Debugging help: Graph anonymization workflow/script

[Brikas]

Hey, does anyone have an already built workflow for anonymizing a graph?

EDIT: here you go, just made one with Claude: https://github.com/Brikas/logseq-graph-anonymizer

Would love to send my big graph or attach it so performance issues I have.

I'd imagine something like this

• Retains
  • number of pages
  • exact block tree structure, and block count for every page
  • markdown block semantics, such as links, formating
  • block ids, all other default logseq block properties + Tine custom ones.
  • ideally page sizes in kb (but could be an approximate match)
• Replaces
  • generally by some human-readable anonymized determisnist text conservion. Same input returns same anonymized text, so reffences and graph connection is mostly preserved. But one change in char produces completely different output
    • some awareness for if a text is a web/local link, it could keep it a weblink, just anonymize it too by generating a random web/local link
  • all page names
  • all block content
  • all assets (or just placed stubs)