project-guide 2.7.1__tar.gz → 2.7.2__tar.gz

{project_guide-2.7.1 → project_guide-2.7.2}/.gitignore

@@ -32,6 +32,8 @@ build/
 /site/
 
 # project-guide
-docs/project-guide/**
-!docs/project-guide/go.md
-docs/project-guide/**/*.bak.*
+/docs/project-guide/.metadata.yml
+/docs/project-guide/README.md
+/docs/project-guide/developer/
+/docs/project-guide/templates/
+/docs/project-guide/**/*.bak.*

{project_guide-2.7.1 → project_guide-2.7.2}/.project-guide.yml

@@ -1,8 +1,8 @@
 version: '2.0'
-installed_version: 2.6.0
+installed_version: 2.7.2
 target_dir: docs/project-guide
 metadata_file: .metadata.yml
-current_mode: default
+current_mode: debug
 test_first: false
 pyve_version: pyve version 2.6.2
 project_name: project-guide

{project_guide-2.7.1 → project_guide-2.7.2}/CHANGELOG.md

@@ -7,6 +7,23 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.7.2] - 2026-05-19
+
+Bug fix: `project-guide git-push` (and version/phase detection) silently dropped sub-numbered story IDs (`J.m.1`, `J.m.2`, …). When the latest `[Done]` story used the sub-numbered form, the wrapper fell back to the previous bare-letter heading and reported it as "already committed," blocking the push of the new story.
+
+### Fixed
+- **`project_guide/stories.py:_STORY_RE`** — extended the story-ID character class from `[A-Z]\.[a-z]+` to `[A-Z]\.[a-z]+(?:\.\d+)?` so headings of the form `### Story J.m.1: ...` parse correctly. Without this, `_read_done_stories()` silently filtered such headings out and the wrapper picked the wrong "last [Done] story."
+- **`project_guide/cli.py:_COMMIT_SUBJECT_STORY_ID_RE`** — extended in the same shape so the already-committed check recognizes commit subjects like `J.m.1: ...`. Without this fix the duplicate-detection path would have a second-layer hole even after the stories.md side was fixed.
+- **`project_guide/actions.py:_VERSION_RE`** — extended in the same shape so `detect_latest_version()` no longer under-reports the highest version when the latest story uses the sub-numbered form.
+
+### Documented
+- **`_phase-letters.md`** — added a "Sub-numbered stories" subsection covering both use cases: pre-implementation split (`J.m` was planned but split into `J.m.1`, `J.m.2` before any work; bare `J.m` heading dropped) and post-implementation follow-up (`J.m` shipped, then a bug or follow-on feature lands as `J.m.1` before proceeding to `J.n`). Flat single-level only — no cascading like `J.m.1.1`.
+
+### Tests
+- New `tests/test_stories.py` covering `_STORY_RE`, `_read_done_stories()`, and `derive_commit_message()` — `stories.py` had no direct unit tests prior to this story (the test-coverage gap that let the bug ship in P.k).
+- New `tests/test_cli.py` cases for both `git-push` scenarios (post-impl follow-up with bare `J.m` present; pre-impl split with no bare `J.m`) and the round-trip "sub-numbered story is already committed" path.
+- New `tests/test_actions.py` case asserting `detect_latest_version()` picks up the version on a sub-numbered heading.
+
 ## [2.7.1] - 2026-05-11
 
 Compatibility fix for IDE-integrated LLM @-mention / fuzzy-search. The v2.6.0–v2.7.0 gitignore block used a clean `<target>/**` + `!<target>/go.md` negation pair — correct per the `.gitignore` spec, but **several IDE tools** (Cursor, parts of the VS Code fork ecosystem, certain LSP-based search backends) implement a subset of gitignore semantics that does not honor re-include negation. Those tools applied the broad `**` rule, hid `go.md`, and defeated the IDE-LLM-visibility constraint that's the entire reason `go.md` is tracked. P.l switches to a negation-free explicit-list form so simplistic parsers handle it reliably.

