All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
Richard Vaelle (@flyingfalling) fixed a major, where prior to this release the returned samples dataframe had too many rows (by the number of events), the values in these rows were undefined/unitialized values which, depending on the system, could be 0's, values close to 0, previous memory-usages o, or any other value. This can be very problematic, if by chance the same memory layout is reused, because then the initialized values could be eyetracking data from a previous call to pyedfread. This should be a quite visible error through.
- Some bugfixes to get tests working on windows
- Some small fixes in the readme
- Bugfix for
parse_msgwhere astartwithwas fixed tostartswith
So far, blink-events were not explicitly returned. Rather, a blink column existed, indicating that within an event, a blink occurred. Two breaking changes:
- blink renamed to contains_blink
- Blink events are returned explicitly, using ENDBLINK event type from Eyelink. I (Benedikt Ehinger) noticed that the ENDBLINK and STARTBLINK start-times differed by 1 sample, but I have no explanation for this and just use the ENDBLINK data.