bgruening / galaxytools

:microscope::books: Galaxy Tool wrappers
MIT License
115 stars 222 forks source link

Upgrade Galaxy ML to 0.10.0 #1313

Closed anuprulez closed 11 months ago

anuprulez commented 1 year ago

Fix failing tests for almost all ML tools

ping @bgruening @kxk302

bgruening commented 11 months ago

Thanks @anuprulez for working on this!

anuprulez commented 11 months ago

I think all tests passed except for a few where file size > 1MB

tools/sklearn/test-data/ml_vis01.html
tools/sklearn/test-data/ml_vis02.html
tools/sklearn/test-data/ml_vis03.html
tools/sklearn/test-data/ml_vis04.html
tools/sklearn/test-data/ml_vis05.html
tools/sklearn/test-data/model_fit01
tools/sklearn/test-data/searchCV03

These files are more than 1 MB in size but less than 2 MB. The first 5 test-data files contain Plotly plots which are usually bigger. What do you think about these files?

ping @bgruening @kxk302 @paulzierep Thanks!

paulzierep commented 11 months ago

I did close to nothing, wow @anup, thats a lot of work !

bgruening commented 11 months ago

Wow, nice. Lets not care about the output size here. I will deploy those tools manually.

If you are all ok, I will merge and deploy?

paulzierep commented 11 months ago

I think all tests passed except for a few where file size > 1MB

tools/sklearn/test-data/ml_vis01.html
tools/sklearn/test-data/ml_vis02.html
tools/sklearn/test-data/ml_vis03.html
tools/sklearn/test-data/ml_vis04.html
tools/sklearn/test-data/ml_vis05.html
tools/sklearn/test-data/model_fit01
tools/sklearn/test-data/searchCV03

These files are more than 1 MB in size but less than 2 MB. The first 5 test-data files contain Plotly plots which are usually bigger. What do you think about these files?

ping @bgruening @kxk302 @paulzierep Thanks!

the ml_vis*.html tests and model_fit01 only compare the size, we could just check the size, without using the actual file, using has_size or compare with (md5/checksum) ? any preference @anuprulez @bgruening ? the searchCV03 is used as input, that one we would need to reduce.

anuprulez commented 11 months ago

I would love to see this merged :+1:

Once these tools are available, we should verify all ML GTN tutorials with the latest version of these tools.

bgruening commented 11 months ago

I pushed them, hopefully all are uploaded.

I got this error "Repository [stacking_ensemble_models] does not exist in the targeted Tool Shed." .. what tool is that. The ID looks strange. There is a bug somewhere here.

kxk302 commented 11 months ago

I did close to nothing, wow @anup, thats a lot of work !

Same here! Just got back from vacation and was catching up with everything. Thanks a lot @anuprulez!

anuprulez commented 11 months ago

stacking_ensemble_models

It could be this tool [sklearn_stacking_ensemble_models](https://toolshed.g2.bx.psu.edu/repository/browse_repository?id=27b9dc6b09687b74) but in bgruening/galaxytools/tools/sklearn, there is a tool with stacking_ensemble_models ID.

I hope they both are same!