core: GET /classifier-jobs/:jobId/best-detections that can return best detections in a classifier job. Query parameters:
streams - which stream_ids to
by_date - value of 'true' will return best detection in each date within each stream (4 days, 3 streams = 12 n_per_stream detections), value of 'false' or missing value will return just the best detections within the jobstream (3 streams = 3 * n_per_stream)
start - start date from which to search for best detections by their start date (should only be used with by_date=true)
end - end of the date period where the best detections should be searched by their start date (should only be used with by_date=true)
review_statuses - a list of review_statuses that are acceptable to find
n_per_stream - a number of detections per group. Maximun is 10.
Caveat:
When using 'review_statuses' option, notice that n_per_stream is taken BEFORE applying the 'review_statuses' filter, so if you have
reviewed best N detections, try requesting n_per_stream with n+1 to get remaining detections
Also updated hook that runs on job completion, to save the best detections in a finished job. is not set (or is false), the best n_per_stream (up to 10) detections are returned for each stream in the job
✅ DoD
📝 Summary
/classifier-jobs/:jobId/best-detectionsthat can return best detections in a classifier job. Query parameters:streams- which stream_ids toby_date- value of 'true' will return best detection in each date within each stream (4 days, 3 streams = 12 n_per_stream detections), value of 'false' or missing value will return just the best detections within the jobstream (3 streams = 3 * n_per_stream)start- start date from which to search for best detections by their start date (should only be used with by_date=true)end- end of the date period where the best detections should be searched by their start date (should only be used with by_date=true)review_statuses- a list of review_statuses that are acceptable to findn_per_stream- a number of detections per group. Maximun is 10. Caveat: When using 'review_statuses' option, notice thatn_per_streamis taken BEFORE applying the 'review_statuses' filter, so if you have reviewed best N detections, try requestingn_per_streamwith n+1 to get remaining detectionsAlso updated hook that runs on job completion, to save the best detections in a finished job. is not set (or is false), the best
n_per_stream(up to 10) detections are returned for each stream in the jobOrder of results: by_date=true: day1, stream1, bestTop1 day1, stream1, bestTop2 day1, stream1, bestTopN... day1, stream2, bestTop1 day1, stream2, bestTop2 day1, stream2, bestTopN... day1, streamN..., bestTopN... day2, stream1, bestTop1 day2, stream1, bestTop2 day2, streamN..., bestTopN... ... dayN..., streamN..., bestTopN... ...
by_date=false stream1, bestTop1 stream1, bestTop2 stream1, bestTopN... stream2, bestTop1 stream2, bestTopN... streamN..., bestTopN...
📸 Examples
Put screenshots or response/request examples here!
🛑 Problems
💡 More ideas
Write any more ideas you have