Open YoniSchirris opened 1 year ago
From slack:
Jonas: I guess outputting it as json will help right
Yoni: a meta.json with slidescore_server: str, slidescore_study_id: int, slide_mapping: dict ? Certainly
Jonas: Yes all in a file not sure if meta.json is the proper name
While working on it, I realized that it might happen that users download multiple setudies to a single directory, which would then overwrite the slidescore_url
or slidescore_study_id
. Hence, I added an additional layer of unique identifiers using thte url and study id, leading to
{
"https://rhpc.nki.nl/slidescore/": {
"876": {
"slidescore_url": "https://rhpc.nki.nl/slidescore/",
"slidescore_study_id": 876,
"slide_filename_to_study_image_id_mapping": {
"PD10010a_HE-2014-12-07_14.37.01.ndpi": 98528,
"PD10011a_HE-2014-12-07_15.02.40.ndpi": 98337
}
}
}
}
Which may seem a bit redundant. However, I added the slidescore_url
and slidescore_study_id
keys as a means of "documentation".
The name slide_filename_to_study_image_id_mapping
clearly states that the image_id
that the filename is mapped to is dependent on the study
, hopefully.
@AjeyPaiK Do you use the old way of saving the ids? I could still save it the old way.
@jonasteuwen With the current set-up I can't save multiple slidescore slides that have the same filename.
I'd say this is not really supposed to happen, unless there's duplicates which we don't care about. But maybe somehow it could happen. Should we handle this?
Is your feature request related to a problem? Please describe. Documentation is improved. e.g.
/mnt/archive/data/pathology/youngboost_TNBC_HE
contains a TNBC subset of slides, but it is unclear where these slides come from. Also, it's unknown how to upload slides from the directory alone.Describe the solution you'd like
study.txt
with required info for slidescore to upload annotatationsDescribe alternatives you've considered