# Regular aeon Developer Meeting (early) - Jul 26, 2024
###### tags: `aeon-regular-dev-meeting`

#### Jul 26, 2024 @ 12:00PM UTC
#### Previous meeting: https://hackmd.io/@aeon-toolkit/rJxdHdvN0

Present: MM AG SS

### Topics

- [SS] `fit`/`predict`/`fit_predict` in anomaly detection
- Currently, `PyODAdapter` does only support _unsupervised_ usage, however, `semi-supervised` requested
- For `unsupervised` should `fit` just ignore input or raise an exception to prevent unexpected behavior
- [#1837](https://github.com/aeon-toolkit/aeon/issues/1837)
- --> change `PyODAdapter` to also support _semi-supervised_ usage, like k-Means
- [SS] Type Hinting
- Some existing hints are incorrect, e.g. `Type[BaseEstimator]` references the class, while `BaseEstimator` references an instance of this class (https://github.com/aeon-toolkit/aeon/blob/16a19bb8aaece13389c5f920c8f574113fe439aa/aeon/classification/sklearn/_rotation_forest_classifier.py#L109)
- Annotating functions with types from soft-dependencies. Should we adopt [PEP 563](https://peps.python.org/pep-0563/)?
- There is no automated type checking
- [MM] Tutorials
- Let us know if anything needs to be included in `aeon` proper before these so we can review and merge
- MM working on EEG dataset for KDD

### Needs decision

N/A

### Needs attention (reviews)

N/A

### Action items

- [x] SS: change `PyODAdapter` to also support _semi-supervised_ usage, like k-Means
- [x] SS: Create issue for type hinting issue `Type[BaseEstimator]`
- [x] MM: documentation on how to add type hints (example for soft dependencies from [PyODAdapter](https://github.com/aeon-toolkit/aeon/blob/main/aeon/anomaly_detection/_pyodadapter.py))
- [X] MM: Create issue for MyPy and creare pre-commit configuration

### Previous Action items

- [x] ML to send MM evaluation example
- [x] MM to set up regular drop-in meetings for students

### Next meeting date

Aug 9th, 2024