gitmole 0.6.4__tar.gz → 0.6.5__tar.gz

{gitmole-0.6.4 → gitmole-0.6.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gitmole
-Version: 0.6.4
+Version: 0.6.5
 Summary: Offline git repository analysis with a terminal report: hotspots, coupling, ownership, code age, secrets, repo health.
 License: MIT
 Project-URL: Homepage, https://github.com/antvinni/gitmole

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole/__init__.py

@@ -1,3 +1,3 @@
 """gitmole: offline git repository analysis with a terminal report."""
 
-__version__ = "0.6.4"
+__version__ = "0.6.5"

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole/coupling.py

@@ -32,12 +32,13 @@ def _components(pairs: list) -> list:
     return list(groups.values())
 
 
-def clusters(pairs: list, min_files: int = 4) -> tuple:
+def clusters(pairs: list, min_files: int = 4, min_density: float = 0.8) -> tuple:
     """Split `pairs` into (groups, rest). Pairs whose two files share a directory are gathered per
-    directory and then into connected groups; a group touching at least `min_files` distinct files
-    becomes one cluster with the file and pair counts, the weakest degree and the mean of the pairs'
-    average revisions. Every other pair comes back unchanged, in its original order. Two unrelated
-    pairs in one directory are not a cluster. Clusters are largest first."""
+    directory and then into connected groups; a group of at least `min_files` distinct files with at
+    least `min_density` of the possible pairs present becomes one cluster with the file and pair
+    counts, the weakest degree and the mean of the pairs' average revisions. Every other pair comes
+    back unchanged, in its original order. Two unrelated pairs in one directory, or a chain of pairs,
+    are not a cluster: "each other" has to be true. Clusters are largest first."""
     by_dir = defaultdict(list)
     for p in pairs:
         if _dir(p["entity"]) == _dir(p["coupled"]):
@@ -46,7 +47,8 @@ def clusters(pairs: list, min_files: int = 4) -> tuple:
     for directory, ps in by_dir.items():
         for component in _components(ps):
             files = {p["entity"] for p in component} | {p["coupled"] for p in component}
-            if len(files) < min_files:
+            possible = len(files) * (len(files) - 1) / 2
+            if len(files) < min_files or len(component) < min_density * possible:
                 continue
             groups.append({"dir": directory, "files": len(files), "pairs": len(component), "degree": min(p["degree"] for p in component),
                            "average-revs": round(sum(p["average-revs"] for p in component) / len(component))})

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole/findings.py

@@ -100,6 +100,14 @@ def _source_ownership(report: dict) -> list:
     return [r for r in report.get("ownership") or [] if not (filetypes.is_test_path(r["entity"]) or filetypes.is_vendor_path(r["entity"]))]
 
 
+def _present_areas(report: dict, rows: list, build=knowledge.areas) -> list:
+    """Areas built from the ownership rows of directories that still exist, then only those areas that
+    exist themselves: a directory the history knows but HEAD does not (the layout before a move to
+    src/ or crates/) is nowhere to pair anyone on. `build` is knowledge.areas or a wrapper of it."""
+    tree = _tree(report)
+    return [a for a in build(knowledge.present_rows(rows, tree)) if knowledge.in_tree(a["area"], tree)]
+
+
 def bus_factor(report: dict, threshold: float = 0.7, min_lines: int = 200) -> list:
     """One author owns most of the surviving code (whole history). The areas named in the advice
     come from lines added, which `--since` windows, so the advice says so when it applies."""
@@ -111,7 +119,7 @@ def bus_factor(report: dict, threshold: float = 0.7, min_lines: int = 200) -> li
     if lines / total <= threshold:
         return []
     theirs = []
-    for a in knowledge.areas(_source_ownership(report)):
+    for a in _present_areas(report, _source_ownership(report)):
         owned = dict(a["owners"]).get(name, 0)
         if a["lines"] >= min_lines and owned / a["lines"] >= 0.8:
             theirs.append((a["area"], round(100 * owned / a["lines"])))
@@ -263,7 +271,9 @@ def reverts(report: dict, min_share: float = 0.05, min_count: int = 5, warn_shar
 
