sdrp
commented
4 years ago Before any of the above, need to get writeToCSV working again, which also involves re-writing the JSON decoder function now that it takes an array of JSON objects after the multipage update
General idea is that we want the platform to save as much data about user workflow as possible, and also save all captured/digitized images in a common-sense way so they can be accessed and re-evaluated later.
Concrete Tasks: 1) Final Annotation: Get the "Save" button the review page to save a CVS/TSV in a clear format. This is already somewhat there.
2) Digitization Doc / JSON Annotation: "Save" button should also save the filled in JSON annotation of the form, with two fields for the value. One is "Original Annotation" and the other is "Saved Annotation". Any difference between the two means that the user manually altered/corrected something during the review step. Final JSON that is saved should also contain metadata at the topmost level. This includes # of seconds and frames until a good alignment was found, the alignment score of the chosen image, how long the user spent on the "Review" page, the number of errors/revisions made, the final end-to-end digitization time, and really anything else of value that can be captured.
3) The actual image that was captured and digitized. This should be saved in a specific folder outside of
/backend/staticand be given a name that contains identifying information (ex. registration number, date, and page of record).