Implement ROC, PR, Cost curves
Nobody has claimed this yet.
- Dominant language
- PHP
- Stars
- 755
- Forks
- 128
- PR merge metrics
- No merged PRs in 30d
Description
Vishal volunteered to implement ROC curves on the run pages. I'm opening this issue to keep track of progress, and invite others to join.
The ViperCharts lib is ideal for doing this: http://viper.ijs.si/about/
We had an earlier implementation but this is not supported anymore:
https://github.com/openml/OpenML/blob/master/openml_OS/views/pages/frontend/r/subpage/run.php#L194
One important difference is that previously, we used the ViperCharts API (running in Ljubljana) to generate the plots and then pull them in. This time around, I would run ViperCharts on our server, as this is one dependency less to worry about.
Note: This should ideally be done in a JS-only way: get the data via ElasticSearch and push them to the ViperCharts library, so plots are created on the fly. I want to avoid PHP dependencies since the frontend will very likely be rewritten in Django+React. Just tell me what data you need for generating the plots, and I will add them to ElasticSearch.
Here is an example for doing JS-only plots for datasets:
https://github.com/openml/OpenML/blob/master/openml_OS/views/pages/frontend/d/javascript.php
The current JS for the run page contains a lot of legacy code, which I'll clean up asap:
https://github.com/openml/OpenML/blob/master/openml_OS/views/pages/frontend/r/javascript.php
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with openml_OS/views/pages/frontend/r/javascript.php and the dataset example in openml_OS/views/pages/frontend/d/javascript.php, then review the earlier run-page implementation. Determine the Elasticsearch data needed for ViperCharts and how to run the library on the server. Done means ROC, PR, and cost curves render on run pages through JavaScript without a PHP dependency.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- javascript
- Domain
- data-visualization, frontend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100