 
 def knowledge_islands(report: dict, min_lines: int = 200, min_share: float = 0.9) -> list:
-    areas = knowledge.areas(_source_ownership(report))
+    """Areas of the tree written almost entirely by one person. Areas that no longer exist are left
+    out, of the islands and of the total they are measured against."""
+    areas = _present_areas(report, _source_ownership(report))
     islands = knowledge.islands(areas, min_lines=min_lines, min_share=min_share)
     if not islands:
         return []
@@ -329,9 +339,10 @@ def _loss_people(by_person: dict, total: int) -> str:
 
 
 def _loss_areas(report: dict, names: set, source_rows: list) -> list:
-    """Areas at 200+ lines where 80%+ of the surviving code is theirs, tagged live or not and
-    sorted live-first: that is where the gap bites soonest."""
-    theirs = [a for a in loss.areas(source_rows, names) if a["lines"] >= 200 and a["lost_share"] >= 0.8]
+    """Areas at 200+ lines where 80%+ of the surviving code is theirs, still in the tree, tagged live
+    or not and sorted live-first: that is where the gap bites soonest."""
+    theirs = [a for a in _present_areas(report, source_rows, build=lambda rows: loss.areas(rows, names))
+              if a["lines"] >= 200 and a["lost_share"] >= 0.8]
     for a in theirs:
         a["live"] = _is_live(a["area"], report.get("age") or [])
     theirs.sort(key=lambda a: (not a["live"], -a["lines"], a["area"]))   # a live area first: that is where the gap bites

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole/knowledge.py

@@ -7,6 +7,16 @@ from collections import Counter, defaultdict
 ROOT = "(root files)"
 
 
+def in_tree(area: str, tree: dict) -> bool:
+    """Whether any tracked file sits under `area` (a directory prefix ending in "/", or ROOT). With
+    no tree listing every area counts: there is nothing to judge by."""
+    if not tree:
+        return True
+    if area == ROOT:
+        return any("/" not in path for path in tree)
+    return any(path.startswith(area) for path in tree)
+
+
 def _area(entity: str, depth: int) -> str:
     dirs = entity.split("/")[:-1]
     if not dirs:
@@ -14,6 +24,20 @@ def _area(entity: str, depth: int) -> str:
     return "/".join(dirs[:depth]) + "/"
 
 
+def top_area(entity: str) -> str:
+    """The top-level directory of a path, or ROOT."""
+    return _area(entity, 1)
+
+
+def present_rows(rows: list, tree: dict) -> list:
+    """Ownership rows for files whose top-level directory still exists in the tree. Filtering the rows
+    before areas are built keeps a vanished layout (the src/ before a move to crates/) from inflating
+    the total and hiding that one directory now holds almost everything."""
+    if not tree:
+        return rows
+    return [r for r in rows if in_tree(top_area(r["entity"]), tree)]
+
+
 def _aggregate(rows: list, depth: int) -> list:
     lines, per_author = Counter(), defaultdict(Counter)
     for r in rows:

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole/leaks.py

@@ -33,10 +33,13 @@ RAW_FIELDS = ("Secret", "Match", "Line", "Message", "Attributes")
 
 # Shapes that cannot be a live secret: a version string (5.0.0-1667386184.dfbbb54), a token shortened
 # with an ellipsis, a whole-value template marker (your-project-id, <your-token>, XXXX-XXXX, changeme),
+# a whole value that is one of the words every example uses (`password: 'hello'` in a doc comment),
 # and a key block whose body holds no key material. Every rule is about the whole value; nothing is
 # skipped by prefix, since a public and a private key of the same service often share one.
 _VERSION = re.compile(r"^\d+\.\d+\.\d+(?:[-+][0-9A-Za-z.-]+)?$")
 _MARKER = re.compile(r"^(<[^<>]+>|x+|(?:x{2,}[-_ ]?)+|your[-_][\w-]+|change[-_]?me|replace[-_]?me)$", re.I)
+_EXAMPLE_WORDS = {"password", "passwd", "pass", "secret", "hello", "hey", "test", "example", "sample", "dummy", "foo", "bar",
+                  "baz", "admin", "root", "user", "123456", "12345678", "123456789", "abc123", "qwerty", "letmein", "welcome"}
 _KEY_BLOCK = re.compile(r"-----BEGIN [A-Z ]*KEY-----(.*?)-----END [A-Z ]*KEY-----", re.S)
 _KEY_MATERIAL = 64   # a real body is hundreds of base64 characters; a template has dots or a few x's
 
