Files
hermes-agent/website/scripts/extract-plugins.py
teknium1 4c2c20f6e7 feat(website): plugin catalog shows when each entry was added/updated and can sort by it
extract-plugins.py derives addedAt/updatedAt per entry from one `git log
--name-status` over plugin-catalog/ (committer dates, so rebase-merged PRs
are stamped when they land; renames carry the original addedAt). A shallow
checkout or missing git yields null dates instead of wrong ones, and
deploy-site.yml now checks out with fetch-depth: 0 so the deploy has the
history.

The /docs/plugins page gains a Sort control (Most starred / Newest /
Recently updated) and an "Added … · Updated …" line on every card; Updated
is hidden while it equals Added.
2026-09-20 20:18:58 -07:00

336 lines
14 KiB
Python

#!/usr/bin/env python3
"""Extract plugin-catalog entries into website/static/api/plugins.json.
Feeds the Plugin Catalog page at /docs/plugins (website/src/pages/plugins/).
Data source: ``plugin-catalog/*.yaml`` at the repo root — one YAML file per
catalog entry (see plugin-catalog/README.md for the entry schema), plus
``plugin-catalog/removed.yaml`` listing plugins pulled from the catalog.
No network, no crawling: the catalog is human-merged data in the checkout.
Graceful degradation: when ``plugin-catalog/`` does not exist yet (the
catalog PR may not have merged), we emit an EMPTY catalog list and zeroed
meta counts and exit 0 so the docs build stays green. The page renders a
"catalog is just getting started" state.
Outputs (both under website/static/api/, CDN-served at /docs/api/):
- ``plugins.json`` — list of catalog entries for the page (camelCase)
- ``plugins-meta.json`` — counts by tier + generatedAt + removedCount
- ``plugin-catalog.json`` — ``{"entries": [raw YAML mappings], "removed": [...]}`` in the loader's own
schema; installed Hermes clients fetch this for live catalog refresh
(``hermes_cli.plugin_catalog.LIVE_CATALOG_URL``) so new entries and removals reach them without
updating. Emitting it here means the docs deploy IS the publish step — no second pipeline.
"""
from __future__ import annotations
import argparse
import json
import re
import subprocess
import sys
from collections import Counter
from datetime import datetime, timezone
from pathlib import Path
from urllib.parse import urlsplit
import yaml
REPO_ROOT = Path(__file__).resolve().parents[2]
DEFAULT_CATALOG_DIR = REPO_ROOT / "plugin-catalog"
DEFAULT_OUTPUT_DIR = REPO_ROOT / "website" / "static" / "api"
# Written by fetch-plugin-stars.py (at most one GitHub probe per day); absent → no ranking data.
DEFAULT_STARS_FILE = DEFAULT_OUTPUT_DIR / "plugin-stars.json"
_GITHUB_REPO_RE = re.compile(r"^https://github\.com/([^/\s]+)/([^/\s#?]+?)(?:\.git)?/?$")
CATALOG_TIERS = ("official", "community")
CATALOG_CATEGORIES = ("desktop", "memory", "platform", "web", "tools", "voice", "automation", "models", "general")
SHA_RE = re.compile(r"^[0-9a-f]{40}$")
# Keep in sync with scripts/validate_plugin_catalog.py (cosmetic fields attached to the pin).
VERSION_RE = re.compile(r"^[A-Za-z0-9][A-Za-z0-9._+-]{0,31}$")
IMAGE_HOSTS = ("raw.githubusercontent.com", "github.com")
IMAGE_HOST_SUFFIX = ".githubusercontent.com"
def _log(msg: str) -> None:
print(f"[extract-plugins] {msg}", file=sys.stderr)
def _is_allowed_image_url(url: str) -> bool:
parts = urlsplit(url)
host = (parts.hostname or "").lower()
return parts.scheme == "https" and bool(host) and (host in IMAGE_HOSTS or host.endswith(IMAGE_HOST_SUFFIX))
def _cosmetic(value, accept, file_name: str, entry: str, key: str) -> str:
"""A cosmetic field is dropped, never fatal: the site must not lose an entry a reviewer merged."""
text = str(value or "").strip()
if text and not accept(text):
_log(f"{file_name} ({entry}): dropping invalid {key} {text!r}")
return ""
return text
def _str_list(value) -> list[str]:
if isinstance(value, str):
return [value] if value.strip() else []
if isinstance(value, list):
return [str(x) for x in value if x]
return []
def _normalize_capabilities(raw) -> dict:
raw = raw if isinstance(raw, dict) else {}
return {
"providesTools": _str_list(raw.get("provides_tools")),
"providesHooks": _str_list(raw.get("provides_hooks")),
"providesMiddleware": _str_list(raw.get("provides_middleware")),
"requiresEnv": _str_list(raw.get("requires_env")),
}
def load_stars(stars_file: Path) -> dict[str, int]:
"""``{"owner/repo": stars}`` from fetch-plugin-stars.py's cache; empty when missing/unreadable."""
try:
data = json.loads(stars_file.read_text(encoding="utf-8"))
except (OSError, ValueError):
return {}
raw = data.get("stars") if isinstance(data, dict) else None
return {str(k): int(v) for k, v in raw.items() if isinstance(v, (int, float))} if isinstance(raw, dict) else {}
def _repo_stars(repo: str, stars: dict[str, int]) -> int | None:
m = _GITHUB_REPO_RE.match(repo)
return stars.get(f"{m.group(1)}/{m.group(2)}") if m else None
def load_git_dates(catalog_dir: Path) -> dict[str, dict[str, str]]:
"""``{"<file>.yaml": {"addedAt": iso, "updatedAt": iso}}`` from the catalog's git history.
