Skip to content

CHANGELOG.md

Latest commit

 

History

History
441 lines (301 loc) · 21.9 KB

CHANGELOG.md

File metadata and controls

441 lines (301 loc) · 21.9 KB
Raw

Unreleased

  • Fix: docstrings of height(), mouse() and to_parquet() did not match the code (#265)

v1.0.1

  • Fix: add missing bundle.css to wheel

v1.0.0

  • Feat: add order() for determining the draw order of points (#190 via #237)
  • Feat: use categorical color encoding in histogram bars (#97)
  • Refactor: move toolbar UI from Python ipywidgets to TS React. This enables full Marimo support 🎉
  • Chore: add straight option to connect() for straight line connections (#130)
  • Chore: verify Python 3.14 support (label placement not yet available due to geoindex-rs lacking 3.14 wheels)
  • Fix: Unused categories cause wrong colors/legend (#189 via #233)
  • Fix: Wrong coloring when switching color_by between categorical/continuous (#191 via #233)
  • Fix: Exception when updating data with changed categories (#197 via #233)

v0.22.3

  • Fix: geoindex-rs import compatibility (#222)

v0.22.2

  • Fix: a bug related to installation with default dependencies (#204)

v0.22.1

  • Fix: make optional package import more robust (#199)

v0.22.0

  • Feat: add scatter.label() function for displaying text labels
  • Feat: addLabelPlacement class that allows to statically-resolved label placement and preserve results
  • Feat: add new color helper methods: brighten, darken, saturate, desaturate
  • Feat: add 2D color map support via pycolormap-2d
  • Refactor: Make seaborn dependency for contour annotations optional. You can install it via pip install "jupyter-scatter[annotation-extras]" or pip install "jupyter-scatter[all]"

v0.21.1

  • Feat: automatically assign gray to undefined/NA values for categorical data (#186)
  • Fix: properly undefined/NA values in categorical data (#185)
  • Fix: tooltip issue with non-existing tooltip (#187)
  • Fix: ensure tooltip_histogram == False works as expected
  • Fix: ensure a static opacity is not shown in the tooltip

v0.21.0

  • Feat: add support for brush and rectangle lasso types.
  • Feat: expose lasso selection polygon via scatter.widget.lasso_polygon_selection
  • Feat: enable merging of lasso selection upon holding down the meta key and deselected points upon holding down the alt key.
  • Feat: allow specifying which properties should feature a histogram via scatter.tooltip(histograms=['color', 'property_a'])
  • Fix: background histogram render bug
  • Fix: ensure NAs are skipped for the histogram
  • Fix: ensure that the tooltip is hidden on lasso start
  • Fix: ensure that the color, opacity, and size order is reset when one switches from a categorical to a continuous color map. The reset is skipped when order == 'reverse'.

v0.20.0

  • Feat: add support for showing a scatter plot in full-screen mode. In full-screen mode you can easily customize and up-scale the resolution for better exporting in PNG.
  • Feat: add ability to change the point size zoom scale function via scatter.size(scale_function='asinh')
  • Feat: add scatter.camera(is_fixed=True) to allow fixing the camera (i.e., disable pan/zoom) #176
  • Feat: add CLI for quick-starting a demo via uvx jupyter-scatter demo
  • Fix: properly alpha blend when saving a view as PNG
  • Chore: switch from ESLint+Prettier to Biome #170
  • Chore: migrate from hatch to uv #169
  • Chore: bump minimum Python version to 3.9

v0.19.1

  • Fix: revert back to pd.api.types.is_string_dtype from pd.StringDType because the two function, albeit looking semantically identical, do not do the same thing. The former detects object and string types while the latter only strictly detects string types. This is confusing as pd.Series(['a', 'b']) is of type object. So pd.StringDtype.is_dtype(pd.Series(['a', 'b'])) == False but pd.api.types.is_string_dtype(pd.Series(['a', 'b'])) == True. Wat?!

v0.19.0

  • Feat: add contour line annotations #163
  • Refactor: rename the reset_view argument of scatter.data() to zoom_view for clarity #162
  • Docs: add API docs for reset_scales, zoom_view, and animate of scatter.data()
  • Fix: don't use external view domain (only the view) #161

v0.18.1

  • Fix: re-enable point transition and make it explicit via scatter.options(transition_points=True, transition_points_duration=3000)
  • Fix: Correctly render axes grid lines upon resizing
  • Fix: Reapply point color map when setting color_by
  • Fix: Do not set tooltip titles if the corresponding elements are undefined.