@@ -54,7 +57,7 @@ def digest(value: str, key: bytes) -> str:
 
 def is_placeholder(value: str) -> bool:
     value = (value or "").strip()
-    if _VERSION.match(value) or value.endswith("...") or value.endswith("…") or _MARKER.match(value):
+    if _VERSION.match(value) or value.endswith("...") or value.endswith("…") or _MARKER.match(value) or value.lower() in _EXAMPLE_WORDS:
         return True
     m = _KEY_BLOCK.search(value)
     if m:

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole/render.py

@@ -502,7 +502,16 @@ def knowledge_section(report: dict, full: bool = True, width=None) -> dict:
     """Ownership by area of the tree: who wrote most of each directory, gone owners marked."""
     months = report["meta"].get("gone_months", loss.DEFAULT_MONTHS)
     gone = {g["name"] for g in loss.gone(report, months)}
-    areas = loss.areas(report.get("ownership") or [], gone)   # every area the map showed before, tests included
+    rows_all = report.get("ownership") or []   # every area the map showed before, tests included
+    areas = loss.areas(rows_all, gone)
+    hidden_note = None
+    tree = (report.get("size") or {}).get("files") or {}
+    if full is not True and tree:
+        # a directory the history knows but HEAD does not is a layout that no longer exists; the rows are
+        # filtered before the areas are built so a vanished layout cannot hide that one directory now dominates
+        areas = [a for a in loss.areas(knowledge.present_rows(rows_all, tree), gone) if knowledge.in_tree(a["area"], tree)]
+        hidden = len({knowledge.top_area(r["entity"]) for r in rows_all if not knowledge.in_tree(knowledge.top_area(r["entity"]), tree)})
+        hidden_note = f"{hidden} historical area{'s' if hidden != 1 else ''} hidden{HIDDEN_SUFFIX}" if hidden else None
     limit = _limit("Knowledge map", full)
     rows = []
     for a in areas[:limit]:
@@ -512,7 +521,7 @@ def knowledge_section(report: dict, full: bool = True, width=None) -> dict:
     columns = [("area", PATH), ("lines added", RIGHT), ("authors", RIGHT), ("lost", RIGHT), ("main owner", {}), ("second", {})]
     if full is not True:
         columns, rows = _keep(columns, rows, ["area", "lines added", "main owner", "second"])
-    notes = [c for c in (_more(len(areas), limit),) if c]
+    notes = [c for c in (_more(len(areas), limit), hidden_note) if c]
     if gone:
         notes.append(f"gone = no commits in the {months} months before {report['meta'].get('last_date')}"
                      + ("; gone and lost are measured over the whole history" if report["meta"].get("since") else ""))

{gitmole-0.6.4 → gitmole-0.6.5}/gitmole.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gitmole
-Version: 0.6.4
+Version: 0.6.5
 Summary: Offline git repository analysis with a terminal report: hotspots, coupling, ownership, code age, secrets, repo health.
 License: MIT
 Project-URL: Homepage, https://github.com/antvinni/gitmole

{gitmole-0.6.4 → gitmole-0.6.5}/tests/test_coupling.py

@@ -45,8 +45,24 @@ class Clusters(unittest.TestCase):
         self.assertEqual([g["dir"] for g in groups], ["lib/", "(root files)"])
         self.assertEqual(rest, [])
 
