Switchyard lets LLM applications route traffic across models and providers while preserving native OpenAI and Anthropic API compatibility - enabling flexible model selection, benchmarking, and cost/performance optimization.
Rust
#!/usr/bin/env python3
# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
"""Fetch GitHub metadata for a list of repo URLs and output a craft-repos-style CSV.
Usage:
# 1. Create a text file with one GitHub URL per line, e.g.:
# https://github.com/pallets/jinja
# https://github.com/psf/requests
#
# 2. Run:
# python3 scripts/fetch_repo_details.py /path/to/urls.txt --out /path/to/output.csv
# python3 scripts/fetch_repo_details.py https://github.com/pallets/jinja --out /path/to/output.csv
#
# Requires: gh CLI authenticated (gh auth login)
# Output columns: short_name, github_repo, github_url, stars, license, domain, description
"""
from __future__ import annotations
import argparse
import csv
import json
import re
import subprocess
import sys
from pathlib import Path
# First-match wins — order matters
DOMAIN_RULES = [
(
"AI/ML",
[
"llm",
"gpt",
"neural",
"nlp",
"embedding",
"rag",
"agent",
"voice",
"speech",
"asr",
"tts",
"machine learning",
"machine-learning",
"scikit",
"sklearn",
"deep learning",
"deep-learning",
"pytorch",
"tensorflow",
"keras",
"xgboost",
"classifier",
"regression",
"clustering",
],
),
(
"Web/API",
[
"web",
"http",
"api",
"server",
"fastapi",
"django",
"flask",
"websocket",
"rest",
"graphql",
"socket",
"proxy",
"tunnel",
"vpn",
"socks",
"wsgi",
"asgi",
],
),
("CLI/TUI", ["cli", "terminal", "tui", "command-line", "shell", "prompt", "console", "repl", "curses"]),
(
"Data/DB",
[
"database",
"sql",
"data",
"pandas",
"csv",
"json",
"parser",
"parse",
"format",
"markdown",
"text",
"pdf",
"extract",
"index",
"search",
"protobuf",
],
),
(
"Infra/System",
[
"docker",
"deploy",
"cloud",
"infra",
"system",
"process",
"thread",
"async",
"queue",
"task",
"worker",
"archive",
"compress",
"audio",
"video",
"stream",
"codec",
],
),
]
def classify_domain(short_name: str, description: str, topics: list[str]) -> str:
text = (short_name + " " + (description or "") + " " + " ".join(topics or [])).lower()
for domain, keywords in DOMAIN_RULES:
if any(kw in text for kw in keywords):
return domain
return "Other"
def _parse_github_url(url: str) -> str:
m = re.match(r"https?://github\.com/([^/]+)/([^/]+?)(?:\.git)?/?$", url)
if not m:
sys.exit(f"ERROR: not a valid bare GitHub repo URL: {url}\n Expected: https://github.com/owner/repo")
return f"{m.group(1)}/{m.group(2)}"
def fetch_repo(github_repo: str) -> dict | None:
result = subprocess.run(
[
"gh",
"api",
f"repos/{github_repo}",
"--jq",
'[.name, (.description // ""), .stargazers_count,'
' (.license.spdx_id // "None"), .html_url, (.topics // [])]',
],
capture_output=True,
text=True,
)
if result.returncode != 0:
print(f" ERROR {github_repo}: {result.stderr.strip()}", file=sys.stderr)
return None
try:
name, description, stars, license_id, url, topics = json.loads(result.stdout)
except (json.JSONDecodeError, ValueError) as e:
print(f" ERROR {github_repo}: unexpected API response: {e}", file=sys.stderr)
return None
return {
"name": name,
"description": description,
"stars": stars,
"license": license_id,
"url": url,
"topics": topics,
}
def main() -> None:
parser = argparse.ArgumentParser(description="Fetch GitHub metadata and output craft-repos-style CSV.")
parser.add_argument("input", help="GitHub URL or file of URLs (one per line)")
parser.add_argument(
"--out", type=Path, default=None, help="Output CSV (default: stdout for URL, <input>.csv for file)"
)
args = parser.parse_args()
if args.input.startswith("http"):
urls = [args.input.strip()]
out_file: Path | None = args.out
else:
urls_file = Path(args.input)
out_file = args.out or urls_file.with_suffix(".csv")
with urls_file.open() as f:
urls = [line.strip() for line in f.read().splitlines() if line.strip().startswith("http")]
rows = []
failed = []
for url in urls:
github_repo = _parse_github_url(url)
print(f"Fetching {github_repo}...", file=sys.stderr)
info = fetch_repo(github_repo)
if not info:
failed.append(url)
continue
domain = classify_domain(info["name"], info["description"], info["topics"])
rows.append(
{
"short_name": info["name"],
"github_repo": github_repo,
"github_url": info["url"],
"stars": info["stars"],
"license": info["license"],
"domain": domain,
"description": info["description"],
}
)
fieldnames = ["short_name", "github_repo", "github_url", "stars", "license", "domain", "description"]
if out_file:
with open(out_file, "w", newline="") as f:
writer = csv.DictWriter(f, fieldnames=fieldnames)
writer.writeheader()
writer.writerows(rows)
print(f"\nWrote {len(rows)}/{len(urls)} repos to {out_file}", file=sys.stderr)
else:
writer = csv.DictWriter(sys.stdout, fieldnames=fieldnames)
writer.writeheader()
writer.writerows(rows)
if failed:
print(f"\nERROR: {len(failed)} fetch(es) failed:", file=sys.stderr)
for u in failed:
print(f" {u}", file=sys.stderr)
sys.exit(1)
if __name__ == "__main__":
main()