Currently CC-0 which is really bad for code.

@drseb reported:

When evaluating the E4 error, it would be good to know how many tabs were in the header compared to the row.

To integrate this with CI tools, right now, I have to manually deal with pipeline failure (if tsvalid prints nothing, its fine, else its fail). Not sure what is best here, but i think returning 1 is not a bad idea.

This is a really useful lib!

@hrshdhgd -- we could really have used this for our NER project with Mayo where we kept getting these broken TSVs that messed things up.

I am starting to use csvkit more and more - I am wondering if it makes sense to donate some functionality there, or if this is more specific to our use cases:

https://github.com/wireservice/csvkit

There are various others like mlr etc, but I am tending to coalesce around csvkit

[no response expected for some time and ok to close as won't-do]

@jamesaoverton

You can start from the README.md, and then take a bit of a look around if you have time. I think it would be good if you could look at the project organisation a little bit as well.

• I have done a huge deal of work on the QC stuff, black, flake8, code cleanliness etc.
• The encoding error check does not do what we thought it would (which is pointing out bad characters) - turns out you can stick anything into a utf-8 encoded file. We probably need to iterate over this a bit, but do have some tests in place.
• Let's use this for a while now, and gather issues. Adding checks should be really easy at this point.

You can close this issue once you are happy with the general organisation of the code, and then we should start using the tool in our day to day work and start filing bug reports.