[IOTDB-5728] Implement config parser & model/dataset factory on MLNode

[lichenyu20]

Implement mlnode parser

parse training request of type TCreateTrainingTaskReq
Usage:

    modelId = 'mid_etth1_dlinear_default'
    isAuto = False
    modelConfigs = {
        'task_class': 'forecast_training_task',
        'source_type': 'thrift',
        'dataset_type': 'window',
        'time_embed': 'h',
        'input_len': 96,
        'pred_len': 96,
        'model_name': 'dlinear',
        'input_vars': 7,
        'output_vars': 7,
        'task_type': 'm',
        'kernel_size': 25,
        'learning_rate': 1e-3,
        'batch_size': 32,
        'num_workers': 0,
        'epochs': 10,
        'metric_names': ['MSE', 'MAE']
    }
    queryExpressions = ['root.eg.etth1.**', 'root.eg.etth1.**', 'root.eg.etth1.**']
    queryFilter = '0,1501516800000'
    req = TCreateTrainingTaskReq(
        modelId=str(modelId),
        isAuto=isAuto,
        modelConfigs={k: str(v) for k, v in modelConfigs.items()},
        queryExpressions=[str(query) for query in queryExpressions],
        queryFilter=str(queryFilter),
    )
    data_conf, model_conf, task_conf = parse_training_request(req)