feat(ui): add necromancer score metric (fixes #45)#280
Conversation
git-miku
Bot
added
the
needs-triage
git-miku
Bot
added
size/S
|
Tip 👋 Hey @Diwakar-odds — Miku's on it. Here are the most useful commands for this PR:
Note 🎓 This repo is part of ECSOC26. Run Note ⭐ Star this repo to unlock all commands. Say 📖 All commands🤖 AI-powered — 14 commands
🔧 Issue & PR management — 18 commands
🎉 Community & utility — 9 commands
|
PR Analysis & Label RequestThis PR introduces the Necromancer Score (Issue #45) to the TUI dashboards. It calculates deep historical patterns (resurrecting 6-month dormant projects) and seamlessly integrates these visual metrics across all three main dashboard components. ECSoC26 Label Justification:
Please review when you have a moment. Thank you! |
|
| Filename | Overview |
|---|---|
| termstory/insights.py | Adds necromancer scoring logic, with the project-level helper depending on enough session history to see dormant gaps. |
| termstory/tui.py | Displays the new metric in dashboard headers, but filtered views can miss valid revivals because they pass only the visible sessions. |
| tests/test_tui.py | Updates async TUI tests and mocks avatar fetching for the changed dashboard rendering. |
Reviews (2): Last reviewed commit: "fix(ui): resolve test failures and marku..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| days_diff = (session_dt - last_seen_dt).days | ||
|
|
||
| # Check for resurrection: > 180 days dead + meaningful session | ||
| if days_diff > 180: |
There was a problem hiding this comment.
When a project has been inactive for exactly 180 days, this strict check returns no resurrection even though the feature is described as counting 180+ day dormancy. The TUI then shows a necromancer score that is too low for that boundary case, while the related project necromancer calculation treats the same threshold inclusively.
| if days_diff > 180: | |
| if days_diff >= 180: |
bitflicker64
left a comment
bitflicker64
left a comment
There was a problem hiding this comment.
Hey @Diwakar-odds, picking up #45 and #38 back-to-back is good momentum. But this PR has the same three blockers as #282 and a few of its own, so it needs another pass before merge.
What's blocking
-
CI is red on all four Python versions (
MarkupError: auto closing tag ('[/]') has nothing to closeintest_tui_landing_page_after_onboarding). The cause is a markup bug in the new header line, not a logic bug: the f-string attermstory/tui.py:934ends with...[/dim][/]— that's one extra[/]because the wrapping[/]from the previous pattern was retained. Compare to the line it replaces (PROJECTS:row) which balances cleanly. Same bug class as the DAILY CLASS header in #282. Fix: drop the trailing[/], or drop the leading[bold cyan]from theNECROMANCER:segment if you intended the structure from #282. Three sites have the bug (lines 934, 1167, 1366). -
Duplicate logic with
calculate_project_necromancer_score. Acalculate_project_necromancer_score(sessions, projects) -> Dict[str, Any]already exists onmainattermstory/insights.py:578(line 621 on the PR head). It's wired intoformatter.pyviaformat_necromancer_scoreand used incli.py. The PR adds a near-duplicate under a slightly different name (calculate_necromancer_score, singular "project" dropped, noprojectsparameter, returnsint). Two implementations of the same metric will drift. Greptile flagged a related concern but didn't surface the duplicate itself. -
The two implementations disagree on the threshold. This new helper uses
days_diff > 180(strict, off-by-one for exactly 180 days, which is what the issue describes). The existingcalculate_project_necromancer_scoreusesgap >= 180 * 24 * 3600(inclusive). Pick one and make the codebase consistent — preferably use the existing helper since it already has tests and a formatter.
Suggested fix shape
- Delete this new
calculate_necromancer_score. Reuse the existingcalculate_project_necromancer_scorefrom the formatter path (.get("score", 0)) in the three header sites instead. - Fix the trailing
[/]in all three header lines. - If you want the "meaningful session" filter (>5 mins OR >10 commands) on top of the existing helper, add it as a post-filter in the formatter or wrap it in a thin helper that calls the existing one — don't fork the implementation.
