Refactor documentation #171
Conversation
| @@ -6,47 +6,49 @@ | |||
|
|
|||
There was a problem hiding this comment.
It would be good to explain on the home page what ClinicaDL does in a few words and how it is related with Clinica. This paragraph should also mention in what context ClinicaDL was initially developed as some functions are quite Alzheimer-specific (e.g. restrict or get labels in tsvtools).
There was a problem hiding this comment.
Answered in e0c51a8
| @@ -0,0 +1,21 @@ | |||
| !!! warning "Coming soon" | |||
There was a problem hiding this comment.
Is it useful to make this page available now? It's for the reconstruction case?
There was a problem hiding this comment.
It is useful to explain the output structure.
I added the purpose of this command at the beginning of the file in 66372d5
| - `map.npy` is the numpy array corresponding to the saliency maps and can be loaded with `numpy.load`. | ||
|
|
||
| The output tree for the `individual` level is quite similar, except that one occlusion is created | ||
| per session. Then the output tree also includes the `participant_id` and `session_id`. |
There was a problem hiding this comment.
Wrong formatting (no list displayed).
There was a problem hiding this comment.
I could not change this part as it seems that it disappeared?...
There was a problem hiding this comment.
Hi @14thibea ,
Thanks for this huge PR!
I have several comments because this contribution mix mainly two things: interface changes of the MAPS class and documentation.
- My initial though is that some changes would be in conflict with PR#172, so maybe @ravih18 should be aware of the eventual modifications introduced by this PR.
- New models where uploaded to the CI agents but there are still some problems with the MAPSManager class (see here and example).
Concerning the documentation:
- I'm not sure of the titles for the sections. I would prefer shorter and self-explanatory titles (in some way what has been done in Clinica documentation). I needed to scroll to the end of the Home section to understand that these titles corresponds to the functions proposed by the software.
- We need to uniform where and when we write ClinicaDL and
clinicadl: The first for the full text and the second one only for the commands.
I added some other comments inside the code.
I propose to merge this PR once the CI problems are solved and once we are agree of the section titles (and in general, about the structure of the documentation). In that way, the PR#172 can integrate modifications and progress in parallel. Other modifications (if necessary) will be done in additional PRs.
…AD-DL into refactor_documentation
Co-authored-by: mdiazmel <[email protected]>
Co-authored-by: mdiazmel <[email protected]>
…AD-DL into refactor_documentation
* Add Contribute section * Update README to clarify AD-DL separation * Rename classify in predict * Fix typo * Add default network * Add visualization function * Fix tests * Fix interpretation * Update documentation * Fix random-search + detach analysis * Add new sections on contribution * Fix interpret * Fix typo * Refactor interpretation and prediction * Fix cli + add interpret test * Train fix * Train refactoring * Fix typo * Minor fixes * Fix interpret * Change title in README * Remove config file at third level * Answer to the review comments * Explain AD context * Fix predict * Apply suggestions from code review Co-authored-by: mdiazmel <[email protected]> * Update docs/Train/Details.md Co-authored-by: mdiazmel <[email protected]> * Improve documentation accoding to review comments * Add docstring to meta-maps * Execute predict without GPU * Remove meta-maps test * Fix Jenkinsfile Co-authored-by: mdiazmel <[email protected]>
Updates on MAPS structure coming soon...
Some modifications were added to preserve previous functionalities:
modelbecomes an option and a default architecture exists for all modes (it depends only on the task manager)