Simplify the pipeline procedure by passing everything through a single argument conversion_pipeline #298

Activity

1. shollendEODC

   added

   enhancementNew feature or requestNew feature or request

   questionFurther information is requestedFurther information is requested

   on Oct 1, 2026
2. emmanuelmathot commented on Oct 1, 2026

   

   emmanuelmathot

   on Oct 1, 2026

   Contributor

   More actions

   If you mean we should remove the auto detection of the product type. Agreed!
3. shollendEODC commented on Oct 2, 2026

   

   shollendEODC

   on Oct 2, 2026

   CollaboratorAuthor

   More actions

   i would go even further and remove the *_optimized parameters in eopf_geozarr/cpm/writer.py:

   from:

   def write(
           ...
           generic_rechunker: bool | None = None,
           s2_optimized: bool | None = None,
           s3_olci_optimized: bool | None = None,
           ...
       ) -> Any:    
       ...
       selected_pipeline = select_pipeline(
               dtree,
               force=self._resolve_forced_pipeline(
                   dtree,
                   generic_rechunker=generic_rechunker,
                   s2_optimized=s2_optimized,
                   s3_olci_optimized=s3_olci_optimized,
               ),
           )
       ...
   

   To:

   def write(
           ...
           processing_pipeline: str,
           ...
       ) -> Any:
       selected_pipeline = verify_pipeline(processing_pipeline)  # -> check and validate processing_pipeline matches the input dtree
   

   Something along the lines of this; it would also align with the S2Type simplified detection for data product, which could be extended to all other products.
4. emmanuelmathot commented on Oct 2, 2026

   

   emmanuelmathot

   on Oct 2, 2026

   Contributor

   More actions

   I am generally in favor but a separate PR would allow us to check that further
5. shollendEODC commented on Oct 2, 2026

   

   shollendEODC

   on Oct 2, 2026

   CollaboratorAuthor

   More actions

   Definitely, this will have to done in its own PR, as also the test routing and eodc/developmentseed's pipeline will have to be adapted. I don't assign this a high priority, getting the products out soon is more important, but I want to keep it in a backlog for optimizing and refactoring the code.
6. shollendEODC

   self-assigned this

   on Oct 2, 2026