inscopix / py_isx Goto Github PK

support for VesselSets

• write classes for isx.VesselSet
• structure VesselSet classes and methods like it is in the original isx (IDPS) implementation
• use the isx.CellSet in py_isx as a reference

For reading "movie_u8.isxd" and "movie_longer_than_3_min.isxd" on unix operating systems, Movie.read(file_path).get_frame_data(0).dtype results in dtype('int64') , however on Windows, the dtype is dtype('int32'). Is this the intended behavior?

• indicate in pyproject.toml that the version number is always "dev"
• version number in setup.py should be read dynamically from tag
• action should only run when we make a tag/release

It's atypical for a Python package to have all of it's relevant code specified in the __init__.py file: https://github.com/inscopix/py_isx/blob/main/isx/__init__.py

Would it be possible to break this into submodules?

sklearn is usually a good reference for how to structure a Python package

isxd movies can have invalid frames of the following type: dropped, cropped, blank.

They are not stored on disk, the only indication of them is in the json metadata. Currently this api breaks when trying to access the last frame of a movie with dropped frames because it's not correctly recalculating the "true" frame index on disk.

We need to update the Timing class to have these properties and update the read/write functions to support these invalid frames.

I'm not sure if this is a duplicate of #13, but what is the recommended way to get the frame times from a cellset file?

It looks like the CellSet.get_cell_image_data method is missing from the docs.

Both the API distributed with IDPS and the PyPi package share the package name 'isx'
This can cause ambiguity as to which package is installed when trying to automate package installation and dependency checking using setup tools or another method. Perhaps the API shipped with IDPS should be renamed to isx_api or something similar in future releases.

Cheers!

problem

setup.py file should contain something like this:

license_files = ('LICENSE.txt',),

but this isn't being created by our automatic generation of the setup.py file

should support reading mulit-plane movies

Merge the small tests into a single action