pylts Docs

A thin, reusable pydantic wrapper over limited litestream functions: makes it easier to use litestream inside a python script to handle the replicate and restore aspects of an sqlite database in an s3 bucket.

Usage

Python

# assumes installation via pip install pylts / poetry add pylts
>>> from pylts import ConfigS3 # inherits from pydantic.BaseSettings
>>> from pathlib import Path
>>> d = Path().cwd() / "data" # where to place database
>>> stream = ConfigS3(folder=d) # assumes .env variables based on <s3> bucket
>>> stream.dbpath # constructed local <db> path based on folder
>>> stream.s3 # replica s3:// url in aws <s3>
>>> stream.restore() # subprocess: 'litestream restore -o <db> <s3>'
>>> stream.timed_replicate() # subprocess: 'litestream replicate <db> <s3>'

Flow

Placement

---
title: From Local to Bucket to Production
---
flowchart LR
  subgraph fly.io
    volume
    app
  end
  subgraph aws
    s3(s3 bucket)--pylts restore volume-->app
  end
  subgraph local
    db--"litestream replicate (args)"-->s3
    code--fly deploy-->app
  end

Environment	Task
local	Set up repo with a sqlite.db (db)
local	With pydantic-parsed values, litestream replicate this created db to s3
local	Deploy the repo to fly.io
prod	Pydantically restore db from s3 to fly.io

Retrieval

The database that is (re)stored in the volume (fly.io) may be updated either by cron jobs or by adding, deleting entries directly. How do we ensure these updates are persisted to the s3 bucket?

---
title: From Production to Bucket to Local
---
flowchart LR
  subgraph fly.io
    volume
    app
  end
  subgraph aws
    app--pylts replicate volume-->s3
  end
  subgraph local
    s3--"litestream restore (args)"-->db
    code
  end