feat: benchmark(5) Benchmark.py串接前端

[FernandeChant]

目標

1. 修改Benchmark.py
2. 串app.py

   預計規劃

3. 修改Benchmark.py: 原本只有產出我們自己的fine_tune_model報告，現在修改為fine_tune_model + raw model
4. 串app.py：在app.py讀取jsonl檔案和fine_tune_model的名稱，並呼叫Benchmark.py的function

預期結果

app.py

`@app.route('/', methods=['GET', 'POST'])
 def BenchmarkReport():
     if request.method == 'POST':
         uploaded_file = request.files['jsonl_file']
         name = request.form['name']

         with NamedTemporaryFile(delete=False, suffix='.jsonl') as temp_file:
             temp_path = temp_file.name
             uploaded_file.save(temp_path)
         # Generate answers using the provided dataset_path and model_name
         dataset_path = temp_path`

Benchmark.py

`def generateAnswers(dataset_path, model_names):
    questions = []
    model_answers = {model_name: [] for model_name in model_names}
    ideal_answers = []

......
def evalWithRubric(test_set, model_names):
    rubric_evaluations = {model_name: [] for model_name in model_names}
    questions = test_set['Question']
    model_answers = test_set['ModelAnswers']
    ideal_answers = test_set['IdealAnswer']`

[FernandeChant]

已解決+PR