Add script for fetching historical versions

fetch_history.py

@@ -0,0 +1,63 @@
+import csv
+import json
+import os
+import sys
+
+import requests
+import requests_ratelimiter
+import tqdm
+
+
+def get_commits(session, repo, path):
+    query = {"path": path}
+    headers = {
+        "Accept": "application/vnd.github+json",
+        "Authorization": "Bearer " + os.environ["GITHUB_TOKEN"],
+        "X-GitHub-Api-Version": "2022-11-28",
+    }
+
+    try:
+        r = session.get(
+            "https://api.github.com/repos/" + repo + "/commits",
+            params=query,
+            headers=headers,
+            timeout=10,
+        )
+    except (
+        requests.exceptions.ConnectionError,
+        requests.exceptions.ReadTimeout,
+    ):
+        # Skip on request error
+        return set()
+    else:
+        # Get the commit hashes
+        return set(c["sha"] for c in r.json())
+
+
+def main():
+    # Initialize a new session
+    session = requests.Session()
+    adapter = requests_ratelimiter.LimiterAdapter(per_second=2)
+    session.mount("http://", adapter)
+    session.mount("https://", adapter)
+
+    with open("repos.csv", "r") as csvfile:
+        # Count number of rows and reset
+        reader = csv.DictReader(csvfile)
+        rows = sum(1 for row in reader)
+        csvfile.seek(0)
+
+        reader = csv.DictReader(csvfile)
+        for row in tqdm.tqdm(reader, total=rows):
+            # Remove github.com/ from the beginning and fetch commits
+            repo = row["repository"].split("/", maxsplit=1)[1]
+            commits = get_commits(session, repo, row["path"])
+
+            # Write the collected commits
+            obj = {"repository": repo, "path": row["path"], "commits": list(commits)}
+            json.dump(obj, sys.stdout)
+            sys.stdout.write("\n")
+
+
+if __name__ == "__main__":
+    main()