{project_guide-2.7.1 → project_guide-2.7.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: project-guide
-Version: 2.7.1
+Version: 2.7.2
 Summary: Stay organized and in control with adaptive LLM workflow prompts.
 Project-URL: Homepage, https://github.com/pointmatic/project-guide
 Project-URL: Documentation, https://pointmatic.github.io/project-guide/

{project_guide-2.7.1 → project_guide-2.7.2}/docs/specs/stories.md

@@ -308,6 +308,40 @@ The new canonical form lists every gitignored top-level entry explicitly so no n
 - An opt-back-in flag (`--gitignore-style=negation`). YAGNI until someone asks; the new form is strictly better for the documented IDE-LLM-visibility constraint.
 - A `project-guide check` integrity rule that detects "consumer has tracked-but-should-be-ignored files under `target_dir`". Defer until there's a second integrity rule worth shipping (see Future > Integrity & Validation).
 
+### Story P.m: v2.7.2 Recognize sub-numbered story IDs (`J.m.1`) in regex sites [Done]
+
+Reported bug: in a consumer project with the heading sequence `… J.l [Done], J.m.1 [Done]`, `project-guide git-push` printed `"Story J.l is already committed. Use 'git-push' directly for any follow-up commit."` and refused to push, even though `J.m.1` was the actual just-completed story.
+
+**Root cause:** three regexes encoded the story-ID shape as `[A-Z]\.[a-z]+`, which silently fails to match the sub-numbered form. `stories.py:_STORY_RE` was the proximate cause — `_read_done_stories()` filtered the `J.m.1` heading out entirely, leaving `J.l` as the "last `[Done]`," and the commit-subject check then correctly observed that `J.l` had been committed. The other two sites (`cli.py:_COMMIT_SUBJECT_STORY_ID_RE`, `actions.py:_VERSION_RE`) had the same hole and were fixed in the same story to avoid a half-fix where a future code path re-introduced the bug from a different angle.
+
+**Sub-numbered form** — `<phase>.<letter>.<digit>+`, flat single-level only (no cascading like `J.m.1.1`). Two use cases per the developer's intent:
+
+- **Pre-implementation split:** `J.m` is planned but its scope is judged too large before any work begins; the heading is split into `J.m.1`, `J.m.2` and the bare `J.m` heading is dropped. Sequence: `… J.l, J.m.1, J.m.2, J.n, …`.
+- **Post-implementation follow-up:** `J.m` ships, then a bug or follow-on feature must land before proceeding to `J.n`; the follow-up is added as `J.m.1` (and may cascade to `J.m.2`, `J.m.3`, …). Sequence: `… J.l, J.m, J.m.1, J.m.2, …, J.n, …`.
+
+Both scenarios are exercised by the new test suite.
+
+- [x] `project_guide/stories.py:_STORY_RE` — extend capture group to `[A-Z]\.[a-z]+(?:\.\d+)?` with comment cross-referencing `_phase-letters.md`
+- [x] `project_guide/cli.py:_COMMIT_SUBJECT_STORY_ID_RE` — matching extension; updated comment to reflect the new shape so future readers see why the two regexes must move together
+- [x] `project_guide/actions.py:_VERSION_RE` — matching extension; `detect_latest_version()` now picks up versions on sub-numbered headings
+- [x] `project_guide/templates/project-guide/templates/modes/_phase-letters.md` — new "Sub-numbered stories" subsection documenting both scenarios and the flat-only constraint
+- [x] Ran `project-guide update` to propagate the template change into `docs/project-guide/templates/modes/_phase-letters.md` (installed copy)
+- [x] New `tests/test_stories.py` — first direct unit-test coverage for `stories.py` (the gap that let the bug ship in P.k). Covers `_STORY_RE` matches (plain, sub-numbered, multi-digit sub-number), `_read_done_stories()` for both scenarios, `derive_commit_message()` with sub-numbered ID, and the round-trip through `cli._COMMIT_SUBJECT_STORY_ID_RE`
+- [x] `tests/test_cli.py` — new cases for both `git-push` scenarios (post-impl follow-up; pre-impl split with no bare `J.m`) and the "sub-numbered story is already committed" path that surfaces the second-layer regex hole
+- [x] `tests/test_actions.py` — new case asserting `detect_latest_version()` picks up the version on a sub-numbered heading
+- [x] Bump `project_guide/version.py`: `__version__ = "2.7.2"`
+- [x] Bump `pyproject.toml`: `version = "2.7.2"`
+- [x] Add `## [2.7.2] - 2026-05-19` entry to `CHANGELOG.md` framed as a bug fix with test-coverage backfill
+
+**Migration:** none. The change is purely permissive — existing plain-letter IDs (`J.l`, `J.m`) continue to match. Consumers who never used sub-numbered IDs see no behavior change; consumers who did will see `git-push` and `detect_latest_version()` start handling those headings correctly.
+
+**Why the docs change rides this story:** the sub-numbered form was already in use by consumers but undocumented in `_phase-letters.md`. Extending the regex without documenting the form would leave the next maintainer staring at an unexplained `(?:\.\d+)?` tail. Doc + code travel together.
+
+**Out of scope:**
+- Multi-level cascading (`J.m.1.1`). The developer's intent is flat; YAGNI until a consumer asks. Adding it later is a strict superset (`(?:\.\d+)+` instead of `(?:\.\d+)?`) and would be backwards compatible.
+- A `project-guide check` rule that warns when a sub-numbered ID appears without a preceding plain-letter heading in non-pre-split contexts. The two valid scenarios cover everything seen in the wild; defer until there's evidence of misuse.
+- Sub-numbered phase letters (`AA.1`). Phases don't carry the same pre-impl-split / post-impl-followup workflow that motivates the story-level sub-numbering, and no consumer has asked.
+
 ---
 
