Zotero is slow in a large document on macOS

adomasven

This discussion was created from comments split from: Zotero is extremely slow in long Word document.

Skelbti

Hi,

I'm having the same problem with my 800-page PhD dissertation. I've lost many days of work waiting for Zotero to update all the citations (over 1,000). I use Word on macOS, but the same thing happens on my Windows PC. I know Word isn't ideal for long documents either, but perhaps the integration between Zotero and Word could be improved.

I'll provide my Debug ID for this issue once the process is complete. For now, the log stays at these lines:

(3)(+0000000): Integration: MacWord16-refresh:'/Applications/Microsoft Word.app/' invoked

(3)(+0000000): Integration: Instantiating MacWord plugin handler for command refresh with doc /Applications/Microsoft Word.app/

(3)(+0000060): Running /usr/bin/osascript '-e' 'tell application "Microsoft Word" to time to GMT'

(4)(+0001786): ZoteroMacWordIntegration: getDocumentData

(3)(+0000037): Moving window to 560, 519

(4)(+0000105): ZoteroMacWordIntegration: getFields

(3)(+0010458): Running /usr/bin/sw_vers '-productVersion'

EDIT: It stays at this line for more than 20 minutes now.

Skelbti

My Debug ID for a complete refresh : D1810492742

adomasven

Integration: Retrieved 865 fields in 3804.766; 0.22734643865089207 fields/second
Integration: MacWord16-refresh:'/Applications/Microsoft Word.app/' complete in 6543.82s

It's taking over an hour to retrieve 865 citations in your document, and the full update took nearly 2 hours. This seems very slow even for Word for Mac standards, but that largely depends on the complexity of the document (number of figures, footnotes, tables, bookmarks, etc), the speed of your Mac, whether your system is overloaded with other work, and whether you keep Word in front of other windows (macOS reduces the speed of windows running in the background to improve speed). It'd be interesting to see a Debug ID for refreshing this document on Windows, it should definitely be much faster.

Otherwise, your best bet is to split this document into chapters and work on their citations separately, and then merge the document in the end and do a final refresh there.

Skelbti

My document is complex, containing over 1,000 footnotes, 200 figures, and 50 tables. My computer is a 2023 MacBook Pro with an M2 Pro chip, and Activity Monitor shows that the RAM and CPU are not being used to their full potential. Therefore, I doubt that my computer's processing power is the problem.

I have already submitted my Ph.D. dissertation, but I need to make some changes for the publisher. I am now at the point in the process where splitting the document into multiple files is no longer an option.

I'll try the refresh process on my Windows desktop PC and post the debug ID here to see if there's a significant difference.

Perhaps the architecture of the Word connector needs to be changed? Is there a way to not parse every citation each time, only the ones that changed? Even the pipeline through AppleScript seems to act like a bottleneck. However, I'm not a developer, so these are just simple ideas. Thanks again for Zotero!

adomasven

Yes, the pipeline via AppleScript is indeed a bottleneck, but we don't have a better API available to us for Word for Mac.