From 7c91fe1e4b49d30c88f5a3d433ed0dd9cee8545d Mon Sep 17 00:00:00 2001
From: chopratejas <chopratejas@gmail.com>
Date: Tue, 14 Apr 2026 15:59:49 -0700
Subject: [PATCH] Fix `headroom learn` failing on project paths with
 underscores (#159)

_component_tokenizations only split on `-` and `.`, so directory names
like `my_project` could never be reconstructed from the dash-encoded
slug. Add `_` as a separator so the greedy decoder matches snake_case
directory names correctly.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 headroom/learn/plugins/claude.py |  8 ++--
 tests/test_learn/test_scanner.py | 64 ++++++++++++++++++++++++++++++++
 2 files changed, 68 insertions(+), 4 deletions(-)

diff --git a/headroom/learn/plugins/claude.py b/headroom/learn/plugins/claude.py
index df9e7caac..049b57788 100644
--- a/headroom/learn/plugins/claude.py
+++ b/headroom/learn/plugins/claude.py
@@ -375,9 +375,9 @@ def _component_tokenizations(component: str) -> list[list[str]]:
 
     add([component])
 
-    for separator in ("-", ".", None):
+    for separator in ("-", ".", "_", None):
         if separator is None:
-            tokens = [token for token in re.split(r"[-.]", component) if token]
+            tokens = [token for token in re.split(r"[-._]", component) if token]
         else:
             tokens = [token for token in component.split(separator) if token]
         add(tokens)
@@ -385,9 +385,9 @@ def _component_tokenizations(component: str) -> list[list[str]]:
     if component.startswith(".") and len(component) > 1:
         hidden_component = component[1:]
         add(["", hidden_component])
-        for separator in ("-", ".", None):
+        for separator in ("-", ".", "_", None):
             if separator is None:
-                tokens = [token for token in re.split(r"[-.]", hidden_component) if token]
+                tokens = [token for token in re.split(r"[-._]", hidden_component) if token]
             else:
                 tokens = [token for token in hidden_component.split(separator) if token]
             add(["", *tokens])
diff --git a/tests/test_learn/test_scanner.py b/tests/test_learn/test_scanner.py
index 0739c9a96..48aaa003d 100644
--- a/tests/test_learn/test_scanner.py
+++ b/tests/test_learn/test_scanner.py
@@ -103,6 +103,38 @@ class TestGreedyPathDecode:
         result = _greedy_path_decode(tmp_path, ["my", "cool", "project", "nosync", "headroom"])
         assert result == tmp_path / "my-cool-project.nosync" / "headroom"
 
+    # ---- Underscore tests (issue #159) ----
+
+    def test_single_underscore_in_dirname(self, tmp_path: Path) -> None:
+        """Directory name contains one literal underscore (e.g. my_project)."""
+        _make_dirs(tmp_path, "my_project")
+        result = _greedy_path_decode(tmp_path, ["my", "project"])
+        assert result == tmp_path / "my_project"
+
+    def test_multiple_underscores_in_dirname(self, tmp_path: Path) -> None:
+        """Directory name contains multiple underscores (e.g. my_cool_project)."""
+        _make_dirs(tmp_path, "my_cool_project")
+        result = _greedy_path_decode(tmp_path, ["my", "cool", "project"])
+        assert result == tmp_path / "my_cool_project"
+
+    def test_underscore_nested_path(self, tmp_path: Path) -> None:
+        """Nested path like org/my_project should decode correctly."""
+        _make_dirs(tmp_path, "org/my_project")
+        result = _greedy_path_decode(tmp_path, ["org", "my", "project"])
+        assert result == tmp_path / "org" / "my_project"
+
+    def test_mixed_underscore_and_hyphen_in_dirname(self, tmp_path: Path) -> None:
+        """Directory with both hyphens and underscores (e.g. my-cool_project)."""
+        _make_dirs(tmp_path, "my-cool_project")
+        result = _greedy_path_decode(tmp_path, ["my", "cool", "project"])
+        assert result == tmp_path / "my-cool_project"
+
+    def test_underscore_dir_containing_hyphen_subdir(self, tmp_path: Path) -> None:
+        """Path like my_app/sub-module — underscore parent + hyphen child."""
+        _make_dirs(tmp_path, "my_app/sub-module")
+        result = _greedy_path_decode(tmp_path, ["my", "app", "sub", "module"])
+        assert result == tmp_path / "my_app" / "sub-module"
+
     def test_nonexistent_path_returns_none(self, tmp_path: Path) -> None:
         result = _greedy_path_decode(tmp_path, ["does", "not", "exist"])
         assert result is None
@@ -236,6 +268,38 @@ class TestDecodeProjectPath:
         else:
             assert result is None or result == project
 
+    def test_underscore_dirname_via_greedy(self, users_tmp: Path) -> None:
+        """my_project — underscore in directory name (issue #159).
+
+        Claude Code encodes /Users/foo/org/my_project as
+        -Users-foo-org-my-project.  Simple replace gives
+        …/org/my/project which does not exist, so the greedy decoder
+        must reconstruct my_project from tokens ['my', 'project'].
+        """
+        project = users_tmp / "org" / "my_project"
+        project.mkdir(parents=True)
+
+        encoded = "-" + str(project)[1:].replace("/", "-")
+        result = _decode_project_path(encoded)
+
+        if str(users_tmp).startswith("/Users/"):
+            assert result == project
+        else:
+            assert result is None or result == project
+
+    def test_multi_underscore_dirname_via_greedy(self, users_tmp: Path) -> None:
+        """my_cool_project — multiple underscores (issue #159)."""
+        project = users_tmp / "my_cool_project"
+        project.mkdir(parents=True)
+
+        encoded = "-" + str(project)[1:].replace("/", "-")
+        result = _decode_project_path(encoded)
+
+        if str(users_tmp).startswith("/Users/"):
+            assert result == project
+        else:
+            assert result is None or result == project
+
     def test_windows_drive_letter_pattern(self) -> None:
         """Encoded name -C-MQ2-macros should detect Windows drive letter."""
         import sys