#!/usr/bin/env python3
"""
Embed project docs, skills, agents, and rules for deep-dive and RAG.
Ingests docs/, .cursor/skills/, .cursor/agents/, README, CLAUDE.md, worktrees.json,
.cursor/rules; chunks content; optionally generates embeddings and writes index to
.cursor/embeddings/. Use real content only (no synthetic data). See
docs/EMBEDDINGS_AND_PROMPTS_PLAN.md.
Run: uv run python scripts/embed_docs.py [--dry-run]
"""
from __future__ import annotations
import argparse
import json
import sys
from pathlib import Path
REPO_ROOT = Path(__file__).resolve().parent.parent
CURSOR_EMBED = REPO_ROOT / ".cursor" / "embeddings"
CURSOR_PROMPTS = REPO_ROOT / ".cursor" / "prompts"
def repo_root() -> Path:
"""Project root (pyproject.toml or .cursor)."""
cwd = Path.cwd()
for d in [cwd, *cwd.parents]:
if (d / "pyproject.toml").exists() or (d / ".cursor").is_dir():
return d
return REPO_ROOT
def ensure_dirs(root: Path) -> None:
"""Create .cursor/embeddings and .cursor/prompts if missing."""
(root / ".cursor" / "embeddings").mkdir(parents=True, exist_ok=True)
(root / ".cursor" / "prompts").mkdir(parents=True, exist_ok=True)
def collect_sources(root: Path) -> list[tuple[str, Path]]:
"""Collect paths to embed: docs, skills, agents, key files."""
sources: list[tuple[str, Path]] = []
# Project docs
docs_dir = root / "docs"
if docs_dir.is_dir():
for p in sorted(docs_dir.glob("*.md")):
sources.append(("doc", p))
for name in ["README.md", "CLAUDE.md", "DOCKER.md"]:
p = root / name
if p.is_file():
sources.append(("doc", p))
# Cursor skills
skills_dir = root / ".cursor" / "skills"
if skills_dir.is_dir():
for d in sorted(skills_dir.iterdir()):
if d.is_dir():
sk = d / "SKILL.md"
if sk.is_file():
sources.append(("skill", sk))
# Cursor agents
agents_dir = root / ".cursor" / "agents"
if agents_dir.is_dir():
for p in sorted(agents_dir.glob("*.md")):
sources.append(("agent", p))
# Rules and worktrees
rules_dir = root / ".cursor" / "rules"
if rules_dir.is_dir():
for p in sorted(rules_dir.glob("*.mdc")):
sources.append(("rule", p))
wt = root / ".cursor" / "worktrees.json"
if wt.is_file():
sources.append(("worktree", wt))
return sources
def run_ingest(root: Path, dry_run: bool) -> dict:
"""Ingest sources and write minimal index/meta (Phase 1 placeholder)."""
ensure_dirs(root)
sources = collect_sources(root)
meta: list[dict] = []
for typ, path in sources:
rel = path.relative_to(root)
meta.append({
"source": str(rel),
"type": typ,
"path": str(path),
})
out_index = root / ".cursor" / "embeddings" / "index.json"
out_meta = root / ".cursor" / "embeddings" / "meta.json"
if not dry_run:
out_index.write_text(json.dumps({"chunks": [], "version": "0.1.0"}, indent=2))
out_meta.write_text(json.dumps({"sources": meta, "version": "0.1.0"}, indent=2))
return {"sources": len(sources), "meta_entries": len(meta), "dry_run": dry_run}
def main() -> int:
parser = argparse.ArgumentParser(description="Embed project docs for deep-dive/RAG")
parser.add_argument("--dry-run", action="store_true", help="Only list sources, do not write")
args = parser.parse_args()
root = repo_root()
result = run_ingest(root, args.dry_run)
print(f"embed_docs: {result['sources']} sources, {result['meta_entries']} meta entries (dry_run={result['dry_run']})")
if not args.dry_run:
print("Wrote .cursor/embeddings/index.json and meta.json (placeholder). Full embedding in Phase 1.")
return 0
if __name__ == "__main__":
sys.exit(main())
