Closed followingell closed 2 years ago
Just checking, is anyone available to review this please @sweco, @mrshu?
Is there anything that I can add to the PR to make the review process easier for yourselves?
Thanks for the PR @followingell -- I'll take a closer look at it later today 🙂
@followingell Is there any specific reason for
PurePath
here as opposed to usingPath
that has already been in use?
@mrshu Not sure where the above comment has gone(?), regardless responding here rather than via email:
I think this StackOverflow answer sums it up quite nicely. Essentially, PurePath
just performs string-like operations whereas Path
can also do I/O operations which we don't need here. As such, I chose to utilise the simpler, parent class.
If you'd rather I just use Path
then I'm happy to do so.
@mrshu FYI for now I have finished with changes. As such, ready for review 👍
Thanks again @followingell , this was now released in 0.3.0
!
Issue
Summary
I added the ability to perform
compute
commands on only a subset of tables by adding a--model-path-filter
option. This means that a subset of models can be selected based upon theiroriginal_file_path
value (taken from themanifest.json
artifact).This functionality means that
dbt-coverage
can now be used in monolithic dbt projects which contain sub-projects owned by different teams. Before adding model selection functionality, usingdbt-coverage
would not have been useful/advisable in such a structure because another, unrelated team may decrease the overall coverage, which can then block PR merging (shoulddbt-coverage
have been integrated as part of a CI/CD pipeline for example).README.md
below:Note: this is a relatively 'rough' solution and there are likely many improvements that could be made to my code / far more elegant implementations that would achieve the same functionality. Please, feel free to suggest changes!
Testing
jaffle_shop
'testing project' and have not encountered issues so far.