fix: pin image versions, fix DB credentials, add hf_models placeholder

- docker-compose.yml: pin postgres to v16 (postgres:latest is v18+
  which changed data dir layout from /data to versioned subdir,
  causing container exit code 1 on existing volumes)
- docker-compose.yml: inject AIRFLOW__DATABASE__SQL_ALCHEMY_CONN via
  environment so password comes from .env at runtime instead of being
  hardcoded in the Dockerfile; fix default password mismatch
  (was 'airflow', now consistent with POSTGRES_PASSWORD default)
- airflow/Dockerfile: pin to apache/airflow:2.10.4 (was :latest);
  remove hardcoded SQL_ALCHEMY_CONN — now injected by docker-compose
- streamlit_app/Dockerfile: add --start-period=30s to HEALTHCHECK
  (Streamlit takes ~30s to boot; without it depends_on:healthy could
  fail prematurely before the app is ready)
- hf_models/.gitkeep: track placeholder so the directory exists on
  fresh clones — required by the text-completion-llm-service bind mount
- .gitignore: allow hf_models/.gitkeep (hf_models/* + !.gitkeep)
- Makefile: quickstart now runs mkdir -p hf_models to ensure the dir
  exists before docker compose build
- README.md: fix Grafana credentials (was 'admin/admin', now points to
  .env); add pinned versions to Tech Stack table

Author: VTvito

.gitignore

@@ -30,7 +30,8 @@ coverage.xml
 .ruff_cache/
 
 # ── Model cache ──
-hf_models/
+hf_models/*
+!hf_models/.gitkeep
 
 # ── Benchmark artifacts ──
 benchmark/data/

Makefile

@@ -11,6 +11,7 @@ quickstart: ## One-command setup: env → build → start → load demo data
 	@echo "══════════════════════════════════════════════════════"
 	@test -f .env || (cp .env.example .env && echo "✓ Created .env from .env.example")
 	@test -f .env && echo "✓ .env already exists"
+	@mkdir -p hf_models && echo "✓ hf_models/ directory ready (HuggingFace model cache)"
 	docker compose build
 	docker compose up -d
 	@echo "⏳ Waiting for services to be healthy..."

README.md

@@ -55,7 +55,7 @@ docker exec -it airflow airflow users create \
 |---|---|---|
 | **Streamlit** (AI Pipeline Builder) | http://localhost:8501 | — |
 | **Airflow** | http://localhost:8080 | admin / admin |
-| **Grafana** | http://localhost:3000 | admin / admin |
+| **Grafana** | http://localhost:3000 | admin / *GF_SECURITY_ADMIN_PASSWORD from .env* |
 | **Prometheus** | http://localhost:9090 | — |
 
 ### Try a Demo Pipeline
@@ -303,7 +303,7 @@ See [`.env.example`](.env.example) for all available variables including databas
 | Orchestration | Apache Airflow |
 | AI Agent | OpenAI / HuggingFace Transformers |
 | UI | Streamlit |
-| Containers | Docker, Docker Compose |
+| Containers | Docker, Docker Compose (PostgreSQL 16, Airflow 2.10.4) |
 | Monitoring | Prometheus + Grafana |
 | Testing | pytest, ruff |
 | CI/CD | GitHub Actions |

airflow/Dockerfile

@@ -1,8 +1,8 @@
-# Usa l'immagine ufficiale di Apache Airflow
-FROM apache/airflow:latest
+# Official Apache Airflow image (pinned for reproducibility)
+FROM apache/airflow:2.10.4
 
-# Imposta le variabili di ambiente per il database e Prometheus
-ENV AIRFLOW__DATABASE__SQL_ALCHEMY_CONN=postgresql+psycopg2://airflow:airflow@postgres/airflow
+# Metrics: export Airflow stats to Prometheus via statsd-exporter
+# DB connection string is injected at runtime by docker-compose (see AIRFLOW__DATABASE__SQL_ALCHEMY_CONN)
 ENV AIRFLOW__METRICS__STATSD_ON=True
 ENV AIRFLOW__METRICS__STATSD_HOST=statsd-exporter
 ENV AIRFLOW__METRICS__STATSD_PORT=9125

docker-compose.yml

@@ -1,6 +1,6 @@
 services:
   postgres:
-    image: postgres:latest
+    image: postgres:16
     container_name: postgres
     environment:
       POSTGRES_USER: ${POSTGRES_USER:-airflow}
@@ -24,6 +24,8 @@ services:
       context: ./airflow
       dockerfile: Dockerfile
     container_name: airflow
+    environment:
+      - AIRFLOW__DATABASE__SQL_ALCHEMY_CONN=postgresql+psycopg2://${POSTGRES_USER:-airflow}:${POSTGRES_PASSWORD:-change-me-strong-password}@postgres/${POSTGRES_DB:-airflow}
     depends_on:
       postgres:
         condition: service_healthy

hf_models/.gitkeep

@@ -16,7 +16,7 @@ ENV PYTHONPATH="/app"
 
 EXPOSE 8501
 
-HEALTHCHECK --interval=30s --timeout=10s --retries=3 \
+HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \
     CMD curl -f http://localhost:8501/_stcore/health || exit 1
 
 CMD ["streamlit", "run", "streamlit_app/app.py", \