v0.18.0

  • Feat: add support for line-based annotations via scatter.annotations()
  • Feat: the exported and saved images now include a background color instead of a transparent background. You can still still export or save images with a transparent background by holding down the ALT key while clicking on the download or camera button.
  • Refactor: When saving the current view as an image via the camera button on the left side bar, the image gets saved in scatter.widget.view_data as a 3D Numpy array (shape: [height, width, 4]) instead of a 1D Numpy array. Since the shape is now encoded by the 3D numpy array, scatter.widget.view_shape is no longer needed and is removed.
  • Fix: hide button for activating rotate mouse mode as the rotation did not work (which is easily fixable) and should not be available when axes are shown as the axes are not rotateable. However rotating the plot without rotating the axis leads to incorrect tick marks.
  • Fix: VSCode integration by updating regl-scatterplot to v1.10.4 (#37)

v0.17.2

v0.17.1

  • Fix: regression preventing tooltip from showing up (#141)

v0.17.0

  • Feat: add scatter.show_tooltip(point_idx)
  • Fix: reset scale & norm ranges upon updating the data via scatter.data()
  • Fix: ensure scatter.axes(labels=['x_label', 'y_label']) works properly (#137)

v0.16.1

  • Fix: preserve filter state upon changing visual encoding (#134)

v0.16.0

BREAKING CHANGES:

The following list of helper widgets for configuring the scatter are removed from the scatter.widget as they are unmaintained, undocumented, and incomplete. If you relied on any of those UI widgets, please see jscatter/widget.py v0.15.0 on how they were created.

  • scatter.widget.mouse_mode_widget
  • scatter.widget.lasso_initiator_widget
  • scatter.widget.selection_widget
  • scatter.widget.hovering_widget
  • scatter.widget.color_widgets
  • scatter.widget.color_by_widget
  • scatter.widget.color_map_widget
  • scatter.widget.height_widget
  • scatter.widget.background_color_widget
  • scatter.widget.background_image_widget
  • scatter.widget.lasso_color_widget
  • scatter.widget.lasso_min_delay_widget
  • scatter.widget.lasso_min_dist_widget
  • scatter.widget.color_widget
  • scatter.widget.color_selected_widget
  • scatter.widget.color_hover_widget
  • scatter.widget.opacity_widget
  • scatter.widget.selection_outline_width_widget
  • scatter.widget.size_widget
  • scatter.widget.selection_size_addition_widget
  • scatter.widget.reticle_widget
  • scatter.widget.reticle_color_widget
  • scatter.widget.download_view_widget
  • scatter.widget.save_view_widget
  • scatter.widget.reset_view_widget

Additionally, the following helper methods are removed as they are unnecessary.

  • scatter.widget.options() (simply listed out all above removed widgets)
  • scatter.widget.select() (same as scatter.selection = list_of_point_indices)
  • scatter.widget.use_cmap() (same as passing the cmap name to scatter.color(map=cmap_name))

Other Changes:

  • Feat: Add basic support for x/y time scale via Scatter(data=df, x='x', x_scale='time', y='y', y_scale='time')
  • Docs: Add API documentation for scatter.widget
  • Docs: Add description for x/y scales
  • Docs: Add description for connected scatterplots
  • Fix: Match numerical and string IDs properly in compose(match_by='XYZ')
  • Fix: Ensure that the domain and histograms match by avoiding missing categorical indices
  • Fix: Ignore NaNs when computing histograms
  • Fix: Warn when data contains NaNs and replace them with zeros
  • Fix: Show correctly ordered color encoding in legend
  • Fix: Ensure the widget's x and y scale domains are updated properly
  • Fix: Ensure the widget's color, opacity, and size titles are updated properly
  • Fix: Ensure the widget's axes titles are updated properly
  • Fix: Include normalization in data dimension name
  • Fix: Allow rendering a single axis instead of enforcing either none or both axis
  • Fix: Rely on pre-normalized data to get bin ID
  • Fix: Connect order
  • Fix: X/Y scale domain bug
  • Fix: Connected point bugs

v0.15.1

  • Fix: Remove an unused widget property that causes an issue with newer version of anywidget (#117)

v0.15.0

  • Feat: Add support for histograms in the tooltip (#96)
  • Feat: Add support for non-visualized properties in the tooltip (#96)
  • Fix: Allow mixing custom and DataFrame-based data (#89)
  • Fix: Improve the tooltip positioning to avoid the tooltip being cut off unnecessarily
  • Fix: Properly redraw axes on resize (#108)
  • Fix: Incorrect axes scale domain (#107)
  • Fix: Use custom regl-scatterplot option on creating a Scatter instance (#106)
  • Fix: Broken link to properties in docstrings (#110)

