WandB integration: from spacy.cli.train import train doesn't work

There is no train function in spacy.cli.train module.

Hi I am running train-search script as described here for the WandB integration https://github.com/explosion/projects/tree/v3/integrations/wandb

import typer
from pathlib import Path
from spacy.cli.train import train


def main(config_dir: Path, default_config: str, results_dir: Path):
    """Run all config files instead of the default one.
    Ideally, these runs are parellellized instead of run in sequence."""
    for config_file in config_dir.iterdir():
        if config_file.name != default_config:
            output_path = results_dir / config_file.stem
            if not output_path.exists():
                output_path.mkdir()
            train(config_path=config_file, output_path=output_path)


if __name__ == "__main__":
    typer.run(main)

I am getting the following error:

TypeError: train() got an unexpected keyword argument ‘config_path’

I suspect that the error is due to the fact that there is no train function in spacy.cli.train. So I changed this line

from spacy.cli.train import train to

from spacy.cli.train import train_cli and changed the train() to train_cli in the script.

which now gives the following error:

TypeError: train_cli() missing 1 required positional argument: ‘ctx’

Any idea how to make train-search work?

Thank you!

Environment

• spaCy version: 3.0.0rc2
• Platform: Darwin-20.1.0-x86_64-i386-64bit
• Python version: 3.7.9
• Pipelines: en_core_web_trf (3.0.0a0)

[svlandeg]

This should be fixed by https://github.com/explosion/projects/pull/24!

[scellat]

Nice!!! Thank you for the update!