One ``git log`` over the directory, newest first: the first commit seen touching a file is
its last update, the last one is when it entered the catalog. Renames (``R``) carry the
old path's history onto the new name so a renamed entry keeps its original addedAt.
Committer dates, not author dates: rebase-merged PRs are stamped when they LAND on main,
which is when the entry actually became installable.
Empty when git is unavailable or the checkout is shallow — a depth-1 clone would report
every entry as added in the tip commit, which is worse than no dates (deploy-site.yml
checks out with ``fetch-depth: 0`` for this reason).
"""
if not catalog_dir.is_dir():
return {}
try:
shallow = subprocess.run(
["git", "rev-parse", "--is-shallow-repository"],
cwd=catalog_dir, capture_output=True, text=True, check=True, timeout=30,
).stdout.strip()
if shallow == "true":
_log("shallow checkout: skipping addedAt/updatedAt (need fetch-depth: 0)")
return {}
log = subprocess.run(
["git", "log", "--format=%x00%cI", "--name-status", "--relative", "-M", "--", "."],
cwd=catalog_dir, capture_output=True, text=True, check=True, timeout=120,
).stdout
except (OSError, subprocess.SubprocessError) as e:
_log(f"git history unavailable, skipping addedAt/updatedAt ({e})")
return {}
dates: dict[str, dict[str, str]] = {}
alias: dict[str, str] = {} # old file name → current name, for renamed entries
def touch(name: str, when: str) -> None:
if "/" in name or not name.endswith(".yaml"):
return
while name in alias:
name = alias[name]
rec = dates.setdefault(name, {"addedAt": when, "updatedAt": when})
rec["addedAt"] = when # newest-first walk: the last write wins = oldest commit
when = ""
for line in log.splitlines():
if line.startswith("\x00"):
when = line[1:].strip()
continue
parts = line.split("\t")
if len(parts) < 2 or not when:
continue
if parts[0].startswith("R") and len(parts) == 3:
old, new = parts[1], parts[2]
touch(new, when)
alias[old] = new
continue
touch(parts[-1], when)
return dates
def load_catalog_entries(catalog_dir: Path, stars: dict[str, int] | None = None,
dates: dict[str, dict[str, str]] | None = None) -> list[dict]:
"""Parse all ``*.yaml`` files (except removed.yaml) into page entries.
Entries missing any of name/repo/sha are skipped with a stderr log —
a malformed community entry must never break the docs deploy.
``dates`` is ``load_git_dates()`` output keyed by file name; absent → null addedAt/updatedAt.
"""
entries: list[dict] = []
stars = stars or {}
dates = dates or {}
if not catalog_dir.is_dir():
return entries
for path in sorted(catalog_dir.glob("*.yaml")):
if path.name == "removed.yaml":
continue
try:
raw = yaml.safe_load(path.read_text(encoding="utf-8"))
except (yaml.YAMLError, OSError) as e:
_log(f"skipping {path.name}: unreadable YAML ({e})")
continue
if not isinstance(raw, dict):
_log(f"skipping {path.name}: not a mapping")
continue
name = str(raw.get("name") or "").strip()
repo = str(raw.get("repo") or "").strip()
sha = str(raw.get("sha") or "").strip().lower()
missing = [
field
for field, value in (("name", name), ("repo", repo), ("sha", sha))
if not value
]
if missing:
_log(f"skipping {path.name}: missing required field(s) {', '.join(missing)}")
continue
if not SHA_RE.match(sha):
_log(f"skipping {path.name} ({name}): sha is not a 40-hex commit pin")
continue
tier = str(raw.get("tier") or "community").strip().lower()
if tier not in CATALOG_TIERS:
_log(f"{path.name} ({name}): unknown tier {tier!r}, treating as community")
tier = "community"
category = str(raw.get("category") or "desktop").strip().lower()
if category not in CATALOG_CATEGORIES:
_log(f"{path.name} ({name}): unknown category {category!r}, treating as general")
category = "general"
entries.append({
"name": name,
"description": str(raw.get("description") or "").strip(),
"repo": repo,
"sha": sha,
"shaShort": sha[:7],
"tier": tier,
"category": category,
"maintainer": str(raw.get("maintainer") or "").strip(),
"subdir": str(raw.get("subdir") or "").strip(),
"requiresHermes": str(raw.get("requires_hermes") or "").strip(),
"platforms": _str_list(raw.get("platforms")),
"capabilities": _normalize_capabilities(raw.get("capabilities")),
"docsUrl": str(raw.get("docs_url") or "").strip(),
"version": _cosmetic(raw.get("version"), VERSION_RE.match, path.name, name, "version"),
"image": _cosmetic(raw.get("image"), _is_allowed_image_url, path.name, name, "image"),
"installCommand": f"hermes plugins install {name}",
"stars": _repo_stars(repo, stars),
"addedAt": dates.get(path.name, {}).get("addedAt"),
"updatedAt": dates.get(path.name, {}).get("updatedAt"),
})
