CSV Ingest - fallback to folder name if folder path is not complete #28
Conversation
jrsndl
changed the title
antirotor
added
type: enhancement
| if missing_paths: | ||
| folder_ids_by_name: Dict[str, str] = { | ||
| folder_entity["path"]: folder_entity["id"] | ||
| for folder_entity in ayon_api.get_folders( | ||
| project_name, | ||
| folder_names=missing_paths, fields={"id", "path"} | ||
| ) | ||
| } | ||
| folder_ids.update(folder_ids_by_name) |
There was a problem hiding this comment.
I think this is interesting, but I also think we should exclude any that already have / inside the folder paths to ensure we disallow implicitly matching another folder by name if a folder path was initially passed.
Also, I think we may be better off being even more explicit and allowing a dedicated folder name column instead of folder path so that if a user uses that it's clear that the column allows matching any existing folder?
@jakubjezek001 thoughts?
Note also that this code is buggy - because below we're also using folder_ids_by_path to query the task entities - however those found by folder name are not included in that lookup.
There was a problem hiding this comment.
The folder path column should be sufficient. Let's not add another column if the information can be extracted from the folder path.
|
This PR conflicts with the issue here: #24. The issue suggests creating nonexistent hierarchies. The fallback mechanism tries to identify and handle misconfigured entries in the CSV, even if they are incorrect. The CSV data is assumed to be correct, and any errors indicate a failure in the previous data assembly step. The original CSV ingest design treats it as the source of truth, expecting perfect data. Therefore, nonexistent folder paths should be prepared as we do in Editorial. |
That's clear. Having said that, it means this PR is the wrong approach and this should be closed. Yes? |
That is what I would suggest, but lets wait for a @jrsndl opinion. |
…me_from_folder_path
|
Closing this PR - it can be always reopened in future in case it will be needed. |
The CSV Folder Path column should be first matched against project folder paths, and the failed items agains project folder names. Maybe this should be optional so the current behavior is preserved.
See issue 27
closes #27