Suzen Fylke

I'd like to do the Swedish translation.

I was planning on working on it over the weekend. Is that okay? If you need it sooner you can reassign it.

Yes; if a component (operator) fails, Airflow can retry it: https://airflow.apache.org/docs/1.10.6/_api/airflow/operators/index.html?highlight=retry#package-contents

Yes, we're using InteractiveContext for local debugging. The use case is running components one by one as you iterate on them before running a pipeline end-to-end using LocalDagRunner then KubeflowDagRunner...

Changing to `beam_pipeline_args = list(beam_pipeline_args or self.beam_pipeline_args)` fixes this. There are no changes to the original list and the labels are well-formed now.

Hi @singhniraj08, I have a workaround, but it would be nice to have this fix merged. Should I submit a pull request for this?

Hi @singhniraj08, I just opened a PR with the fix.

There's an issue in tfx about converting BulkInferrer outputs to CSV for better visualization: https://github.com/tensorflow/tfx/issues/2478. It's not completely relevant to this work, but it might be if this project will...

+1 There's a [TODO](https://github.com/tensorflow/tfx/blob/master/tfx/components/example_gen/csv_example_gen/executor.py#L158) to consider allowing users to configure parsing parameters, and the [function used](https://github.com/tensorflow/tfx/blob/master/tfx/components/example_gen/csv_example_gen/executor.py#L204) already allows passing a delimiter, so making this change is mostly a matter of...

@UsharaniPagadala This issue isn't resolved yet. The delimiter is still hard-coded: https://github.com/tensorflow/tfx/blob/master/tfx/components/example_gen/csv_example_gen/executor.py#L199