 ## Future

{project_guide-2.7.1 → project_guide-2.7.2}/project_guide/actions.py

@@ -127,9 +127,11 @@ class Artifact:
             action=action,
         )
 
-# Matches `### Story <Phase>.<sub>: vMAJOR.MINOR.PATCH ...`
+# Matches `### Story <Phase>.<sub>: vMAJOR.MINOR.PATCH ...` — the optional
+# `(?:\.\d+)?` tail covers sub-numbered IDs (`J.m.1`, `J.m.2`, …) per the
+# `_phase-letters.md` "Sub-numbered stories" rule.
 _VERSION_RE = re.compile(
-    r"^###\s+Story\s+[A-Za-z]+\.[a-z]+:\s+v(\d+)\.(\d+)\.(\d+)",
+    r"^###\s+Story\s+[A-Za-z]+\.[a-z]+(?:\.\d+)?:\s+v(\d+)\.(\d+)\.(\d+)",
     re.M,
 )
 

{project_guide-2.7.1 → project_guide-2.7.2}/project_guide/cli.py

@@ -1626,8 +1626,9 @@ def heal(no_input: bool):
 
 # Regex matching the story-ID prefix of a commit-subject line — used to detect
 # which stories have already been committed (Story P.k). Mirrors the
-# `[A-Z]\.[a-z]+` shape of `_STORY_RE` in `stories.py`.
-_COMMIT_SUBJECT_STORY_ID_RE = __import__("re").compile(r"^([A-Z]\.[a-z]+):\s")
+# `[A-Z]\.[a-z]+(?:\.\d+)?` shape of `_STORY_RE` in `stories.py` (the optional
+# `.\d+` tail covers sub-numbered IDs like `J.m.1`).
+_COMMIT_SUBJECT_STORY_ID_RE = __import__("re").compile(r"^([A-Z]\.[a-z]+(?:\.\d+)?):\s")
 
 
 def _get_committed_story_ids() -> set[str]:

{project_guide-2.7.1 → project_guide-2.7.2}/project_guide/stories.py

@@ -17,8 +17,11 @@ from dataclasses import dataclass, field
 from pathlib import Path
 
 # Matches: ### Story N.a: v2.4.0 Some Title [Done]
+# The optional `(?:\.\d+)?` tail captures sub-numbered IDs (`J.m.1`, `J.m.2`, …)
+# used for pre-implementation splits or post-implementation follow-ups. See
+# `_phase-letters.md` (Sub-numbered stories).
 _STORY_RE = re.compile(
-    r"^### Story ([A-Z]\.[a-z]+): (.+) \[(Done|In Progress|Planned)\]\s*$",
+    r"^### Story ([A-Z]\.[a-z]+(?:\.\d+)?): (.+) \[(Done|In Progress|Planned)\]\s*$",
     re.MULTILINE,
 )
 

{project_guide-2.7.1 → project_guide-2.7.2}/project_guide/templates/project-guide/templates/modes/_phase-letters.md

@@ -14,6 +14,15 @@ Within a phase, stories use lowercase letters following the same scheme: `A.a`,
 
 Examples: `A.a`, `A.b`, …, `A.z`, `A.aa`, `A.ab`, ….
 