v0.14.3

  • Fix: don't return color, opacity, or size settings when labeling is defined
  • Fix: prevent x-axis label to be cut off at the bottom
  • Fix: axes label color in dark mode

v0.14.2

  • Fix view synchronization when axes are not shown
  • Fix y-padding size determination
  • Fix stale channel value getter for the tooltip

v0.14.1

  • Fix: update color, opacity, and size scales as the domains update
  • Fix: auto-reset x and y scale domains upon updating the x and y data
  • Fix: use better number formatter for the legend

v0.14.0

  • Add the ability to show a tooltip upon hovering over a point via scatter.tooltip(true) (#86)
  • Fix axes updating of linked scatter plots when panning and zooming (#87)
  • Fix missing x-axes of linked scatter plots (#84)
  • Fix a type in the return value of scatter.xy()

v0.13.1

  • Fix: Prevent resetting the filter upon color, size, or opacity changes
  • Fix: Upon changing the associated data frame via scatter.data(new_df), reapply color, size, and opacity settings

v0.13.0

  • Add ability to specify titles when composing multiple scatter plots by passing tuples of (Scatter, title) to compose() or link()
  • Fix: Add docstrings to compose() and link()
  • Fix: Optimize height of the legend
  • Fix: Check if axes are enabled before updating them when the x or y scale changes
  • Fix: Merge point selections on SHIFT instead of activating the lasso as SHIFT interferes with Jupyter Lab
  • Fix: Allow to call scatter.zoomTo() with the same points multiple times
  • Fix: Unfilter when calling scatter.filter(None)
  • Fix: Properly listen to changes when setting custom regl-scatterplot options via scatter.options()

v0.12.6

  • Fix distributed build by ensuring that jscatter/bundle.js is included in the build
  • Fix categorical encoding for partial data

v0.12.5

Warning: do not use this version! The distributed build is broken. Use v0.12.6 instead. 🙏

  • Ensure that the default point colors respect the background when setting both at the same time during initialization. I.e., in the following scenario, the point color will be set to white by default as the background color was set to black:

    jscatter.plot(data=df, background_color='black')

  • Fix an issue when working with views of a pandas DataFrame where not all categorical data is present

  • Loosen strictness of rows and cols of compose() to allow having empty cells in the grid

v0.12.4

  • Respect the dictionary key-value order of categorical encoding maps in the legend. E.g., the following categorical color map legend will read C, then B, and finally A:

    scatter.legend(True)
scatter.color(map=dict(C='red', B='blue', A='pink'))

  • Update third-party JS libraries

v0.12.3

  • Fix incorrect legend for categorical coloring

v0.12.2

  • Update regl-scatterplot to v1.6.9

v0.12.1

  • Fix the ordering of the legend's value labels for continuous encodings such that high to low values are order top to bottom (#70)

v0.12.0

  • Add support for referencing points by the Pandas DataFrame's index via Scatter(data_use_index=True) or scatter.data(use_index=True). This is useful for synchronizing the selection or filtering of two Scatter instances that operate on different data frames with point correspondences. (#62

    import jscatter
import numpy as np
import pandas as pd

df1 = pd.DataFrame(
    data=np.random.rand(16, 2),
    index=[chr(65 + x) for x in range(16)],
    columns=['x', 'y']
)
df2 = pd.DataFrame(
    data=np.random.rand(8, 2),
    index=[chr(76 - x) for x in range(8)],
    columns=['x', 'y']
)

s1 = jscatter.Scatter(data=df1, data_use_index=True, x='x', y='y', x_scale=[0, 2], y_scale=[0, 1])
s2 = jscatter.Scatter(data=df2, data_use_index=True, x='x', y='y', x_scale=[-1, 1], y_scale=[0, 1])

def on_selection_change(change):
    s2.selection(s1.selection())

s1.widget.observe(on_selection_change, names='selection')

jscatter.compose([s1, s2])
    shorter.mp4

  • Avoid unregistering all observers when calling jscatter.compose() such that external observers remain registered

  • Fix undefined this in codec preventing the scatter.selection() from working correctly (#66)

v0.11.0

  • Add scatter.filter([0, 1, 2]) for filtering down points. Filtering down to a specific subset of points is much faster than than updating the underlying data (#61)
  • Add scatter.data(df) to allow rendering new data points without having to re-initialize the scatter instance (#61)
  • Add the ability to automatically zoom to filtered points via Scatter(zoom_on_filter=True) or scatter.zoom(on_filter=True)
  • Add lasso on long press and make it the default. The behavior can be changed via Scatter(lasso_on_long_press=False) or scatter.lasso(on_long_press=False)
  • Updated regl-scatterplot to v1.6

v0.10.0

  • Add support for automatic zooming to selected points via scatter.zoom(on_selection=True)
  • Fix view synchronization issue
  • Add remove() to the JS widget to ensure that the scatterplot is destroyed in ipywidgets v8.

