#!/bin/bash
# Initial corpus build on the production node (run ON the node from ~/apps/ai-atlas). Idempotent; connectors are conditional
# (ETag / content hash) so re-running only fetches what changed. Priority 0 first (labs, hub, research, providers), then the rest.
set -euo pipefail
cd "$(dirname "$0")/.."
export PATH="/opt/homebrew/opt/postgresql@17/bin:/opt/homebrew/bin:$PATH"
export DATABASE_URL=${DATABASE_URL:-postgresql+asyncpg://aiatlas:aiatlas@127.0.0.1:5432/aiatlas}
export REDIS_URL=${REDIS_URL:-redis://127.0.0.1:6379/5}
export AIA_DATA_DIR=${AIA_DATA_DIR:-$HOME/ai-atlas-data}
export AIA_LOG_JSON=0
AIA=.venv/bin/aia

$AIA migrate | tail -1
$AIA seed | tail -1
echo "== all enabled connectors, once each, in priority order"
$AIA crawl --priority 9 || true
$AIA quality | tail -1
$AIA stats | head -40
$AIA status