"""
Index a local skill into the skill-retriever data store.
Adds to metadata.json, FAISS vectors, and id_mapping.
Usage:
uv run python scripts/index_local_skill.py
"""
import json
import sys
from pathlib import Path
from datetime import datetime, timezone
import numpy as np
import faiss
from fastembed import TextEmbedding
# --- Config ---
SKILL_DIR = Path.home() / "repos" / "skill-retriever" / "skills" / "science-plots"
DATA_DIR = Path.home() / ".skill-retriever" / "data"
COMPONENT_ID = "AnthonyAlcaraz/claude-skills/skill/science-plots"
def parse_frontmatter(text):
"""Extract YAML frontmatter from markdown."""
lines = text.strip().split("\n")
if lines[0].strip() != "---":
return {}, text
end = -1
for i in range(1, len(lines)):
if lines[i].strip() == "---":
end = i
break
if end == -1:
return {}, text
fm = {}
for line in lines[1:end]:
if ":" in line:
key, val = line.split(":", 1)
fm[key.strip()] = val.strip().strip('"').strip("'")
body = "\n".join(lines[end + 1:]).strip()
return fm, body
def main():
# 1. Read skill content
skill_path = SKILL_DIR / "SKILL.md"
if not skill_path.exists():
print(f"ERROR: {skill_path} not found")
sys.exit(1)
raw = skill_path.read_text(encoding="utf-8")
fm, body = parse_frontmatter(raw)
# 2. Build metadata entry
entry = {
"id": COMPONENT_ID,
"name": "science-plots",
"component_type": "skill",
"description": fm.get("description", "Create publication-quality scientific figures using SciencePlots + matplotlib"),
"tags": ["matplotlib", "scientific-plotting", "ieee", "nature", "latex", "data-visualization"],
"author": fm.get("author", "AnthonyAlcaraz"),
"version": fm.get("version", "1.0"),
"last_updated": datetime.now(timezone.utc).isoformat(),
"commit_count": 1,
"commit_frequency_30d": 0.0,
"raw_content": raw,
"parameters": {},
"dependencies": [],
"tools": [],
"source_repo": "AnthonyAlcaraz/claude-skills",
"source_path": "skills/science-plots/SKILL.md",
"category": "data-visualization",
"install_url": None,
"security_risk_level": "safe",
"security_risk_score": 0.0,
"security_findings_count": 0,
"has_scripts": True
}
# 3. Add to metadata.json
meta_path = DATA_DIR / "metadata.json"
metadata = json.loads(meta_path.read_text(encoding="utf-8"))
# Remove existing entry if present (idempotent)
metadata = [m for m in metadata if m["id"] != COMPONENT_ID]
metadata.append(entry)
meta_path.write_text(json.dumps(metadata, indent=2, default=str), encoding="utf-8")
print(f"Metadata: added ({len(metadata)} total components)")
# 4. Generate embedding
print("Generating embedding with BAAI/bge-small-en-v1.5...")
model = TextEmbedding(model_name="BAAI/bge-small-en-v1.5")
embeddings = list(model.embed([raw]))
vector = np.array(embeddings[0], dtype=np.float32).reshape(1, -1)
print(f"Embedding shape: {vector.shape}")
# 5. Add to FAISS index
index_path = DATA_DIR / "vectors" / "faiss_index.bin"
mapping_path = DATA_DIR / "vectors" / "id_mapping.json"
index = faiss.read_index(str(index_path))
mapping = json.loads(mapping_path.read_text(encoding="utf-8"))
id_to_int = mapping["id_to_int"]
int_to_id = mapping["int_to_id"]
# Determine next integer ID
if id_to_int:
next_int = max(id_to_int.values()) + 1
else:
next_int = 0
# Remove existing if present (rebuild needed for true removal, but we can just skip re-adding)
if COMPONENT_ID in id_to_int:
print(f"Already in FAISS at index {id_to_int[COMPONENT_ID]}, updating mapping only")
else:
# IndexIDMap requires add_with_ids
ids = np.array([next_int], dtype=np.int64)
index.add_with_ids(vector, ids)
id_to_int[COMPONENT_ID] = next_int
int_to_id[str(next_int)] = COMPONENT_ID
print(f"FAISS: added at index {next_int} (total vectors: {index.ntotal})")
faiss.write_index(index, str(index_path))
mapping_path.write_text(json.dumps(mapping, indent=2), encoding="utf-8")
print("Vectors saved")
# 6. Add to graph.json
graph_path = DATA_DIR / "graph.json"
graph = json.loads(graph_path.read_text(encoding="utf-8"))
# Add node if not present
existing_ids = {n["id"] for n in graph.get("nodes", [])}
if COMPONENT_ID not in existing_ids:
graph["nodes"].append({
"id": COMPONENT_ID,
"label": "science-plots",
"component_type": "skill"
})
print(f"Graph: added node ({len(graph['nodes'])} total)")
else:
print("Graph: node already exists")
graph_path.write_text(json.dumps(graph, indent=2, default=str), encoding="utf-8")
print(f"\nDone! '{COMPONENT_ID}' indexed into skill-retriever.")
if __name__ == "__main__":
main()