+### Sub-numbered stories
+
+A story may carry an optional numeric suffix — `J.m.1`, `J.m.2`, … — appended after the sub-letter. Sub-numbers are flat (no cascading like `J.m.1.1`) and start at `1`. Two situations use them:
+
+- **Pre-implementation split.** When `J.m` is planned but its scope is judged too large before any work begins, the heading is split into `J.m.1`, `J.m.2`, … and the bare `J.m` heading is dropped. Sequence becomes `…, J.l, J.m.1, J.m.2, J.n, …`.
+- **Post-implementation follow-up.** When `J.m` ships but a bug or follow-on feature must land before proceeding to `J.n`, the follow-up is added as `J.m.1` (and may cascade to `J.m.2`, `J.m.3`, …). Sequence becomes `…, J.l, J.m, J.m.1, J.m.2, …, J.n, …`.
+
+Sub-numbered stories follow normal Version Cadence: each one that ships code takes its own bump.
+
 ### Continuing across archive boundaries
 
 When `stories.md` is archived (via `archive_stories` mode), the fresh `stories.md` starts empty — but phase letters do **not** reset. To determine the next phase letter:

{project_guide-2.7.1 → project_guide-2.7.2}/project_guide/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "2.7.1"
+__version__ = "2.7.2"

{project_guide-2.7.1 → project_guide-2.7.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "project-guide"
-version = "2.7.1"
+version = "2.7.2"
 description = "Stay organized and in control with adaptive LLM workflow prompts."
 readme = "README.md"
 license = "Apache-2.0"

{project_guide-2.7.1 → project_guide-2.7.2}/tests/test_actions.py

@@ -150,6 +150,19 @@ def test_detect_latest_version_raises_when_none_found():
         detect_latest_version(text)
 
 
+def test_detect_latest_version_recognizes_subnumbered_story_id():
+    """Sub-numbered story IDs (J.m.1, used for pre-impl splits or post-impl
+    follow-ups) must contribute to the latest-version computation. The
+    plain-letter regex silently dropped them, causing `detect_latest_version`
+    to under-report the highest version when the latest story used the
+    `.NN` form."""
+    text = (
+        "### Story J.l: v0.68.0 prior story [Done]\n\n"
+        "### Story J.m.1: v0.70.0 follow-up after J.m [Done]\n"
+    )
+    assert detect_latest_version(text) == (0, 70, 0)
+
+
 # ---------------------------------------------------------------------------
 # detect_latest_phase_letter
 # ---------------------------------------------------------------------------

{project_guide-2.7.1 → project_guide-2.7.2}/tests/test_cli.py

@@ -2811,4 +2811,87 @@ def test_git_push_propagates_child_exit_code(runner, tmp_path, monkeypatch):
         assert result.exit_code == 7, result.output
 
