The previous page Ideal Academic Workflow, Illustrated showed what could be possible. We are actually very close to this already, but there are some roadblocks. I had to cheat a bit to get this to work just right. Some of these items are my fault, or quirks of this particular article, etc. but I think it is useful to see all of the things I had to fudge - especially as fixing some of them will improve the workflow for everyone. (See especially step 4!)

1. I already had the citation in my database, so Zotero created a duplicate. It has duplicate detection so this can be cleaned up afterwords, but ideally it should detect that it is already there and let me know.
2. The Citation Manager Plugin initially added the full date (with author and month) instead of just year year in numerous places, I had to clean this up. It also didn’t capture the journal in my template, that is my fault - I probably have to update my template.
3. The link to the PDF was broken because of the hashtag in the title. I had to copy the file, remove the hashtag, and then add it to Logseq manually.
4. A number of issues arose with annotations. My ideal, text, followed by a Pandoc Citekey and a ref link to the highlight in the PDF is not supported. I had to do it in four steps: 1. Highlight, 2. copy text, 3. copy ref, 4. turn ref into link so the text doesn’t appear twice (you can’t select the text in the ref, so this is essential), 5. Add the citekey using the Citation Manager Plugin, 6. add the page number (some PDF readers are able to do this automatically).

It may be that in the future one would just annotate in Zotero and then pull the annotations into Logseq, but even then some of these issues should be dealt with, in order to ensure that the next stage of the workflow works well.