-    def test_a_partial_clique_counts_the_files_it_touches(self):
-        pairs = [pair("d/a", "d/b"), pair("d/b", "d/c"), pair("d/c", "d/e")]   # 4 files, 3 pairs, not every pair present
-        [g], rest = coupling.clusters(pairs)
-        self.assertEqual((g["files"], g["pairs"]), (4, 3))
+    def test_a_chain_is_not_a_cluster(self):
+        pairs = [pair("d/a", "d/b"), pair("d/b", "d/c"), pair("d/c", "d/e")]   # 4 files, 3 of 6 possible pairs: connected, not a clique
+        groups, rest = coupling.clusters(pairs)
+        self.assertEqual(groups, [])
+        self.assertEqual(rest, pairs)
+
+    def test_a_near_complete_clique_is_a_cluster(self):
+        pairs = clique("d", ["a", "b", "c", "e", "f"])   # 5 files, 10 pairs
+        eight = pairs[:8]
+        [g], rest = coupling.clusters(eight)
+        self.assertEqual((g["files"], g["pairs"]), (5, 8))
         self.assertEqual(rest, [])
+        groups, rest = coupling.clusters(pairs[:7])   # 7 of 10 is under the 80% floor
+        self.assertEqual(groups, [])
+        self.assertEqual(len(rest), 7)
+
+    def test_two_unrelated_pairs_in_one_directory_are_not_a_cluster(self):
+        pairs = [pair("a", "b"), pair("c", "d")]
+        groups, rest = coupling.clusters(pairs)
+        self.assertEqual(groups, [])
+        self.assertEqual(rest, pairs)

{gitmole-0.6.4 → gitmole-0.6.5}/tests/test_findings.py

@@ -145,6 +145,14 @@ class BusFactor(unittest.TestCase):
         self.assertEqual(findings.bus_factor(r)[0]["advice"], "Pair someone with Ann on core/ first; it is 95% theirs since 2025-01-01.",
                          "ownership is windowed while the headline share is not")
 
+    def test_areas_no_longer_in_the_tree_are_not_named_in_the_advice(self):
+        own = [{"entity": "flask/a.py", "author": "Ann", "added": 5000, "deleted": 0},   # the pre-src/ layout
+               {"entity": "src/a.py", "author": "Ann", "added": 900, "deleted": 0},
+               {"entity": "src/b.py", "author": "Bob", "added": 50, "deleted": 0}]
+        tree = {"files": {"src/a.py": {"code": 1, "complexity": 0}, "src/b.py": {"code": 1, "complexity": 0}}}
+        f = findings.bus_factor(report(theseus_authors={"Ann": 79, "Bob": 21}, ownership=own, size=tree))
+        self.assertEqual(f[0]["advice"], "Pair someone with Ann on src/ first; it is 95% theirs.")
+
     def test_vendored_trees_are_not_named_in_the_advice(self):
         own = [{"entity": "vendor/github.com/x/a.go", "author": "Ann", "added": 500000, "deleted": 0},
                {"entity": "core/a.py", "author": "Ann", "added": 900, "deleted": 0},
@@ -438,6 +446,19 @@ class KnowledgeIslands(unittest.TestCase):
         self.assertEqual(f[0]["advice"], "Pair someone with Bob on core/ first; it is the largest at 300 lines.")
         self.assertNotIn("tests/", f[0]["detail"])
 
+    def test_areas_no_longer_in_the_tree_are_not_islands(self):
+        own = [{"entity": "src/a.rs", "author": "Ann", "added": 30000, "deleted": 0},   # moved to crates/ years ago
+               {"entity": "grep-printer/a.rs", "author": "Ann", "added": 20000, "deleted": 0},
+               {"entity": "crates/core/a.rs", "author": "Bob", "added": 300, "deleted": 0}]
+        tree = {"files": {"crates/core/a.rs": {"code": 300, "complexity": 1}}}
+        f = findings.knowledge_islands(report(ownership=own, size=tree))
+        self.assertEqual(f[0]["advice"], "Pair someone with Bob on crates/core/ first; it is the largest at 300 lines.",
+                         "with the vanished directories gone, crates/ holds everything and the map descends into it")
+        self.assertNotIn("src/", f[0]["detail"])
+        self.assertIn("100% of all lines added", f[0]["detail"], "lines in vanished directories are not in the denominator")
+        f = findings.knowledge_islands(report(ownership=own))
+        self.assertIn("src/", f[0]["detail"], "without a tree listing every area counts")
+
     def test_vendored_trees_are_not_islands(self):
         own = [{"entity": "vendor/github.com/x/a.go", "author": "Ann", "added": 500000, "deleted": 0},
                {"entity": "web/node_modules/y/b.js", "author": "Ann", "added": 90000, "deleted": 0},
@@ -511,6 +532,18 @@ class KnowledgeLoss(unittest.TestCase):
         self.assertIn("Areas mostly theirs: old/ (100%), docs/ (100%)", f[0]["detail"])
         self.assertEqual(f[0]["advice"], "Pair someone on old/ first; nobody who wrote it is around to ask.")
 
