search

catalystneuro / visual-coding-to-nwb-v2

Conversion of old v1 NWB files from https://registry.opendata.aws/allen-brain-observatory/ to v2 so they can be uploaded to DANDI.
MIT License
0 stars 0 forks source link

Is `ophys_experiment_name` redundant with `session_type`? #5

Closed CodyCBakerPhD closed 11 months ago

CodyCBakerPhD commented 1 year ago

So many custom fields on the outer metadata...

For the prototype file 571099190.nwb

general.session_type = three_session_B

and

general.ophys_experiment_name = 20170224_283278_3StimB

CodyCBakerPhD commented 1 year ago

Well, turns out there is huge variability/heterogeneity on this one

From script: https://github.com/catalystneuro/visual-coding-to-nwb-v2/blob/a2ec802e0617569d359321405376e403e44e6b7d/src/visual_coding_to_nwb_v2/visual_coding_ophys/scripts/check_experiment_name_and_session_type.py

and JSON dump of stripped references: https://github.com/catalystneuro/visual-coding-to-nwb-v2/blob/main/src/visual_coding_to_nwb_v2/visual_coding_ophys/scripts/stripped_experiment_names.json

Most items belong to one of three groups 3Stim{A or B or C2} but even then they are variations in capitalization, misspelling, or disarrangement

image

CodyCBakerPhD commented 1 year ago

Whereas the session_type is much more stable but still some variation for the 'C' group

image

CodyCBakerPhD commented 11 months ago

After more experience with this dataset I can safely say that all of the information contained in the variability of the ophys_experiment_name is captured by other more structured and reliable fields across the file contents

From http://help.brain-map.org/display/observatory/Data+-+Visual+Coding, we can conclude that the session_type is reliable for this and will be included in the session ID for easier human readability