Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Simplify TaggedTable #647

Closed
lars-reimann opened this issue Apr 24, 2024 · 3 comments · Fixed by #680
Closed

Simplify TaggedTable #647

lars-reimann opened this issue Apr 24, 2024 · 3 comments · Fixed by #680
Assignees
@lars-reimann
Labels
enhancement 💡 New feature or request released Included in a release

Comments

@lars-reimann
Copy link
Member

Is your feature request related to a problem?

We've already spent a lot of time overriding Table methods in TaggedTable just to keep the tagging. There are also more issues waiting, e.g. #644. Moreover, we also want to connect other types of input and output data, e.g. images to a table, so TaggedTable is just a special case of a more abstract Dataset concept.

Desired solution

  • Don't inherit Table in TaggedTable.
  • Get rid of all the overridden Table methods.
  • Table.tagColumn can stay. It's now the final step before passing the data to an ML model.
  • Regressor.fit and Classifier.fit could still return a TaggedTable or maybe just a Table?

Possible alternatives (optional)

No response

Screenshots (optional)

No response

Additional Context (optional)

No response
@lars-reimann lars-reimann added the enhancement 💡 New feature or request label Apr 24, 2024
@github-project-automation github-project-automation bot moved this to Backlog in Library Apr 24, 2024
This was referenced Apr 24, 2024
Closed
Closed
@lars-reimann
Copy link
Member Author

There's also a strong correlation between TaggedTable and the InputConversionTable. Both expect names of features and the name of the target.

@Gerhardsa0
Copy link
Contributor

There's also a strong correlation between TaggedTable and the InputConversionTable. Both expect names of features and the name of the target.

I aggree, also with TimeSeries and test:inout_conversion_time_series. Maybe we should let the NN detect the type and then cast, because a lot of information is stored there.
I will make a prototype in my PR #615

@lars-reimann lars-reimann self-assigned this May 1, 2024
@lars-reimann lars-reimann mentioned this issue May 1, 2024
Merged
@lars-reimann lars-reimann linked a pull request May 1, 2024 that will close this issue
feat: rework TaggedTable #680
Merged
lars-reimann added a commit that referenced this issue May 1, 2024
@lars-reimann @megalinter-bot
feat: rework TaggedTable (#680)
db2b613 
Closes #647

### Summary of Changes

* `TaggedTable` is now called `TabularDataset`,
* It is moved from `safeds.data.tabular.containers` to
`safeds.data.labeled.containers`. That's where all dataset classes for
supervised learning will go, like the upcoming `ImageDataset`.
* `TabularDataset` no longer inherits from `Table`.
* `TabularDataset` now has a very small interface. It's only meant to be
used as input for supervised ML models. Table manipulation is now solely
done using the `Table` class.
* `tag_columns` on `Table` is now called `to_tabular_dataset`. This
makes it consistent with other conversion methods and emphasizes that
this is a terminal operation and should only be used once one is done
manipulating the table.
* `TabularDataset` now has a public `to_table` method to get a `Table`
again.

---------

Co-authored-by: megalinter-bot <[email protected]>
@github-project-automation github-project-automation bot moved this from Backlog to ✔️ Done in Library May 1, 2024
lars-reimann pushed a commit that referenced this issue May 1, 2024
@semantic-release-bot
chore(release): 0.22.0 [skip ci]
43717e0 
## [0.22.0](v0.21.0...v0.22.0) (2024-05-01)

### Features

* `is_fitted` is now always a property ([#662](#662)) ([b1db881](b1db881)), closes [#586](#586)
* add `Column.missing_value_count` ([#682](#682)) ([f084916](f084916)), closes [#642](#642)
* Add `InputConversion` & `OutputConversion` for nn interface ([#625](#625)) ([fd723f7](fd723f7)), closes [#621](#621)
* Add hash,eq and sizeof in ForwardLayer ([#634](#634)) ([72f7fde](72f7fde)), closes [#633](#633)
* allow using tables that already contain target for prediction ([#687](#687)) ([e9f1cfb](e9f1cfb)), closes [#636](#636)
* callback `Row.sort_columns` takes four parameters instead of two tuples ([#683](#683)) ([9c3e3de](9c3e3de)), closes [#584](#584)
* rename `group_rows_by` in `Table` to `group_rows` ([#661](#661)) ([c1644b7](c1644b7)), closes [#611](#611)
* rename `number_of_column` in `Row` to `number_of_columns` ([#660](#660)) ([0a08296](0a08296)), closes [#646](#646)
* rework `TaggedTable` ([#680](#680)) ([db2b613](db2b613)), closes [#647](#647)
* show missing value count/ratio in summarized statistics ([#684](#684)) ([74b8a35](74b8a35)), closes [#619](#619)
* specify `extras` instead of `features` in `to_tabular_dataset` ([#685](#685)) ([841657f](841657f)), closes [#623](#623)

### Bug Fixes

* actually use `kernel` of support vector machines for training ([#681](#681)) ([09c5082](09c5082)), closes [#602](#602)

### Performance Improvements

* Faster plot_histograms and more reliable plots ([#659](#659)) ([b5f0a12](b5f0a12))
@lars-reimann
Copy link
Member Author

🎉 This issue has been resolved in version 0.22.0 🎉

The release is available on:

Your semantic-release bot 📦🚀

@lars-reimann lars-reimann added the released Included in a release label May 1, 2024
@lars-reimann lars-reimann mentioned this issue May 5, 2024
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@lars-reimann lars-reimann

Labels
enhancement 💡 New feature or request released Included in a release
Projects
Library
Archived in project
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

feat: rework TaggedTable Safe-DS/Library
2 participants
@lars-reimann @Gerhardsa0