# Most-starred first (unknown = 0), then name so the order is stable; tier does not rank.
entries.sort(key=lambda e: (-(e["stars"] or 0), e["name"]))
return entries
def _stars_fetched_at(stars_file: Path) -> str | None:
try:
data = json.loads(stars_file.read_text(encoding="utf-8"))
except (OSError, ValueError):
return None
return str(data.get("fetched_at")) if isinstance(data, dict) and data.get("fetched_at") else None
def load_removed(catalog_dir: Path) -> list[dict]:
"""``removed:`` list from plugin-catalog/removed.yaml (mappings only)."""
removed_path = catalog_dir / "removed.yaml"
if not removed_path.is_file():
return []
try:
raw = yaml.safe_load(removed_path.read_text(encoding="utf-8"))
except (yaml.YAMLError, OSError) as e:
_log(f"could not read removed.yaml: {e}")
return []
removed = raw.get("removed") if isinstance(raw, dict) else None
return [r for r in removed if isinstance(r, dict)] if isinstance(removed, list) else []
def count_removed(catalog_dir: Path) -> int:
return len(load_removed(catalog_dir))
def load_raw_entries(catalog_dir: Path) -> list[dict]:
"""Raw entry mappings (loader schema, snake_case) for ``plugin-catalog.json``; the client re-validates."""
entries: list[dict] = []
if not catalog_dir.is_dir():
return entries
for path in sorted(catalog_dir.glob("*.yaml")):
if path.name == "removed.yaml":
continue
try:
raw = yaml.safe_load(path.read_text(encoding="utf-8"))
except (yaml.YAMLError, OSError):
continue
if isinstance(raw, dict) and raw.get("name") and raw.get("repo") and raw.get("sha"):
entries.append(raw)
return entries
def main(catalog_dir: Path = DEFAULT_CATALOG_DIR, output_dir: Path = DEFAULT_OUTPUT_DIR,
stars_file: Path | None = None) -> int:
if not catalog_dir.is_dir():
_log(
f"plugin-catalog directory not found at {catalog_dir}; "
"emitting empty catalog (this is expected until the catalog lands)"
)
stars_path = stars_file if stars_file is not None else output_dir / "plugin-stars.json"
stars = load_stars(stars_path)
entries = load_catalog_entries(catalog_dir, stars, load_git_dates(catalog_dir))
removed_count = count_removed(catalog_dir)
by_tier = Counter(e["tier"] for e in entries)
by_category = Counter(e["category"] for e in entries)
meta = {
"generatedAt": datetime.now(timezone.utc).isoformat(),
"total": len(entries),
"byTier": {tier: by_tier.get(tier, 0) for tier in CATALOG_TIERS},
"byCategory": {c: by_category.get(c, 0) for c in CATALOG_CATEGORIES if by_category.get(c)},
"starsFetchedAt": _stars_fetched_at(stars_path),
"removedCount": removed_count,
}
output_dir.mkdir(parents=True, exist_ok=True)
with open(output_dir / "plugins.json", "w", encoding="utf-8") as f:
json.dump(entries, f, separators=(",", ":"), ensure_ascii=False)
with open(output_dir / "plugins-meta.json", "w", encoding="utf-8") as f:
json.dump(meta, f, separators=(",", ":"), ensure_ascii=False)
with open(output_dir / "plugin-catalog.json", "w", encoding="utf-8") as f:
json.dump({"generated_at": meta["generatedAt"], "entries": load_raw_entries(catalog_dir),
"removed": load_removed(catalog_dir)}, f, separators=(",", ":"), ensure_ascii=False)
print(
f"Extracted {len(entries)} plugin catalog entries "
f"({meta['byTier']['official']} official, {meta['byTier']['community']} community, "
f"{removed_count} removed) to {output_dir / 'plugins.json'}"
)
return 0
if __name__ == "__main__":
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--catalog-dir", type=Path, default=DEFAULT_CATALOG_DIR)
parser.add_argument("--output-dir", type=Path, default=DEFAULT_OUTPUT_DIR)
parser.add_argument("--stars-file", type=Path, default=None,
help="plugin-stars.json from fetch-plugin-stars.py (default: <output-dir>/plugin-stars.json)")
args = parser.parse_args()
sys.exit(main(catalog_dir=args.catalog_dir, output_dir=args.output_dir, stars_file=args.stars_file))