vault backup: 2024-04-24 00:32:54

This commit is contained in:
2024-04-24 00:32:54 -05:00
parent 3b0d335b64
commit a760305a5c
5 changed files with 49 additions and 20 deletions

10
Datastore.md Normal file
View File

@@ -0,0 +1,10 @@
# What to use?
Currently (4/24/24) thinking of using primarily NoSQL based storage.
* Not much of this seems to be relational data
* Not as familiar, so good learning opportunity
* Also, I _think_ SQL is more expensive to run in GCP
* Granted, the actual inventory portion could be relational
# What to store?
* Should receipt images be retained? doing so would let us train a data model on them later
* Is a separate document needed to store what line item associations are made? or are we willing to accept adding data to the stored Document AI response?