.robocop not picked up when using API

When using the API as:

    import robocop
    from robocop.config import Config
    from robocop.utils import issues_to_lsp_diagnostic

    config = Config(root=project_root)
    robocop_runner = robocop.Robocop(config=config)
    robocop_runner.reload_config()

    issues = robocop_runner.run_check(ast_model, filename, source)
    diag_issues = issues_to_lsp_diagnostic(issues)
    return diag_issues

The .robocop file is not being loaded based on the project_root.

I debugged a bit and it seems like load_default_config_file is never called. I tried to manually call: config.parse_opts([]), right after creating the Config and it seems to pick things up on the debugger, but I still couldn’t get it to ignore what I expected.

I tested with a file with:

*** Test Cases ***
Test
    Fail

and a .robocop file with:

--exclude missing-doc-testcase
--include missing-doc-suite

But the missing doc in the testcase is still reported.

So, I’m waiting on some pointers on what I’m doing wrong here…

[bhirsz]

Configuration (when calling parse_opts and .robocop existing) is loaded correctly. The problem is that we’re ‘filtering out’ some if the issues - based on config - when reporting the issues. Since we’re not reporting the issues in self.run method it will leave everything (instead of reporting them we return list of issues, that is converted to lsp diagnostic later on). I will rearrange our architecture.

Also if the root is not the Path it will throw an error (I saw you made workaround for that). I will fix that as well.

[bhirsz]

For argument parsing issues I had to create workaround - because argparse that we’re using in Robocp always returns SystemExit. After my changes, if you’re using Robocop API and you have invalid configuration it should return “InvalidArgumentError” exception (found in robocop.exceptions).