feat: read git history with go-git, not the git binary (#15)

bin/tools/gitimport.py

@@ -1,21 +1,12 @@
-"""gitimport - parse `git log` output and turn commits into brain leafs.
+"""gitimport - Ladybug graph writes for Commit/File/Person (no git binary).
 
-Pure, testable functions. Field grammar (see bin/git/import):
-
-  git log --no-merges --name-only \
-      --format='%x1e%H%x1f%an%x1f%ae%x1f%aI%x1f%s'
-
-  0x1e = record separator, 0x1f = field separator.
-  Files: newline-separated lines following each record's subject.
+Commit records come from bin/git/import.go (go-git). This module only MERGEs
+the version graph File-[:HAS_VERSION]->Commit-[:AUTHORED]->Person.
 """
-
 from __future__ import annotations
 
 from dataclasses import dataclass, field
 
-REC_SEP = "\x1e"
-FIELD_SEP = "\x1f"
-
 
 @dataclass
 class Commit:
@@ -26,54 +17,6 @@ class Commit:
     subject: str
     files: list[str] = field(default_factory=list)
 
-    def leaf_text(self, repo: str) -> str:
-        head = f"commit {self.sha[:12]} in {repo} — {self.subject}"
-        body = [head, f"Author: {self.author} <{self.email}>", f"Date: {self.date}"]
-        if self.files:
-            body.append("Changing: " + ", ".join(self.files))
-        return "\n".join(body)
-
-
-def parse_log(text: str) -> list[Commit]:
-    """Parse `git log` output into Commit records.
-
-    Records are separated by 0x1e. A record is fields joined by 0x1f,
-    followed by optional newline-separated file paths inside the next
-    segment (git emits blank line + files after each record).
-    """
-    commits: list[Commit] = []
-    # field records and file lists alternate; simpler: split on REC_SEP,
-    # each chunk = header line, possibly followed by newline + files.
-    for chunk in text.split(REC_SEP):
-        chunk = chunk.strip("\n")
-        if not chunk:
-            continue
-        lines = chunk.split("\n", 1)
-        header = lines[0].split(FIELD_SEP)
-        if len(header) < 5:
-            continue
-        sha, author, email, date, subject = header[:5]
-        files = [ln.strip() for ln in lines[1].splitlines() if ln.strip()] if len(lines) > 1 else []
-        commits.append(Commit(sha=sha, author=author, email=email,
-                              date=date, subject=subject, files=files))
-    return commits
-
-
-def commits_to_leafs(commits: list[Commit], repo: str) -> list[dict]:
-    """Map commits to the leaf shape bin/kb/index expects (source/repo/...)."""
-    out: list[dict] = []
-    for c in commits:
-        out.append({
-            "source": f"{repo}@{c.sha}",
-            "repo": repo,
-            "heading": f"commit {c.sha[:12]} — {c.subject}",
-            "text": c.leaf_text(repo),
-            "type": "commit",
-            "status": "current",
-            "related": ",".join(c.files),
-        })
-    return out
-
 
 GIT_SCHEMA = (
     "CREATE NODE TABLE IF NOT EXISTS Commit (id STRING, repo STRING, subject STRING, "
@@ -114,4 +57,4 @@ def index_commits(conn, commits: list[Commit], repo: str) -> int:
             conn.execute("MATCH (f:File {id:$fid}), (c:Commit {id:$sha}) "
                          "MERGE (f)-[:HAS_VERSION]->(c)",
                          parameters={"fid": f"{repo}:{path}", "sha": c.sha})
-    return len(commits)
+    return len(commits)

bin/tools/test_bin_layout.py

@@ -91,3 +91,26 @@ class BinLayoutTest(unittest.TestCase):
 
     def test_postgres_query_is_shebang(self) -> None:
         self._assert_shebang("bin/postgres/query.go")
+
+    def test_git_import_is_gogit_shebang(self) -> None:
+        self._assert_shebang("bin/git/import.go")
+        py = (ROOT / "bin" / "git" / "import").read_text()
+        self.assertNotIn(
+            '["git"',
+            py,
+            "Python git/import must not subprocess the git binary",
+        )
+        self.assertIn("bin/git/import.go", py)
+
+    def test_gitimport_py_has_no_git_binary(self) -> None:
+        py = (ROOT / "bin" / "tools" / "gitimport.py").read_text()
+        self.assertNotIn("subprocess", py)
+        self.assertNotIn("git log", py)
+
+    def test_gogit_is_direct_go_mod_require(self) -> None:
+        text = (ROOT / "go.mod").read_text()
+        first = text.split("require (")[1].split(")")[0]
+        self.assertRegex(first, r"github.com/go-git/go-git/v5\s+v")
+        for line in first.splitlines():
+            if "go-git/go-git" in line:
+                self.assertNotIn("indirect", line)

bin/tools/test_gitgraph.py

@@ -9,12 +9,6 @@ sys.path.insert(0, str(Path(__file__).resolve().parent))
 import kblib  # noqa: E402
 import gitimport  # noqa: E402
 
-SAMPLE = (
-    "\x1e" + "a1b2c3d" + "\x1f" + "Ada Lovelace" + "\x1f" + "ada@example.com"
-    + "\x1f" + "2026-08-10T12:00:00+01:00" + "\x1f" + "feat: first commit"
-    + "\n\nREADME.md\nsrc/main.c\n"
-)
-
 COMMIT_PERSON_SCHEMA = (
     "CREATE NODE TABLE IF NOT EXISTS Commit (id STRING, repo STRING, subject STRING, "
     "author STRING, email STRING, date STRING, PRIMARY KEY(id))"
@@ -26,6 +20,17 @@ HAS_VERSION_SCHEMA = "CREATE REL TABLE IF NOT EXISTS HAS_VERSION (FROM File TO C
 AUTHORED_SCHEMA = "CREATE REL TABLE IF NOT EXISTS AUTHORED (FROM Commit TO Person)"
 
