First, I've been working with pyf for few days.
It is a very very very good tool yet a little young.
BTW, from the user point of view, it would be insanely great if :
- In the CSV Writer, (and any other file writer) We could choose to
name the files in the storage along the output Name defined in the
tube, or not (some may feel comfortable with those automatic names).
In the first case, existing files could be either left in place or
overwritten. In case of keeping the existing file alive, new file
could be selectively pre/post fixed by a given sequence.
-In any file writer, it would be useful to take subfolders into
account. They would be silently created if not existing yet.
- Submitted tubes could be cancelled or stopped, especially for
handling bugs and acknowledging failure rather than having lots of
pending tubes in the events log we don't know if they are really
pending or crashed.