#!/usr/bin/env bash
# Entrypoint: orchestrate two worktrees (Canvas LMS content + Trustworthy AI presentation)
# or run MCP server. Aligns with GitHub/GitLab workflow tasks.
set -e
WORKTREE_CANVAS="${WORKTREE_CANVAS:-/app/worktrees/canvas-lms-content}"
WORKTREE_TRUSTWORTHY_AI="${WORKTREE_TRUSTWORTHY_AI:-/app/worktrees/trustworthy-ai-presentation}"
case "${1:-orchestrate}" in
server)
exec python3 /app/server.py "${@:2}"
;;
http)
exec python3 /app/server.py --transport streamable-http --port 8000 "${@:2}"
;;
embed)
if command -v uv >/dev/null 2>&1; then
cd /app && uv sync --all-extras 2>/dev/null || true
uv run python3 /app/scripts/embed_docs.py
for dir in /app/mcp/*/; do
[ -d "$dir" ] || continue
name=$(basename "$dir")
uv run python3 /app/scripts/embed_mcp_server.py "$name" 2>/dev/null || true
done
else
python3 /app/scripts/embed_docs.py
for dir in /app/mcp/*/; do
[ -d "$dir" ] || continue
name=$(basename "$dir")
python3 /app/scripts/embed_mcp_server.py "$name" 2>/dev/null || true
done
fi
exit 0
;;
orchestrate|"")
# Run UV-based setup (skills/agents discovery)
if command -v uv >/dev/null 2>&1; then
cd /app
uv sync --all-extras 2>/dev/null || true
uv run python3 /app/scripts/setup_worktree.py 2>/dev/null || true
else
python3 /app/scripts/setup_worktree.py 2>/dev/null || true
fi
# Run orchestration: two worktrees, skeleton/ToC, CI/CD plan, hypothesis/evidence, schema/ToC updates
exec python3 /app/scripts/docker_orchestrate.py \
--worktree-canvas "$WORKTREE_CANVAS" \
--worktree-trustworthy-ai "$WORKTREE_TRUSTWORTHY_AI"
;;
*)
exec "$@"
;;
esac
