Exclude private repositories

[yarikoptic]

For the purpose of this service we should avoid indexing private repositories. Since for some services we might require using API keys, if service returns private ones, we should avoid it "explicitly".

From prior "research" it seems that forgejo has query option to avoid listing private repos, so it should be used while querying such services. And such option is absent on GIN so we might not be able to avoid listing them but we should be able to filter out post-fact using private attribute.

[jwodder]

@yarikoptic For the record, using the below script, I was unable to find any private DataLad-using repositories on GIN or hub.datalad.org that were listed in datalad-repos.json:

#!/usr/bin/env -S pipx run
# /// script
# requires-python = ">=3.11"
# dependencies = ["ghreq ~= 0.1", "python-dotenv ~= 1.0"]
# ///

from __future__ import annotations
import json
import os
from pathlib import Path
from dotenv import load_dotenv
import ghreq

REPO_DATABASE = Path("datalad-repos.json")


def main() -> None:
    load_dotenv()
    with REPO_DATABASE.open() as fp:
        db = json.load(fp)
    db["gin"] = rm_private(
        db["gin"], "https://gin.g-node.org/api/v1/", os.environ["GIN_TOKEN"]
    )
    db["hub_datalad_org"] = rm_private(
        db["hub_datalad_org"],
        "https://hub.datalad.org/api/v1/",
        os.environ["HUB_DATALAD_ORG_TOKEN"],
    )
    with REPO_DATABASE.open("w") as fp:
        print(json.dumps(db, indent=4), file=fp)


def rm_private(repos: list, api_url: str, token: str) -> list:
    public = []
    with ghreq.Client(
        api_url=api_url,
        accept=None,
        api_version=None,
        headers={"Authorization": f"token {token}"},
    ) as client:
        for r in repos:
            if r["status"] == "gone":
                public.append(r)
            else:
                data = client.get(f"/repos/{r['name']}")
                if data["private"]:
                    print(f"Found private repo r['name'] on {api_url}")
                else:
                    public.append(r)
    return public


if __name__ == "__main__":
    main()