 
+def test_git_push_picks_up_subnumbered_last_done_story_scenario_a(runner, tmp_path, monkeypatch):
+    """Scenario A — post-implementation follow-up: ..., J.l, J.m, J.m.1.
+
+    Reported bug: when the last [Done] story is `J.m.1`, the wrapper
+    silently fell back to `J.l` and reported it as already committed.
+    """
+    with runner.isolated_filesystem(temp_dir=tmp_path):
+        _write_stories_md(
+            "### Story J.l: update remaining refs [Done]",
+            "### Story J.m: v0.69.0 Integrate component [Done]",
+            "### Story J.m.1: v0.70.0 Follow-up after J.m [Done]",
+        )
+        _mock_git_push_on_path(monkeypatch)
+        captured: list = []
+        _mock_git_log_subjects(
+            monkeypatch,
+            subjects=[
+                "J.l: update remaining refs",
+                "J.m: v0.69.0 Integrate component",
+            ],
+            git_push_argv_capture=captured,
+        )
+
+        result = runner.invoke(main, ['git-push'])
+
+        assert result.exit_code == 0, result.output
+        assert len(captured) == 1, captured
+        assert captured[0][1] == "J.m.1: v0.70.0 Follow-up after J.m"
+
+
+def test_git_push_picks_up_subnumbered_last_done_story_scenario_b(runner, tmp_path, monkeypatch):
+    """Scenario B — pre-implementation split: ..., J.l, J.m.1, J.m.2 (no bare J.m).
+
+    When J.m's scope is split before implementation, the bare J.m heading
+    never exists. The wrapper's "last [Done]" must be J.m.2.
+    """
+    with runner.isolated_filesystem(temp_dir=tmp_path):
+        _write_stories_md(
+            "### Story J.l: update remaining refs [Done]",
+            "### Story J.m.1: v0.69.0 First half of the split [Done]",
+            "### Story J.m.2: v0.70.0 Second half of the split [Done]",
+        )
+        _mock_git_push_on_path(monkeypatch)
+        captured: list = []
+        _mock_git_log_subjects(
+            monkeypatch,
+            subjects=[
+                "J.l: update remaining refs",
+                "J.m.1: v0.69.0 First half of the split",
+            ],
+            git_push_argv_capture=captured,
+        )
+
+        result = runner.invoke(main, ['git-push'])
+
+        assert result.exit_code == 0, result.output
+        assert len(captured) == 1, captured
+        assert captured[0][1] == "J.m.2: v0.70.0 Second half of the split"
+
+
+def test_git_push_errors_when_subnumbered_story_is_already_committed(runner, tmp_path, monkeypatch):
+    """Already-committed detection must recognize sub-numbered IDs in
+    commit subjects (e.g. `J.m.1: ...`), not just `J.m: ...`. Without
+    this, a sub-numbered story would be re-pushed silently."""
+    with runner.isolated_filesystem(temp_dir=tmp_path):
+        _write_stories_md(
+            "### Story J.m.1: v0.70.0 Follow-up after J.m [Done]",
+        )
+        _mock_git_push_on_path(monkeypatch)
+        captured: list = []
+        _mock_git_log_subjects(
+            monkeypatch,
+            subjects=["J.m.1: v0.70.0 Follow-up after J.m"],
+            git_push_argv_capture=captured,
+        )
+
+        result = runner.invoke(main, ['git-push'])
+
+        assert result.exit_code == 1
+        assert "Story J.m.1 is already committed" in result.output
+        assert captured == [], "git-push must not be invoked when the story is already committed"
+
+
 # --- End Story P.k ----------------------------------------------------------

