External brain GitHub work

• Repository made https://github.com/nishio/nishio-hirokazu-external-brain
• Think about what to do next
  • I just have to think about it because I have a meeting in three minutes.
  • JSON Export, etc. is automated in the translation system.
  • Bring it in and create a backup system first.
  • Maybe this: https://github.com/nishio/etude-github-actions
    • 
    • The billing on DeepL was updated, but what’s stopping you?

    • This repository is over its data quota. Account responsible for LFS bandwidth should purchase more data packs to restore access.

    • Ah, the GitHub LFS.
  • 8/21 email.

    • Git LFS has been disabled on your personal account nishio because you’ve exceeded your data plan by at least 150%. Please purchase additional data packs to cover your bandwidth and storage usage:

    • https://github.com/account/billing/data/upgrade

    • Current usage as of 20 Aug 2024 10:38PM UTC:

    • Bandwidth: 8.21 GB / 100 GB (8%)

    • Storage: 150.09 GB / 100 GB (150%)

    • 
    • 
    • 3 data packs and you’re still overflowing.
    • I made it to four.
    • 
    • 
    • Hmmm, well, once and for all.
      • Do you have 150 GB of JSON?
  • etude-github-actions, as the name suggests, was created to practice GitHub Actions, and has been in production for a long time because it worked better than expected.
    • It’s time to review
    • I think you’re forgetting that we’ve considered this before.
    • etude-github-actions
    • pVectorSearch

Maybe, but I think you’re cramming too many functions into one GitHub Actions.

• 1: Create a way to access the latest JSON without hitting the Export API from Cosense.
• 2a: from (1), translation, import to English Scrapbox
• 2b: From (1), insert to vector DB Now (1+2a) and (1+2b) are running in parallel.
• It is easier to link with various other things if it is completed with only (1)
  • I’m sure there are other Cosense users who would like their Cosense data to be on GitHub in an easily accessible form.

I don’t think you should put too much code in (1).

Regarding Translation

• I’m translating line by line at DeepL right now.
• As of 2024-09-08, when the GPT context length became longer, I think it’s fine to pass it as a JSON list and have it return as a JSON list.

---

This page is auto-translated from /nishio/外部脳GitHub作業 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.