v0.9.0

  • Add support for animated zooming to a set of points via scatter.zoom(pointIndices) (#49)
  • Bump regl-scatterplot to v1.4.1
  • Add support for VSCode and Colab (#37)
  • Fix serving of numpy data for JS client. Use consistent serialization object between JS and Python.

v0.8.0

  • Add support for end labeling of continuous data encoding via a new labeling argument of color(), opacity(), size(), connection_color(), connection_opacity(), and connection_size(). (#46)
  • Fix the incorrect size legend when the size map is reversed (#47)

v0.7.4

  • Adjust widget to be compatible with ipywidgets v8 and jupyterlab-widgets v3 (#39)

v0.7.3

  • Fix broken build of v0.7.2
  • Fix versions of ipywidgets and jupyterlab_widgets to avoid running into incompatibilities (#40)

v0.7.2

Do not use. Build is broken

v0.7.1

  • Take the x-padding into account when setting a fixed width (#36)
  • Make width and height correspond to the canvas' (i.e., inner) dimensions such that Scatter(width=500, height=500) will lead to a canvas of size 500x500 px.

v0.7.0

  • Add support for legends via scatter.legend(True, position='top-right', size='small') (#30)

v0.6.1

  • Remove accidentally added console.log

v0.6.0

  • Add support for axes labels via scatter.axes(labels=True) or scatter.axes(labels=['x-axis', 'y-axis']) (#29)

v0.5.1

  • Fix issues when specifying the color, opacity, or size map via a dict

v0.5.0

  • Breaking changes:
    • For scatter.color(), rename color to default, color_selected to selected, and color_hover and hover
    • For scatter.opacity(), rename opacity to default
    • For scatter.size(), rename size to default
    • For scatter.connection_color(), rename color to default, color_selected to selected, and color_hover and hover
    • For scatter.connection_opacity(), rename opacity to default
    • For scatter.connection_size(), rename size to default
  • Add scatter.opacity(unselected=0.5). This property defines the opacity scaling factor of unselected points and must be in [0, 1]. This scaling is only applied if one or more points are selected.

v0.4.1

  • Fix an issue when dynamically resizing the scatter plot height
  • Fix an issue when switching from categorical to continuous color/opacity/size encoding

v0.4.0

  • Breaking change: Renamed color_active and connection_color_active to color_selected and connection_color_selected respectively for clarity.
  • Add support for axes via Scatter(axes=True, axes_grid=True) or scatter.axes(True, grid=True).
  • Add support for log and power x/y scales via Scatter(x_scale='log') or scatter.y(scale='pow').
  • Add docstrings and type hints

v0.3.4

  • Unset data-driven color, opacity, and size encoding when only a constant value was passed to an encoder function. E.g., after initializing scatter = Scatter(..., color_by='property'), calling scatter.color('red') will automatically unset color_by.

v0.3.3

  • Fix scatter.pixels()

v0.3.2

  • Fix more type hints...

v0.3.1

  • Make type hints backward compatible to Python 3.7

v0.3.0

  • Breaking change: Change the signature of compose() to simplify correspondence mapping of data points.
  • Breaking change: Rename view_pixels to view_data and add faster synchronization from the JS to Python kernel
  • Add ability to link the view of multiple scatter plots via compose(sync_view=True)
  • Add link() as a shorthand for compose(sync_view=True, sync_selection=True, sync_hover=True)
  • Add ability to defined categorical colors using a dictionary. E.g., scatter.color(by='coolness', map=dict(cool='blue', hot='orange'))
  • Fix two issues with the order argument in methods color(), opacity(), size(), connection_color(), connection_opacity(), and connection_size() that prevented it's propper use.
  • Improve the ordering of the default Okabe Ito color map
  • Expose default the Okabe Ito (okabe_ito) and Glasbey (glasbey_light and glasbey_dark) color maps for convenience
  • Automatiecally handle string as categorical data

v0.2.2

  • Fix broken installation of v0.2.1 (#23)

v0.2.1

  • Simplify installation (#16)

v0.2.0

  • Complete rewrite of the Python API to match (#3)

v0.1.2

  • Properly destroy regl-scatterplot on destroy of a widget instance

v0.1.1

  • Fix bunch of typos related to renaming the pypi and npm package

v0.1.0

  • First version