Image scraping + Postgres env vars + search results include images

- _fetch_and_index now extracts img src URLs, filters junk, stores in OpenSearch
- get_db() uses PGHOST/PGUSER/etc env vars for Docker compatibility
- search endpoint returns images + image_count in hits
- Docker run command passes all PG env vars
- Discovery script runs via SSH to CT to avoid VPN issues
This commit is contained in:
drjones
2026-08-05 16:28:21 -07:00
parent 42abcf39f1
commit 2ac65db518
2 changed files with 64 additions and 10 deletions

View File

@@ -56,7 +56,46 @@ services:
timeout: 5s
retries: 3
# ── Database ───────────────────────────────────────────────
postgres:
image: postgres:16-alpine
container_name: postgres
environment:
- POSTGRES_USER=research
- POSTGRES_PASSWORD=ResearchDB2026!
- POSTGRES_DB=research_engine
ports:
- "5432:5432"
volumes:
- postgres_data:/var/lib/postgresql/data
restart: unless-stopped
healthcheck:
test: ["CMD", "pg_isready", "-U", "research"]
interval: 10s
timeout: 5s
retries: 5
# ── API + Dashboard ────────────────────────────────────────
backend:
build: .
container_name: research-backend
network_mode: host
restart: unless-stopped
depends_on:
opensearch:
condition: service_healthy
postgres:
condition: service_healthy
redis:
condition: service_healthy
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/"]
interval: 30s
timeout: 10s
retries: 3
volumes:
yacy_data:
opensearch_data:
redis_data:
postgres_data: