Input parameters

[martinezmatias]

Which are the input arguments that the pipeline would receive?

It comes to my mind:

• dataset
• filter
• model
• query
• evaluationmetric

For each of them, we should put a command line argument. (see issue #9 )

Should we include other input parameters?

[tipeXwins]

Depending on how we define the parameters, this might be considered sufficient. I have doubts about how to define the parameters, particularly when it comes to defining model parameters. We have two options:

1. We can use integer parameters and assign numbers to corresponding models. For example, 1 represents OpenAI, 2 represents HuggingFaceCodeGen, and 3 represents HGIncoder.

2. Alternatively, we can use string parameters and specify the full name of the model to match it. For instance, we can use "OpenAI" to refer to the OpenAI model, "HuggingFaceCodeGen" for the HuggingFaceCodeGen model, and "HGIncoder" for the HGIncoder model.

3. Or there is a better approach?