WIP/RF: Rewrite update_zenodo to use commit count instead of line count

Author: effigies

tools/update_zenodo.py

@@ -1,77 +1,60 @@
 #!/usr/bin/env python3
 """Update and sort the creators list of the zenodo record."""
-import sys
-import shutil
-from pathlib import Path
+import git
 import json
+from subprocess import run, PIPE
+from pathlib import Path
 from fuzzywuzzy import fuzz, process
-import subprocess as sp
 
-# These names should go last
-CREATORS_LAST = ["Gorgolewski, Krzysztof J.", "Ghosh, Satrajit"]
 
-if __name__ == "__main__":
-    contrib_file = Path("line-contributors.txt")
-    lines = []
-    if contrib_file.exists():
-        print("WARNING: Reusing existing line-contributors.txt file.", file=sys.stderr)
-        lines = contrib_file.read_text().splitlines()
-
-    git_line_summary_path = shutil.which("git-line-summary")
-    if not lines and git_line_summary_path:
-        print("Running git-line-summary on nipype repo")
-        lines = sp.check_output([git_line_summary_path]).decode().splitlines()
-        contrib_file.write_text("\n".join(lines))
-
-    if not lines:
-        raise RuntimeError(
-            """\
-Could not find line-contributors from git repository.%s"""
-            % """ \
-git-line-summary not found, please install git-extras. """
-            * (git_line_summary_path is None)
-        )
+def decommify(name):
+    return " ".join(name.split(", ")[::-1])
 
-    data = [" ".join(line.strip().split()[1:-1]) for line in lines if "%" in line]
 
-    # load zenodo from master
-    zenodo_file = Path(".zenodo.json")
-    zenodo = json.loads(zenodo_file.read_text())
-    zen_names = [
-        " ".join(val["name"].split(",")[::-1]).strip() for val in zenodo["creators"]
-    ]
-    total_names = len(zen_names)
+# These names should go last
+CREATORS_LAST = ["Krzysztof J. Gorgolewski", "Satrajit Ghosh"]
 
-    name_matches = []
-    position = 1
-    for ele in data:
-        matches = process.extract(ele, zen_names, scorer=fuzz.token_sort_ratio, limit=2)
-        # matches is a list [('First match', % Match), ('Second match', % Match)]
+if __name__ == "__main__":
+    git_root = Path(git.Repo(".", search_parent_directories=True).working_dir)
+    zenodo_file = git_root / ".zenodo.json"
+
+    zenodo = json.loads(zenodo_file.read_text()) if zenodo_file.exists() else {}
+
+    creator_map = {
+        decommify(creator["name"]): creator for creator in zenodo.get("creators", [])
+    }
+
+    shortlog = run(["git", "shortlog", "-ns"], stdout=PIPE)
+    commit_counts = dict(
+        line.split("\t", 1)[::-1]
+        for line in shortlog.stdout.decode().split("\n")
+        if line
+    )
+
+    committers = []
+
+    # Stable sort:
+    # Number of commits in descending order
+    # Ties broken by alphabetical order of first name
+    for committer, _ in sorted(commit_counts.items(), key=lambda x: (-int(x[1]), x[0])):
+        matches = process.extract(
+            committer, creator_map.keys(), scorer=fuzz.token_sort_ratio, limit=2
+        )
         if matches[0][1] <= 80:
             # skip unmatched names
-            print("No entry to sort:", ele)
+            print("No entry to sort:", committer)
             continue
+        committers.append(matches[0][0])
 
-        idx = zen_names.index(matches[0][0])
-        val = zenodo["creators"][idx]
-
-        if val not in name_matches:
-            if val["name"] not in CREATORS_LAST:
-                val["position"] = position
-                position += 1
-            else:
-                val["position"] = total_names + CREATORS_LAST.index(val["name"])
-            name_matches.append(val)
+    for last_author in CREATORS_LAST:
+        if committers[-1] != last_author:
+            committers.remove(last_author)
+            committers.append(last_author)
 
-    for missing in zenodo["creators"]:
-        if "position" not in missing:
-            missing["position"] = position
-            position += 1
-            name_matches.append(missing)
+    creators = [
+        creator_map.get(committer, {"name": committer}) for committer in committers
+    ]
 
-    zenodo["creators"] = sorted(name_matches, key=lambda k: k["position"])
-    # Remove position
-    for creator in zenodo["creators"]:
-        del creator["position"]
+    zenodo["creators"] = creators
 
-    zenodo_file.write_text("%s\n" % json.dumps(zenodo, indent=2, sort_keys=True))
+    zenodo_file.write_text("%s\n" % json.dumps(zenodo, indent=2))