1 files changed, 133 insertions, 0 deletions
diff --git a/koya b/koya
new file mode 100755
index 0000000..826b344
--- /dev/null
+++ b/koya
@@ -0,0 +1,133 @@
+#!/usr/bin/env python3
+"""koya
+
+Usage:
+  koya --token <token> user1 user2...
+
+Options:
+  -h, --help       show this help message and exit
+  --origin ORIGIN  sourcehut origin URL (default: https://git.sr.ht/)
+  --token TOKEN    account oauth access token (default: None)
+  --delay DELAY    delay between origin requests (seconds) (default: 1)
+"""
+
+import argparse
+import json
+import logging
+import os
+import subprocess
+import sys
+import time
+from pathlib import Path
+from urllib.parse import urljoin, urlparse, quote_plus
+
+import requests
+
+logging.basicConfig(
+    stream=sys.stdout, format='%(asctime)s %(message)s',
+    datefmt='%m/%d/%Y %H:%M:%S')
+log = logging.getLogger('koya')
+log.setLevel(logging.DEBUG)
+
+class Koya:
+    def __init__(self, origin, token, delay):
+        self.origin = origin
+        self.token = token
+        self.delay = delay
+
+        self.session = requests.Session()
+        self.session.headers.update({"User-Agent": "koya/1.0 git.jordan.im/koya"})
+
+    def list_repos(self, user):
+        headers = {
+            "Content-Type": "application/json",
+            "Authorization": f"Bearer {self.token}"}
+        repos = []
+        cursor_obj = ""
+        while True:
+            graphql = f"""
+                query {{
+                  user(username: "{user}") {{
+                    repositories({cursor_obj}) {{
+                      cursor
+                      results {{
+                        name
+                      }}
+                    }}
+                  }}
+                }}"""
+            try:
+                u = urljoin(self.origin, "/query")
+                r = self.session.post(u, headers=headers, json={"query": graphql})
+                r.raise_for_status()
+            except Exception as err:
+                log.error(err)
+                return
+
+            res = json.loads(r.text)
+            if not res["data"].get("user"):
+                return repos
+
+            data = res["data"]["user"]["repositories"]
+            repos.extend(data["results"])
+
+            cursor = data.get("cursor")
+            if cursor:
+                cursor_obj = f'''cursor: "{cursor}"'''
+            else:
+                return repos
+
+            log.info(f"Discovered {len(data['results'])} repositories...")
+            time.sleep(self.delay)
+
+def multi_urljoin(*parts):
+    ret = urljoin(parts[0], "/".join(
+        quote_plus(part.strip("/"), safe="/") for part in parts[1:]))
+    return ret
+
+if __name__ == "__main__":
+    DESC = "koya: archive git repositories from sourcehut user accounts"
+    parser = argparse.ArgumentParser(
+        description=DESC,
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+    parser.add_argument(
+        "users", nargs="+", default=os.getcwd(), help="username(s) to archive")
+    parser.add_argument(
+        "--origin", dest="origin", type=str, action="store",
+        default="https://git.sr.ht/", required=False,
+        help="sourcehut origin URL")
+    parser.add_argument(
+        "--token", dest="token", type=str, action="store",
+        help="account oauth access token", required=True)
+    parser.add_argument(
+        "--delay", dest="delay", type=int, action="store", default=1,
+        help="delay between origin requests (seconds)")
+    args=parser.parse_args()
+
+    host = urlparse(args.origin).hostname
+    if not host:
+        log.error("Unable to parse hostname from origin, exiting...")
+        sys.exit()
+
+    koya = Koya(args.origin, args.token, args.delay)
+    for user in args.users:
+        repos = koya.list_repos(user)
+        if not repos:
+            log.error(f"No repositories discovered, skipping {user}...")
+            continue
+
+        for repo in repos:
+            repo_path = os.path.join(host, user, repo["name"])
+            if os.path.exists(repo_path):
+                fetch = ["git", "fetch", "--all", "--force", "--tags", "--prune"]
+                subprocess.run(fetch, cwd=repo_path)
+
+                time.sleep(args.delay)
+            else:
+                Path(repo_path).mkdir(parents=True, exist_ok=True)
+
+                remote_url = multi_urljoin(args.origin, "~"+user, repo["name"])
+                clone = ["git", "clone", remote_url, repo_path]
+                subprocess.run(clone)
+
+                time.sleep(args.delay)