Onecodex

Changed

- Downgrades `notebook` version pin to 6.4.10

Added

- Adds `coerce_haxis_dates=True` parameter to `plot_metadata()`

Fixed

- Fixes date field type coercion in `plot_metadata()`

Added

- Adds new plotting method `plot_functional_heatmap` to plot a heatmap with specified annotations and metrics of functional profile data

Changed

- Pins `urllib3<2`
- Updates `jupyter-client` version pin to 8.6.0
- Updates `notebook` version pin to 7.0.6
- Updates `nbconvert` minimum version to 6.4.3
- Updates `WeasyPrint` version pin to 60.1
- Removes `jinja2` version pin
- Replaces `imp` module usage with `importlib` in preparation for Python 3.12 support

Fixed

- Fixes `FutureWarning`s generated by pandas, scikit-learn, etc.
- `make test` no longer generates warnings
- Fixes notebook exporter commands in `README.md`

Added

- Adds `name` option to `assets upload` command in experimental API mode

Fixed

- Fixes heatmap plotting for samples without a specified `haxis` value
- Fixes installation by monkeypatching altair-saver for npm >=9

Added

- Adds support for Python 3.11
- Adds `assets upload` command in experimental API mode

Changed

- Handles heatmap display of samples with no abundances calculated
- Excludes samples with no abundances calculated from distance heatmap with abundances metric
- Pins altair to 4.2.2
- Requires jsonschema version >= 3.0
- Improves retry handling
- Pins version of altair-saver that includes a fix for npm >= 9

Fixed

- Fixes unnecessary global loading of numpy and pandas
- Fixes chart concatenation

Removed

- Removes support for Python 3.7
- Removes opening links in new tab functionality (can be added manually to a returned chart with `usermeta`)

Added

- Adds support for testing experimental API
- Adds functionality to `SampleCollection` objects to generate tabular data for `FunctionalRun`s using the `.to_df()` method
- Adds `include_taxa_missing_rank` parameter to `to_df()` method for including taxa that do not have a designated parent at `rank` (will be grouped into a "No <rank>" column) (metrics `readcount_w_children` and `abundance_w_children` are supported)
- Adds support for `readcount_w_children` metric to `plot_bargraph()` when passing `include_taxa_missing_rank` (only `abundance_w_children` was supported previously)
- Adds increased customization of plot legend; `plot_bargraph()`'s `legend` parameter now accepts an `altair.Legend` instance
- Adds `group_by` support to `plot_bargraph`
- Adds support for linking to NCBI taxonomy browser from plots

Changed

- Pins selenium to <4.3.0
- Pins scipy<1.9
- Default plot legend is bigger (displaying up to 40 items)

Fixed

- Fixes metadata fetching to support fields that match neither metadata nor taxa instead of raising an error
- Fixes box plot display bug for samples with same value
- Fixes pandas `SettingWithCopyWarning` previously raised by `to_df` and `plot_bargraph`
- Fixes `plot_bargraph` display bug for empty samples (total count or abundance of zero)
- Fixes `None` artifact in PDFs