extra-model Blog

[22-10-21] Simplifying usage of extra-model

During the month of October it's always a good idea to clean some cobwebs that has been around for a while. In the spirit of Hacktoberfest we've decided to clean up couple of issues that simplify how extra-model is called.

Before, all, but one parameter (--debug) that you could provide to CLI version of extra-model were arguments, which in click parlance meant that they were positional. That in turn meant that you could only change, e.g., embeddings path by first specifying both output path AND output filename. Well, no longer! From now on all, but one, inputs to extra-model (input path), are options which means that they can be set in any order using flags that you can see by running extra-model --help. No need to thank us :).

Another tiny UX improvement is that we check the names of the columns BEFORE running extra-model. With big inputs running extra-model can take hours, so it was a subpar user experience when it would fail after all this time because CommentId column is misspeled. Now it'll fail with descriptive error within seconds.

[18-05-21] Simplifying install of extra-model

We have updated/removed all the dependencies that don't have wheels. In addition, we've switched to Python 3.8 as a base image (we are still testing extra-model in 3.7, 3.8 and 3.9).

Full list of changes is available in the CHANGELOG

[17-03-21] Blog Setup!

extra-model v0.2.0 is released!