Smaller notes
- New helper doesn't filter
is_legacysessions; the existing one does. Same drift concern. last_seen_by_project[s.project_id] = datetime.fromtimestamp(end_ts)updates even for sessions you skip (e.g. None project_id above), but that's fine here because thecontinuehappens before the update. Worth a comment for the next reader.- Wrapped header at
tui.py:1167replacesCOMMITS:withNECROMANCER:—total_commitsis still computed but no longer rendered. Same regression as #282. Add a row, don't replace. NECROMANCER:is 11 chars while most other labels in the column are 9 (PROJECTS:,COMMITS:). The extra spaces after the label compensate, but the row will look misaligned if you ever shorten the label. Minor.- No tests for the new helper. At minimum add a unit test that constructs a session with a 180-day gap and asserts the score.
Happy to approve once the duplicate is removed (or the two are reconciled), the markup is fixed, and CI is green.
| break | ||
| return streak | ||
|
|
||
| def calculate_necromancer_score(sessions: List[Session]) -> int: |
There was a problem hiding this comment.
Duplicate of calculate_project_necromancer_score at termstory/insights.py:621 on this branch. The existing one is already wired into formatter.py:format_necromancer_score and cli.py, has tests in tests/test_insights.py, and returns the dict shape the formatter expects. Two implementations of the same metric will drift — recommended path is to delete this helper and pull .get("score", 0) from the existing one in the three header sites.
| for s in sorted_sessions: | ||
| if s.project_id is None: | ||
| continue | ||
|
|
There was a problem hiding this comment.
Threshold off-by-one. Issue #45 says "6+ months"; this strict > 180 excludes exactly-180-day revivals. The existing calculate_project_necromancer_score uses gap >= 180 * 24 * 3600 (inclusive) and matches the issue text. Reconcile to >= 180, or just reuse the existing helper and drop this one entirely.
| header_lines.append(f"[bold cyan]{avatar_lines[7]}[/] [bold cyan]FOCUS SCORE:[/] [bold green]{fs:.1f}/10.0[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[8]}[/] [bold cyan]PEAK VELOCITY:[/] [dim]{peak_velocity}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]PROJECTS:[/] [dim]{active_projects_count}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score} dead projects revived[/dim][/]") |
There was a problem hiding this comment.
Root cause of the CI failure (MarkupError: auto closing tag ('[/]') has nothing to close on all 4 Python versions). The f-string ends with ...[/dim][/] — that's an extra [/] because the wrapping [/] from the previous pattern was kept. The line it replaces (PROJECTS:[/] ... [dim]{active_projects_count}[/]) balances cleanly with one [/] at the end. Fix: drop the trailing [/], leaving [dim]{necro_score} dead projects revived[/dim].
| header_lines.append(f"[bold cyan]{avatar_lines[7]}[/] [bold cyan]FOCUS SCORE:[/] [bold green]{fs:.1f}/10.0[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[8]}[/] [bold cyan]PEAK VELOCITY:[/] [dim]{peak_velocity}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]COMMITS:[/] [dim]{total_commits}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score} dead projects revived[/dim][/]") |
There was a problem hiding this comment.
Same extra [/] markup bug. Fix mirrors line 934.
| header_lines.append(f"[bold cyan]{avatar_lines[7]}[/] [bold cyan]FOCUS SCORE:[/] [bold green]{fs:.1f}/10.0[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[8]}[/] [bold cyan]PEAK TIME:[/] [dim]{peak_velocity}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]PROJECTS:[/] [dim]{len(projects)}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score} dead projects revived[/dim][/]") |
There was a problem hiding this comment.
Same extra [/] markup bug. Fix mirrors line 934.
| header_lines.append(f"[bold cyan]{avatar_lines[7]}[/] [bold cyan]FOCUS SCORE:[/] [bold green]{fs:.1f}/10.0[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[8]}[/] [bold cyan]PEAK VELOCITY:[/] [dim]{peak_velocity}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]COMMITS:[/] [dim]{total_commits}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score} dead projects revived[/dim][/]") |
There was a problem hiding this comment.
Same regression as PR #282 — wrapped view loses total_commits because the header row is replaced instead of added. total_commits is still computed above this line. Add a second header_lines.append instead of swapping.
git-miku
Bot
added
size/L
|
I've resolved the test suite failures and the UI rendering bugs (the missing closing tags that swallowed the error). The tests for the Necromancer score and the daily/monthly wrapped views now pass consistently. Let me know if further changes are needed! |
| operator = get_operator_handle() | ||
| fs = calculate_focus_score(sessions) | ||
| tod = calculate_time_of_day_distribution(sessions) | ||
| necro_score = calculate_project_necromancer_score(sessions, projects) |
There was a problem hiding this comment.
calculate_project_necromancer_score only detects revivals by comparing consecutive sessions present in that list. If a project was last touched 200 days ago and revived today, the daily view contains only today's session, so the header reports 0 dead projects revived even though the revival meets the 180+ day rule. This metric needs access to the prior activity history for the displayed projects, not just the visible session slice.
|
Thanks for the follow-up. I re-reviewed the current head ( pytest tests/test_insights.py::test_calculate_project_necromancer_score \
tests/test_tui.py::test_tui_landing_page_after_onboarding \
tests/test_tui.py::test_wrapped_view_generation_and_layout -qThat subset passes (
Suggested fix shape: remove the unused helper, compute |
bitflicker64
left a comment
bitflicker64
left a comment
There was a problem hiding this comment.
