answer}. Caption two: {label}\nScore:"
            
            try:
                response = client.completions.create(
                    engine="gpt-3.5-turbo-instruct",
                    prompt=prompt,
                    temperature=0,
                    max_tokens=1,
                )
                eval_result = response.choices[0].text.strip()
                print('eval_result', eval_result)
                score = int(eval_result)
                
            except:
                print("Error: API Calling")
                return

        return score

    target_cols = [col for col in solution_df.columns if col not in [params.submission_id_col, "split"]]
    public_score = _metric(public_solution_df[target_cols], public_submission_df[target_cols])
    private_score = _metric(private_solution_df[target_cols], private_submission_df[target_cols])

    metric_name = "metric1"
    
    metric_dict = {"public_score": {metric_name: public_score},
                   "private_score": {metric_name: private_score}
                   }

    return metric_dict