fix: keyword parsing

[jrasm91]

Found a case where keywords returned a list that included a number (year).

[mattsteg]

Thank you for the fix I was just about to submit a bug report.

It appears that the fix doesn't support hierarchical keywords properly. I have a tag hierarchy with a number (year) as the 2nd tier and the hierarchy was not imported.

[jrasm91]

Thank you for the fix I was just about to submit a bug report.

It appears that the fix doesn't support hierarchical keywords properly. I have a tag hierarchy with a number (year) as the 2nd tier and the hierarchy was not imported.

Hierarchical implies / as a delimiter. Are you using something else?

[mattsteg]

I'm not directly using anything - the files are generated by commercial software Capture One on output and (absent adding additional workflow steps) I get what I get. Short version - C1 uses | in "HierarchicalSubject"

Looking at the Exif, I see the following:

1. Hierarchical Subject uses | as a delimiter
2. Tags List uses / as delimiter, but does not include this hierarchy
3. The XMP on the input files (used for syncing between C1 and digikam - this file is not within view of immich) did not have this hierarchy applied
4. After initiating an XMP sync, the XMP on the input file has added this hierarchy to "Hierarchical Subject" with | delimiters

So the implication is that Capture One writes hierarchical keywords to Hierarchical Subject using | as delimiter.

Given that you're getting other tags correct (which are saved in Tags List with / delimiter by digikam) I think you're probably reading in those correctly and not reading the Hierarchical Subject ones (or discarding them)

Tag data is available in the following places in these files:

1. XMP-acdsee "Categories". This (I believe) I'm adding from digikam for compatibility and not using as far as I know. XML, non-hierarchy
2. XMP-microsoft "LastKeyWordXMP". Added by digikam (I believe) for compatibility. / delimited hierarchy
3. XMP-mediapro "CatalogSets" Again added by digikam for compatibility I believe, | delimited hierarchy
4. XMP-dc "Subject" individual keywords (non-hierarchy) separated by * - this is updated by Capture One
5. XMP-lr "HierarchicalSubject - keyword hierarchy delimited by | - this is what Capture One is updating for hierarchy
6. XMP-digiKam "TagsList" - keyword hierarchy delimited by /, added by digikam when roundtripping.

[jrasm91]

Correct. At the moment we only are importing data for "tags" from Keywords and TagsList.