+    def test_areas_no_longer_in_the_tree_are_not_named(self):
+        r = self._report(theseus_authors={"Ann": 60, "Bob": 40},
+                         age=[{"entity": "flask/a.py", "age-months": 2}, {"entity": "src/x.py", "age-months": 2}],
+                         ownership=[{"entity": "flask/a.py", "author": "Bob", "added": 8000, "deleted": 0},   # the old layout, all Bob's
+                                    {"entity": "src/x.py", "author": "Bob", "added": 300, "deleted": 0},
+                                    {"entity": "app/b.py", "author": "Ann", "added": 900, "deleted": 0}],
+                         size={"files": {"src/x.py": {"code": 1, "complexity": 0}, "app/b.py": {"code": 1, "complexity": 0}}})
+        f = findings.knowledge_loss(r)
+        self.assertIn("Areas mostly theirs: src/ (100%).", f[0]["detail"])
+        self.assertNotIn("flask/", f[0]["detail"])
+        self.assertEqual(f[0]["advice"], "Pair someone on src/ first; nobody who wrote it is around to ask.")
+
     def test_a_live_area_is_preferred_over_a_bigger_idle_one(self):
         r = self._report(theseus_authors={"Ann": 60, "Bob": 40},
                          age=[{"entity": "old/a.py", "age-months": 30}, {"entity": "live/b.py", "age-months": 3}],

{gitmole-0.6.4 → gitmole-0.6.5}/tests/test_knowledge.py

@@ -48,6 +48,36 @@ class OddNames(unittest.TestCase):
         self.assertEqual(areas[0]["lines"], 1000)
 
 
+class InTree(unittest.TestCase):
+    TREE = {"crates/core/a.rs": {}, "crates/ignore/src/b.rs": {}, "build.rs": {}}
+
+    def test_an_area_is_in_the_tree_when_any_tracked_file_sits_under_it(self):
+        self.assertTrue(knowledge.in_tree("crates/", self.TREE))
+        self.assertTrue(knowledge.in_tree("crates/ignore/", self.TREE))
+        self.assertFalse(knowledge.in_tree("src/", self.TREE))
+        self.assertFalse(knowledge.in_tree("crate/", self.TREE), "a prefix of a directory name is not that directory")
+
+    def test_root_files_are_in_the_tree_when_any_file_has_no_directory(self):
+        self.assertTrue(knowledge.in_tree(knowledge.ROOT, self.TREE))
+        self.assertFalse(knowledge.in_tree(knowledge.ROOT, {"crates/core/a.rs": {}}))
+
+    def test_no_tree_listing_means_every_area_counts(self):
+        self.assertTrue(knowledge.in_tree("src/", {}))
+
+    def test_present_rows_drop_vanished_top_level_directories_so_the_survivor_can_dominate(self):
+        rows = [{"entity": "src/a.rs", "author": "Ann", "added": 30000, "deleted": 0},          # the layout before crates/
+                {"entity": "crates/core/a.rs", "author": "Ann", "added": 9000, "deleted": 0},
+                {"entity": "crates/ignore/b.rs", "author": "Bob", "added": 900, "deleted": 0},
+                {"entity": "ci/x.sh", "author": "Ann", "added": 500, "deleted": 0}]
+        tree = {"crates/core/a.rs": {}, "crates/ignore/b.rs": {}, "ci/x.sh": {}}
+        kept = knowledge.present_rows(rows, tree)
+        self.assertEqual([r["entity"] for r in kept], ["crates/core/a.rs", "crates/ignore/b.rs", "ci/x.sh"])
+        self.assertEqual([a["area"] for a in knowledge.areas(kept)], ["crates/core/", "crates/ignore/", "ci/"],
+                         "with src/ gone, crates/ holds over 80% and the map descends into it")
+        self.assertEqual([a["area"] for a in knowledge.areas(rows)], ["src/", "crates/", "ci/"], "the vanished src/ used to hide that")
+        self.assertEqual(knowledge.present_rows(rows, {}), rows)
+
+
 class Islands(unittest.TestCase):
     def test_areas_dominated_by_one_author(self):
         areas = knowledge.areas(OWNERSHIP)