 
+def sample_commit() -> gitimport.Commit:
+    return gitimport.Commit(
+        sha="a1b2c3d",
+        author="Ada Lovelace",
+        email="ada@example.com",
+        date="2026-08-10T12:00:00+01:00",
+        subject="feat: first commit",
+        files=["README.md", "src/main.c"],
+    )
+
+
 class GitGraphTest(unittest.TestCase):
     def setUp(self):
         self.dir = tempfile.mkdtemp()
@@ -42,14 +47,12 @@ class GitGraphTest(unittest.TestCase):
         self.db.close()
 
     def test_index_commits_creates_nodes_and_edges(self):
-        cs = gitimport.parse_log(SAMPLE)
-        gitimport.index_commits(self.conn, cs, "sample-repo")
+        gitimport.index_commits(self.conn, [sample_commit()], "sample-repo")
         rp = self.conn.execute("MATCH (p:Person) RETURN p.name, p.email").get_all()
         self.assertEqual([tuple(r) for r in rp], [("Ada Lovelace", "ada@example.com")])
         rc = self.conn.execute("MATCH (c:Commit) RETURN c.id, c.repo").get_all()
         self.assertEqual(len(rc), 1)
         self.assertEqual(rc[0][1], "sample-repo")
-        # File -[:HAS_VERSION]-> Commit -[:AUTHORED]-> Person
         rf = self.conn.execute(
             "MATCH (f:File)-[:HAS_VERSION]->(c:Commit)-[:AUTHORED]->(p:Person) "
             "RETURN f.path, c.id, p.email").get_all()
@@ -58,7 +61,7 @@ class GitGraphTest(unittest.TestCase):
         self.assertTrue(all(r[2] == "ada@example.com" for r in rf))
 
     def test_index_commits_idempotent(self):
-        cs = gitimport.parse_log(SAMPLE)
+        cs = [sample_commit()]
         gitimport.index_commits(self.conn, cs, "sample-repo")
         gitimport.index_commits(self.conn, cs, "sample-repo")
         n = self.conn.execute("MATCH (c:Commit) RETURN count(*)").get_all()[0][0]
@@ -68,4 +71,4 @@ class GitGraphTest(unittest.TestCase):
 
 
 if __name__ == "__main__":
-    unittest.main()
+    unittest.main()

bin/tools/test_gitimport.py

@@ -1,57 +0,0 @@
-import sys
-import unittest
-from pathlib import Path
-
-sys.path.insert(0, str(Path(__file__).resolve().parent))
-
-import gitimport  # noqa: E402
-
-SAMPLE = (
-    "\x1e" + "a1b2c3d" + "\x1f" + "Ada Lovelace" + "\x1f" + "ada@example.com"
-    + "\x1f" + "2026-08-10T12:00:00+01:00" + "\x1f" + "feat: first commit"
-    + "\n\nREADME.md\nsrc/main.c\n"
-    + "\x1e" + "e4f5a6b" + "\x1f" + "Bob Babbage" + "\x1f" + "bob@example.com"
-    + "\x1f" + "2026-08-11T09:30:00+01:00" + "\x1f" + "fix: typo"
-    + "\n\ndocs/notes.md"
-)
-
-
-class GitparseTest(unittest.TestCase):
-    def test_parses_records(self):
-        cs = gitimport.parse_log(SAMPLE)
-        self.assertEqual(len(cs), 2)
-
-    def test_parses_commit_fields(self):
-        cs = gitimport.parse_log(SAMPLE)
-        c = cs[0]
-        self.assertEqual(c.sha, "a1b2c3d")
-        self.assertEqual(c.author, "Ada Lovelace")
-        self.assertEqual(c.email, "ada@example.com")
-        self.assertEqual(c.date, "2026-08-10T12:00:00+01:00")
-        self.assertEqual(c.subject, "feat: first commit")
-
-    def test_parses_changed_files(self):
-        cs = gitimport.parse_log(SAMPLE)
-        self.assertEqual(cs[0].files, ["README.md", "src/main.c"])
-        self.assertEqual(cs[1].files, ["docs/notes.md"])
-
-    def test_ignores_empty(self):
-        self.assertEqual(gitimport.parse_log(""), [])
-
-    def test_skip_malformed_record(self):
-        self.assertEqual(gitimport.parse_log("\x1eweird\x1e"), [])
-
-    def test_commit_leaf_shape(self):
-        leafs = gitimport.commits_to_leafs(gitimport.parse_log(SAMPLE), "sample-repo")
-        self.assertEqual(len(leafs), 2)
-        lf = leafs[0]
-        self.assertEqual(lf["type"], "commit")
-        self.assertEqual(lf["repo"], "sample-repo")
-        self.assertEqual(lf["source"], "sample-repo@a1b2c3d")
-        self.assertIn("Ada Lovelace", lf["text"])
-        self.assertIn("README.md", lf["related"])
-        self.assertIn("feat: first commit", lf["heading"])
-
-
-if __name__ == "__main__":
-    unittest.main()

bin/tools/test_published_docs.py

@@ -39,6 +39,12 @@ class PublishedDocsTest(unittest.TestCase):
             "mail index is a brain write; README must name bin/brain/index.go",
         )
 
+    def test_readme_git_import_is_gogit(self) -> None:
+        text = (ROOT / "README.md").read_text()
+        self.assertIn("bin/git/import.go", text)
+        self.assertIn("go-git", text)
+        self.assertIn("D19", (ROOT / "PLAN.md").read_text())
+
     def test_docs_do_not_claim_hop_walks(self) -> None:
         paths = [
             ROOT / "README.md",