siftlode/backend/entrypoint.sh

#!/bin/sh
set -e

echo "Applying database migrations..."
alembic upgrade head

echo "Starting Siftlode API..."
exec uvicorn app.main:app --host 0.0.0.0 --port 8000 --log-config log_config.json
feat: M1 foundation — compose stack, FastAPI, Google OAuth, encrypted tokens - docker-compose with Postgres 16 + slim Python API image - FastAPI app with session middleware, health endpoint, static login page - Google OAuth (Authlib) with email invite-list whitelist; admin role support - User + OAuthToken models; refresh tokens encrypted at rest (Fernet) - Alembic migrations, run automatically on container startup - Postgres backup/restore scripts for portability between machines 2026-06-11 01:01:37 +02:00			`#!/bin/sh`
			`set -e`

			`echo "Applying database migrations..."`
			`alembic upgrade head`

chore: rebrand Subfeed -> Siftlode Rename all user-facing references (UI wordmark Sift+lode, titles, app name, legal pages, onboarding wizard, emails, README/docs) and infra paths (/srv/subfeed -> /srv/siftlode, image tag, deploy script, backup filenames). Internal identifiers kept on purpose: Postgres user/db "subfeed", logger namespace, localStorage keys, and the subfeed_pgdata volume (renaming would orphan the migrated production data). 2026-06-14 04:40:22 +02:00			`echo "Starting Siftlode API..."`
chore: structured timestamped logging across the backend - uvicorn --log-config (log_config.json): timestamped formatters for uvicorn access/error and the app's own "subfeed" loggers (ms precision) - Log key activity: startup/shutdown, login/denied, token refresh, YouTube API errors, subscription imports, sync pause/resume - Background sync jobs no longer swallow errors silently — failures in RSS poll, backfill, enrichment, autotag and resync are logged with tracebacks 2026-06-11 04:26:18 +02:00			`exec uvicorn app.main:app --host 0.0.0.0 --port 8000 --log-config log_config.json`