fix: TMDB images on recommendations + Postgres movies seed on HF fast path

fix: like button, TMDB images, pipeline animation

data: commit MovieLens 1M dataset + fix ignore files

fix: add httpx to HF Dockerfile deps (required by tmdb.py)

data: add pre-seeded Redis snapshot for fast HF Spaces cold start

fix: keep data/redis_snapshot.json.gz in docker context for HF build

chore: proper .gitignore and .dockerignore

sync: port TMDB enrichment, bootstrap fixes, frontend refactor from main

removed useless

feat: add HF Spaces deployment (single-container, pre-seeded Redis)

docs: rewrite README with Mermaid diagrams and fix scoring formula

feat: MLOps eval, A/B testing, DLQ, WebSocket, cold-start fallback

updated frontend and solved bugs

fix: Redis AOF persistence volume, api depends_on kafka healthy

feat: cinematic UI redesign (navy+amber palette, genre gradients, skeleton, collapsible simulate panel)

fix: pg.acquire() not await pg.acquire() — PoolConnectionProxy not async ctx mgr

fix: drop enable_idempotence — not supported by kafka-python-ng

fix: event-simulator idle by default (tail -f /dev/null), run on demand via make simulate

fix: apache/kafka:3.9.0 + python:3.12-slim+openjdk-21 spark, python healthcheck for api

docs: design spec and README with resume talking points

fix: HF Spaces — add psycopg2/redis deps, fix source label, remove dead code

feat: HuggingFace Spaces lite container (ALS-only, SQLite, same API contract)

feat: Prometheus scrape config and Grafana dashboard (latency, cache hit)

fix: stub prometheus.yml, grafana dirs, frontend waits for api healthy

feat: docker-compose (KRaft Kafka, Spark local, 9 services) and Makefile

fix: Dockerfiles — use non-root user and pip install from pyproject.toml

chore: Dockerfiles for all services

feat: React frontend (Vite+TypeScript+Tailwind, polling hook, movie cards)

fix: API quality fixes (await redis, Kafka lifecycle, Prometheus dedup, lifespan)

feat: FastAPI service with Redis/Postgres/ALS fallback chain and Prometheus metrics

feat: Kafka event simulator with idempotent producer and event schema

fix: streaming job quality fixes (pg reconnect, watermark docs, edge cases)

feat: Spark Structured Streaming with time-decay re-ranking (foreachBatch)

fix: bootstrap quality fixes (unused param, sim clip, stronger tests)

feat: ALS bootstrap with item similarity and Redis/Postgres seeding

feat: postgresql schema (movies, als_candidates, user_recs)

fix: uv workspace project table, workspace deps in tests, kafka-python-ng

chore: repo scaffolding, uv workspace, project structure