Comparatively, the path chooser in the Batch Attachment Uploader is rather limited:
Collection Object/Cat #
Collection Object/Prev/Exch #
Taxon/Full Name
Collecting Event/GUID
Collecting Event/Field #
Loan/Loan #
Accession/Accession #
Gift/Gift #
Borrow/Invoice #
I reckon it was a matter of including most useful/used/requested attachment tables.
what if one would like to upload and match attachments to another table?
In French Guiana Herbarium (CAY) we have digitalized most dried sheets and named them accordingly to the preparation barcode. Therefore it would be straightforward to link them as Preparation Attachments. Should we send a feature request to update the B.A.U. ? Or rather use the python API to generate our preparation attachments ?
Preparation attachments based on bar code would be our sole uncovered needs for both CAY an NOU Herbarium so far.
We’re considering to include as well in situ pictures in Specify-7 in near future but that is already supported by current B.A.U. capabilities as Collecting Event Attributes.
I would counter that we could negate having to add all of these edge cases by simply extending the attachment uploader to support the current Specify 6 functionality of being able to use a mapping file to upload attachments based on any table and a primary key like catalog number, field number, preparation number, etc.
Yes, that would feel very “natural” to prepare a CSV file with, let’s say in our case catalog number, preparation barcode and a path to the digitalized dried sheet and then use the workbench mapper to map the fields.
Is it what you mean @AndyBentley (I’ve never used Specify-6 so I’m not familiar with its functionalities) ? Somehow a merge of Workbench and Batch Attachment Uploader ?
Yes, in Specify 6 you can create a csv file with two columns - catalog number or other identifier, and image name. This file is then placed in the directory with all the images and used to match images to identifiers in the respective table. If the workbench supported image upload then yes, the workbench could be effectively used for this.