{gitmole-0.6.4 → gitmole-0.6.5}/tests/test_leaks.py

@@ -59,6 +59,13 @@ class Placeholder(unittest.TestCase):
         for value in ["AKIA" + "X" * 16, "6L" + "x" * 38, "ghp_" + "a1" * 18, "yourkey" + "9" * 20]:
             self.assertFalse(leaks.is_placeholder(value), "a marker inside real-looking material is not enough: " + value)
 
+    def test_common_example_words_are_placeholders(self):
+        # `password: 'hello'` in a doc comment, `secret` in a sample config: the words every example uses
+        for value in ["hello", "Hello", "secret", "password", "PASSWORD", "example", "123456", "qwerty", "letmein", "foo", "dummy"]:
+            self.assertTrue(leaks.is_placeholder(value), value)
+        for value in ["hello123", "secret-9f8a7b6c5d4e", "s3cr3t!Passw0rd", "foobarbaz2024"]:
+            self.assertFalse(leaks.is_placeholder(value), "a word inside other material is not a placeholder: " + value)
+
     def test_anything_else_is_taken_seriously(self):
         # built at runtime: a literal in these shapes would trip secret scanners on this very file
         key_id, long_key = "AKIA" + "X" * 16, "6L" + "x" * 38

{gitmole-0.6.4 → gitmole-0.6.5}/tests/test_render.py

@@ -212,6 +212,7 @@ class Report(unittest.TestCase):
     def test_knowledge_map_caption_says_gone_is_measured_over_the_whole_history(self):
         r = sample_report()
         r["meta"].update({"last_date": "2026-09-10", "bots": []})
+        r["size"]["files"]["tests/t.py"] = {"code": 1, "complexity": 0}   # every area in the tree, so nothing is hidden
         r["activity"]["authors"] = {"Ann": {"commits": 1, "added": 0, "deleted": 0, "first": "2025-01-01", "last": "2026-09-01"},
                                     "Bob": {"commits": 1, "added": 0, "deleted": 0, "first": "2025-01-01", "last": "2025-01-01"}}
         def caption(rep):
@@ -636,6 +637,25 @@ class KnowledgeMap(unittest.TestCase):
         r["ownership"] = []
         self.assertIn("no ownership data", rendered(r, []))
 
+    def test_default_map_hides_areas_no_longer_in_the_tree_and_says_so(self):
+        r = sample_report()   # the tree holds static/ files only; tests/t.py in the ownership rows is history
+        r["ownership"].append({"entity": "flask/app.py", "author": "Ann", "added": 4000, "deleted": 0})
+        km = _section_text(rendered(r, [], width=200), "Knowledge map")
+        self.assertIn("static/", km)
+        self.assertNotIn("flask/", km)
+        self.assertNotIn("tests/", km)
+        self.assertIn("2 historical areas hidden; --full shows them", km)
+        full = _section_text(rendered(r, [], width=200, full=True), "Knowledge map")
+        self.assertIn("flask/", full)
+        self.assertNotIn("hidden", full)
+
+    def test_map_without_a_tree_listing_hides_nothing(self):
+        r = sample_report()
+        r["size"]["files"] = {}
+        km = _section_text(rendered(r, [], width=200), "Knowledge map")
+        self.assertIn("tests/", km)
+        self.assertNotIn("historical", km)
+
 
 class Timeline(unittest.TestCase):
     def test_last_twelve_months_per_author_with_dots_for_zero(self):