Skip to content

Commit

Permalink
Merge pull request #17 from pupil-labs/rennis250-patch-1
Browse files Browse the repository at this point in the history
Point to examples.
  • Loading branch information
rennis250 authored Jun 11, 2024
2 parents bd6ef36 + f502114 commit 5d38d62
Showing 1 changed file with 1 addition and 18 deletions.
19 changes: 1 addition & 18 deletions README.rst
Original file line number Diff line number Diff line change
Expand Up @@ -25,21 +25,4 @@
Neon Recording API
*******************************

- As far as I can tell, between 'main' and 'multipart' branches, the requested features in the original notion doc, as well as what was discussed at meetings, is implemented and the majority is tested (I just don't have access to a multipart recording yet). Various implementations of some functionality were tried, to see what feels most ergonomic for a user. Some functionality was tested with different, but equivalent, implementations, to eliminate any possibility of errors and to provide some assurance of stability.

- See 'multipart' branch not only for code to load multipart recordings, but also for code to load different (appropriate) timestamps based on whether recording came from cloud or phone

- Some examples and discussion in the Notion doc involved fixation data. It was not clear what we wanted to do with fixations, since I understood that this library will only be for Native Recording Data. Shall a fixation detector be run when loading the data?

- I tried, but pl-recover-recording will complicate installation of this library. it requires building untrunc from source and instructions for windows and macOS are needed

- Some functions from the original design are not possible without monkey patching routines deep within Python. For example, with the current generator approach to sampling, the following is not possible:
gaze = gaze.sample(between_two_events).to_numpy()
unless we monkey patch the code for generator objects. Earlier approaches to the implementation of Stream that supported this were rejected after evaluation.
So there is a Stream.sampled_to_numpy() method to handle this request.

- I tried one or two ways and searched/thought about it, but loading all frames at once into RAM does not seem feasible.

- TODO: fill in rest of README

- see `initial discussion/design <https://www.notion.so/pupillabs/Neon-Recording-Python-Lib-5b247c33e1c74f638af2964fa78018ff?pvs=4>`_
Check out the Examples for ways to use this library.

0 comments on commit 5d38d62

Please sign in to comment.