Any new process, pipeline, or toolbox will likely come with its metadata and completion schemas. We must decide, implement and document a way to discover and parse them, and/or to switch to the correct schema when using a pipeline.
This is obviously linked to #268.
For Morphologist we also have an additional complexity: the Morphologist schemas are used both for the Morphologist toolbox (with its own project sources), and for testing Capsul through FakeMorphologist and TinyMorphologist. Implementing the schema in Morphologist would break dependencies and capsul tests (capsul should not depend on Morphologist). So for now it is implemented in Capsul, which is not good, and possibly has side effects (see #327). I don't have a good solution for this right now.
Any new process, pipeline, or toolbox will likely come with its metadata and completion schemas. We must decide, implement and document a way to discover and parse them, and/or to switch to the correct schema when using a pipeline. This is obviously linked to #268.
For Morphologist we also have an additional complexity: the Morphologist schemas are used both for the Morphologist toolbox (with its own project sources), and for testing Capsul through
FakeMorphologistandTinyMorphologist. Implementing the schema in Morphologist would break dependencies and capsul tests (capsul should not depend on Morphologist). So for now it is implemented in Capsul, which is not good, and possibly has side effects (see #327). I don't have a good solution for this right now.