Eyekit

=========================

Added
-----

- Added `create_stimuli()` function for creating experimental PNG stimulus images.
- `segment()` method added to `FixationSequence` for segmenting fixation sequence at given time intervals (e.g. page turns or subtitle timings).

=======================

General
-------

- Eyekit is now licensed under the terms of the GPLv3.
- Dropped support for Python 3.7

Added
-----

- Ability to extract individual samples from an ASC file in addition to the fixations by setting `import_samples=True`.
- Ability to place a background image in an `Image` object, which is useful if you want to render participant fixations over a screenshot of the experimental screen.
- Added `render_frame()` method to `Image` for building videos and animations.
- `discard_long_fixations()` method added to `FixationSequence`.
- `shift_x()`, `shift_y()`, and `shift_time()` methods added to `Fixation` and `FixationSequence`.

Changed
-------

- The fixation tagging system now allows arbitrary key-value pairs.
- The `snap_to_lines()` method now returns two measures of data quality, delta and kappa.

=========================

Added
-----

- Option to specify the character encoding when importing ASC or CSV files.

=========================

Added
-----

- Added new line assignment method "slice" based on the algorithm described by [Glandorf & Schroeder (2021)](https://doi.org/10.1016/j.procs.2021.09.069).

=========================

General
-------

- The concept of "zones" will be removed in the future due to the confusing terminology. It will still be possible to mark up IAs in the raw text, but these will no longer be called zones.
- Updated code style to Black v22.

Added
-----

- Added `set_crop_area()` to `Image`, which allows you to specify a particular area of the image to crop (rather than using the limits of any TextBlocks placed on the image).
- The opacity can now be set on line segments drawn using the `draw_line()` method of `Image`.

Changed
-------

- When manually setting the start or end time of a `Fixation`, an error is raised if the start time is after the end time or the end time is before the start time.

Deprecated
----------

- The `TextBlock.word_centers()` method has been deprecated, since this is of limited use. Instead, use something like `[word.center for word in txt.words()]` to accomplish the same thing.
- The `TextBlock.zones()` and `TextBlock.which_zone()` methods have been deprecated.

=======================

General
-------

- Changed development status to beta. Going forward, the aim will be to minimize non backwards compatible changes to the API.
- Bump required Python version to 3.7 and make Numpy an optional dependency.
- Many function and method arguments are now explicitly keyword-only.

Added
-----

- Support for storing pupil size data and importing this information from ASC files.
- Added basic fixation tagging scheme, allowing you to tag fixations with arbitrary identifiers.
- Added `shift_start_time_to_zero` method to `FixationSequence` – shift all timings so that the first fixation starts at time 0.
- The color and radius of a fixation can be set for each fixation independently based on user-defined functions.
- The color of saccade lines can be set separately from the color of fixations.
- Option to set the opacity of fixations.

Changed
-------

- When importing from ASC, 1ms is added to the end time for consistency with SR Research's calculation of duration.
- Fixations are now serialized as dictionaries, which will permit greater flexibility in the future.

Deprecated
----------

- The functions `io.read` and `io.write` have been deprecated in favor of `io.load` and `io.save`.
- The functions `tools.snap_to_lines`, `tools.discard_short_fixations`, and `tools.discard_out_of_bounds_fixations` have been deprecated; these are now methods of `FixationSequence`.

Removed
-------

- `Image.draw_sequence_comparison` has been removed since this can now be accomplished with `Image.draw_fixation_sequence`.