GCSDirectoryLoader needs a way to exclude files or file patterns

[ejstembler]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

If GCSDirectoryLoader is processing a bucket and there's a file is doesn't support, it throws an exception. For example:

The FileType.ZIP file type is not supported in partition.

It would be better if it could be told to skip specified files or file types. Or, alternatively maintain a list of files it skipped so the developer can review afterwards. Or, expose a callback for the developer to handle these cases at runtime.

Motivation

Error while processing files.

Proposal (If applicable)

New parameter:

exclude_file_types: Optional[List[str]] = None

Or

exclude_files: Optional[List[str]] = None

Or new property

def excluded_files(self) -> Optional[List[str]]:
    return self._excluded_files

Or new callback

# filename: str
OnFileSkippedCallback = Callable[[str], None]

[ejstembler]

Another error I just encountered: Workbook is encrypted