Awesome Automated APIs with Automagic REST
summary
A generator introspects Postgres and produces 60,000 endpoints (30,000 tables + aliases), Django models and serializers nightly so schema changes land in version control; permissions defer to LDAP via a shim that queries Postgres ACLs; huge tables (NYSE trades) compress to ~100 TB with a columnar extension; slow counts use count-estimate pagination (query EXPLAIN parsing) to cut responses from minutes to milliseconds; an Django REST framework renderer xlsx exports spreadsheets; publishing time falls from two hours to about three minutes and the system works in production.
AI-written summary · about
chapters (12)
- 0:00Start
- 4:26WRDS: A BRIEF HISTORY
- 7:52DUBIOUS DATABASE DECISIONS
- 12:12FOUNDATION COMPLETE BUILDING THE RESTFUL API.
- 14:32HOW WE BUILT IT: FIELD MAPPING
- 16:59END RESULT: MODELS
- 19:41BIG DATA PROBLEMS
- 22:3260,000 MODELS: SLOW!
- 25:06PERMISSIONS FOR 63.000+
- 26:37SPEEDING UP SLOW COUNT(*)
- 29:30FILTERS ON INDEXED COLUMNS
- 30:59SPREADSHEETS
related talks
47mBuilding Scalable Web Apps
30mREST & GraphQL: Adventures in schema-first API development
31mWorking smarter to fight cancer: using dunder special methods to automate
29mSwiss Army Django: Small Footprint ETL
01h 00mBuilding an API with Apigility
24mA New Look Into APIS - Graphene
21mBuilding Powerful APIs with Django, Django Rest Framework, and OpenAPI
34mAPI Platform: From Rest & GraphQL APIs to state-of-the-art standards in seconds