Closed wbwakeman closed 2 years ago
I think it's likely that this is due to the fact that we have 3 pkl files for this experiment (behavior, mapping and replay; all with running data). My guess is that this code is only finding running data for the 'stim.pkl' which is the mapping portion of our experiment. This is 25 minutes long, which matches the number of running samples that are found. We will need to take the running data from all three pkls (or concatenate the pkl files before running them through this code).
Useful information for Wayne:
lims2=> SELECT wkft.name AS wkft, wkf.filename AS wkf FROM well_known_files wkf JOIN well_known_file_types wkft ON wkft.id=wkf.well_known_file_type_id WHERE wkf.attachable_id = 1125713722 AND well_known_file_type_id IN (1007316668,1105900356,610487715);
wkft | wkf
-----------------------+-----------------------------------------
StimulusPickle | 1125713722_578257_20210901.behavior.pkl
MappingPickle | 1125713722_578257_20210901.mapping.pkl
EcephysReplayStimulus | 1125713722_578257_20210901.replay.pkl
Example of current ECEPHYS_EXTRACT_RUNNING_SPEED input.json file:
{
"sync_h5_path": "/allen/programs/braintv/production/visualbehavior/prod0/specimen_1109521187/ecephys_session_1132595046/1132687142/1132595046_583430_20211006.sync",
"stimulus_pkl_path": "/allen/programs/braintv/production/visualbehavior/prod0/specimen_1109521187/ecephys_session_1132595046/1132687142/1132595046_583430_20211006.mapping.pkl",
"output_path": "/allen/programs/braintv/production/visualbehavior/prod0/specimen_1109521187/ecephys_session_1132595046/EcephysExtractRunningSpeedStrategy/analysis_run_1149574448/1132595046_running_speeds.h5"
}
Issues encountered when doing 'extract running speed' for ecephys. Need to investigate these issues and create tickets to address them.
Issue #1: sync file line label has been updated from
stim_vsync
tovsync_stim
as described in the excel file link in "Sync: PCIe 6612 DAQ" section of this confluence page .https://github.com/AllenInstitute/AllenSDK/blob/master/allensdk/brain_observatory/sync_dataset.py#L96
Issue #2: sync mismatch
https://github.com/AllenInstitute/AllenSDK/blob/master/allensdk/brain_observatory/extract_running_speed/__main__.py#L112
Need to make sure a fix here does not break any two-photon running speed processing
Investigation of this issue can be timeboxed to 3 days. May include running another experiment through this to ensure it happens with other experiments.