adding testing data for distriploy release #1

lrouhier · 2020-07-22T18:39:38Z

This PR aims at uploading the current testing data used in the ivadomed test to github in order to perform a distriploy release.

Needed images and files:

model_unet.pt --> pt file for inference testing (does not really need to be well trained we just test if the code runs)
sub-XXX --> Bids folder with multiple contrast (T2w,T2star and T1w)
derivatives/sub-XXX --> derivatives folder with seg-manual labels for all contrasts, labels-disc-manual labels for T2w, lesion-manual label for all contrasts (dummy seg),
bounding_box.json --> dictionnary to test certain function from the bounding_box.py file
results.csv --> results from automate training to test results analysis.
dataset_description.json --> needed file to descript dataset
participants.csv --> table with subject names and potential metadata
3 config file for train/test/eval command --> We need to separate the 3 (it's json files so it does not take much space)

jcohenadad · 2020-07-22T18:42:30Z

derivatives/labels/sub-test001/anat/sub-test001_T1w_lesion-manual.json

@@ -0,0 +1 @@
+{}


where is this data coming from? is it a public dataset? if not, why not simply using a subject from spine-generic?

It's just the copy of the testing data that was already in ivadomed. I do need to add the new data that will be used in the new tests. There will still be a single subject. I don't think it is a public dataset (this PR ).

I don't think it is a public dataset

can we replace it with another subject then

also: if you don't rebase the unused binaries, please make sure to "squash and merge" for merging this PR

I'm looking into it with the test

Maybe. I am trying with the available csf_seg given in the ivadomed/data_example_spinegeneric repo

ok, i'm getting confused now, i thought you were looking for a file with suffix _lesion-manual?

i would suggest you list, in the PR's body, all the files that are needed by your tests. Then, we can all think and find the best solution for the listed files (in terms of size, availability, need for manual intervention, etc.)

Oh sorry I will list them.

I was using the csf segmentation as a dummy seg. (I renamed it T2w_lesion_manual).

I renamed it T2w_lesion_manual

that’s exactly the kind of stuff i’d like to avoid. it might bring confusion to a future student

@lrouhier: For the previous testing data: what I did was to create a "fake lesion", ie a blob, in the spinal cord (the subject has healthy).

jcohenadad · 2020-07-22T18:43:59Z

participants.tsv

@@ -0,0 +1,2 @@
+participant_id	sex	age	first_name	last_name	date_of_birth	date_of_scan	pathology	data_id	institution_id
+sub-test001	unknown	-	unknown	unknown	-	unknwon	unknwon	unknown	unknown


would be better to use a data from which we know the provenance (e.g. spine-generic)

lrouhier · 2020-07-24T17:44:47Z

Pull request autamoatically Closed by removing the previous commit to replace subject with one from spine-generic. This was done to avoid having a big repository for a single subject. PR #2 will provide similar changes. In summary, nothing was done.

jcohenadad · 2020-07-24T17:45:05Z

~~before closing this PR, could you please document what was done? I don't see any change in this PR, nor the info about the new dataset~~

^ scrap that-- info about closing: #1 (comment)

jcohenadad reviewed Jul 22, 2020

View reviewed changes

lrouhier closed this Jul 24, 2020

lrouhier force-pushed the lr/distriploy_release branch from ef4bc83 to 8b4fa52 Compare July 24, 2020 17:38

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

adding testing data for distriploy release #1

adding testing data for distriploy release #1

lrouhier commented Jul 22, 2020 •

edited

Loading

jcohenadad Jul 22, 2020

lrouhier Jul 22, 2020

jcohenadad Jul 22, 2020

jcohenadad Jul 22, 2020

lrouhier Jul 22, 2020

lrouhier Jul 22, 2020

jcohenadad Jul 22, 2020 •

edited

Loading

lrouhier Jul 22, 2020

jcohenadad Jul 22, 2020

charleygros Jul 22, 2020

jcohenadad Jul 22, 2020

lrouhier commented Jul 24, 2020 •

edited

Loading

jcohenadad commented Jul 24, 2020 •

edited

Loading

		@@ -0,0 +1,2 @@
		participant_id sex age first_name last_name date_of_birth date_of_scan pathology data_id institution_id
		sub-test001 unknown - unknown unknown - unknwon unknwon unknown unknown

adding testing data for distriploy release #1

adding testing data for distriploy release #1

Conversation

lrouhier commented Jul 22, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jcohenadad Jul 22, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lrouhier commented Jul 24, 2020 • edited Loading

jcohenadad commented Jul 24, 2020 • edited Loading

lrouhier commented Jul 22, 2020 •

edited

Loading

jcohenadad Jul 22, 2020 •

edited

Loading

lrouhier commented Jul 24, 2020 •

edited

Loading

jcohenadad commented Jul 24, 2020 •

edited

Loading