project_guide-2.7.2/tests/test_stories.py

@@ -0,0 +1,141 @@
+# Copyright (c) 2026 Pointmatic
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from pathlib import Path
+
+from project_guide.stories import (
+    _STORY_RE,
+    StoryHeading,
+    _read_done_stories,
+    derive_commit_message,
+)
+
+_HEADER = """\
+# stories.md -- testproject (python)
+
+## Phase J: Demo
+
+"""
+
+
+def _write(tmp_path: Path, body: str) -> Path:
+    specs = tmp_path / "docs" / "specs"
+    specs.mkdir(parents=True)
+    (specs / "stories.md").write_text(_HEADER + body, encoding="utf-8")
+    return specs
+
+
+# ---------------------------------------------------------------------------
+# _STORY_RE — direct regex coverage for sub-numbered story IDs
+# ---------------------------------------------------------------------------
+
+
+def test_story_re_matches_plain_letter_id():
+    """Baseline: ordinary `J.l` form still matches."""
+    line = "### Story J.l: update remaining quizazz refs [Done]\n"
+    matches = _STORY_RE.findall(line)
+    assert matches == [("J.l", "update remaining quizazz refs", "Done")]
+
+
+def test_story_re_matches_sub_numbered_id():
+    """Sub-numbered `J.m.1` form (post-impl follow-up or pre-impl split)
+    must be parsed the same way as the plain form."""
+    line = "### Story J.m.1: v0.70.0 Follow-up after J.m [Done]\n"
+    matches = _STORY_RE.findall(line)
+    assert matches == [("J.m.1", "v0.70.0 Follow-up after J.m", "Done")]
+
+
+def test_story_re_matches_multi_digit_sub_number():
+    """Sub-numbers can exceed a single digit (e.g., `.10`, `.11`)."""
+    line = "### Story J.m.10: tenth follow-up [Done]\n"
+    matches = _STORY_RE.findall(line)
+    assert matches == [("J.m.10", "tenth follow-up", "Done")]
+
+
+# ---------------------------------------------------------------------------
+# _read_done_stories — "last [Done]" selection with sub-numbered IDs
+# ---------------------------------------------------------------------------
+
+
+def test_read_done_stories_scenario_a_subnumber_follows_bare_letter(tmp_path):
+    """Scenario A — post-implementation follow-up.
+
+    Sequence: J.l, J.m, J.m.1 (all Done). J.m was implemented, then a
+    follow-up bug/feature surfaced as J.m.1 before proceeding to J.n.
+    The wrapper's "last [Done]" must be J.m.1, not J.m or J.l.
+    """
+    specs = _write(
+        tmp_path,
+        (
+            "### Story J.l: update remaining refs [Done]\n\n"
+            "### Story J.m: v0.69.0 Integrate component [Done]\n\n"
+            "### Story J.m.1: v0.70.0 Follow-up after J.m [Done]\n"
+        ),
+    )
+    headings = _read_done_stories(str(specs))
+    assert headings is not None
+    assert [h.story_id for h in headings] == ["J.l", "J.m", "J.m.1"]
+    assert headings[-1] == StoryHeading(
+        story_id="J.m.1", title="v0.70.0 Follow-up after J.m"
+    )
+
+
+def test_read_done_stories_scenario_b_subnumber_without_bare_letter(tmp_path):
+    """Scenario B — pre-implementation split.
+
+    Sequence: J.l, J.m.1, J.m.2 (all Done). J.m's scope was deemed too
+    large before any implementation began, so it was split into J.m.1
+    and J.m.2 and the bare J.m heading never existed. The wrapper's
+    "last [Done]" must be J.m.2.
+    """
+    specs = _write(
+        tmp_path,
+        (
+            "### Story J.l: update remaining refs [Done]\n\n"
+            "### Story J.m.1: v0.69.0 First half of the split [Done]\n\n"
+            "### Story J.m.2: v0.70.0 Second half of the split [Done]\n"
+        ),
+    )
+    headings = _read_done_stories(str(specs))
+    assert headings is not None
+    assert [h.story_id for h in headings] == ["J.l", "J.m.1", "J.m.2"]
+    assert headings[-1].story_id == "J.m.2"
+
+
+# ---------------------------------------------------------------------------
+# derive_commit_message — round-trip with the cli's commit-subject regex
+# ---------------------------------------------------------------------------
+
+
+def test_derive_commit_message_with_sub_numbered_id():
+    """Output preserves the full sub-numbered ID, including the colon."""
+    heading = StoryHeading(story_id="J.m.1", title="v0.70.0 Follow-up after J.m")
+    assert derive_commit_message(heading) == "J.m.1: v0.70.0 Follow-up after J.m"
+
+
+def test_derive_commit_message_round_trip_with_cli_subject_regex():
+    """The commit-subject already-committed check (cli._COMMIT_SUBJECT_STORY_ID_RE)
+    must recognize the sub-numbered ID it would produce from our own
+    `derive_commit_message`. Without this, a J.m.1 commit subject would
+    not register in the "committed" set, breaking the wrapper's
+    duplicate-detection path even after the stories.md side is fixed.
+    """
+    from project_guide.cli import _COMMIT_SUBJECT_STORY_ID_RE
+
+    subject = derive_commit_message(
+        StoryHeading(story_id="J.m.1", title="v0.70.0 Follow-up after J.m")
+    )
+    m = _COMMIT_SUBJECT_STORY_ID_RE.match(subject)
+    assert m is not None, f"regex failed to match subject: {subject!r}"
+    assert m.group(1) == "J.m.1"