Code Review Summary
Verdict: Request changes — CI-breaking type misuse + large unrelated TUI churn; reuses the wrong return shape for an existing metric.
Intent
Surface a “necromancer” (dead-project revival) score in TUI headers for issue #45.
Critical
-
necro_scoreis a dict, interpolated into Rich markup
calculate_project_necromancer_score(...)returns{"score": int, "resurrections": [...]}. Embedding{necro_score}stringifies that dict; list brackets in the string are interpreted as Rich tags →MarkupError: auto closing tag ('[/]') has nothing to close(matches the known failure mode for this suite).Use the integer:
necro = calculate_project_necromancer_score(sessions, projects) necro_score = necro["score"]
-
Dead / duplicate helper — this PR also adds unused
calculate_necromancer_scoreininsights.pywhile main already has the richercalculate_project_necromancer_score. Delete the new int-only duplicate; call the existing API.
Warnings
- Large rewrite of
render_time_summary(try-wrap, removegenerate_daily_chronicleimport, discard expressions likelen({s.date_str...})andsum(len(s.commits)...)) is unrelated noise for a metric PR — shrink the diff. - Header replaces a row instead of adding capacity → loses PROJECTS / COMMITS in some views (same class of regression as #281/#282). Prefer adding a row or a dedicated stats line.
- Threshold/semantics differ between the unused helper (
> 180days + “meaningful session”) and the existing function (gap >= 180 days, no meaningfulness filter). Pick one definition (issue #45) and stick to the shared helper. - No focused unit tests for the score itself (only TUI churn).
Suggestions
- Wire
necro["score"]only; keep diff to insights (if needed) + 3 header call sites + small pure tests. - Don’t invent a second necromancer implementation.
Verdict
Request changes — fix dict-in-markup crash and drop the duplicate helper before re-review.
| header_lines.append(f"[bold cyan]{avatar_lines[6]}[/] [bold cyan]ACTIVE REPOS:[/] [bold]{active_projects_count} Workspaces[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[7]}[/] [bold cyan]FOCUS SCORE:[/] [bold green]{fs:.1f}/10.0[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[8]}[/] [bold cyan]PEAK VELOCITY:[/] [dim]{peak_velocity}[/]") | ||
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score} dead projects revived[/dim]") |
There was a problem hiding this comment.
Critical: necro_score is the full dict from calculate_project_necromancer_score. str(dict) can include [...] which Rich parses as markup and blows up TUI tests.
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score} dead projects revived[/dim]") | |
| header_lines.append(f"[bold cyan]{avatar_lines[9]}[/] [bold cyan]NECROMANCER:[/] [dim]{necro_score['score']} dead projects revived[/dim]") |
(Or better: assign necro_score = calculate_project_necromancer_score(... )["score"] once above and keep {necro_score}.) Same fix needed at the other two header sites.
| break | ||
| return streak | ||
|
|
||
| def calculate_necromancer_score(sessions: List[Session]) -> int: |
There was a problem hiding this comment.
Blocking: This new calculate_necromancer_score is unused and duplicates (with different thresholds/semantics) the existing calculate_project_necromancer_score later in this file. Delete it and call the existing helper from the TUI.
| proj_name = "Other" | ||
| project_seconds[proj_name] += s.duration_seconds | ||
| len({s.date_str for s in sessions if s.start_time}) | ||
|
|
There was a problem hiding this comment.
This expression result is discarded (len({...}) with no assignment). Looks like leftover from the large re-indent. Please drop dead code and keep the PR scoped to the metric.
|
/triage |
git-miku
Bot
added
triage/triaged
|
miku ✅ Marked as triaged ( |
|
:miku /check-star |
git-miku
Bot
added
the
do-not-merge/needs-star
|
miku @Diwakar-odds still hasn't starred this repo. Applied Star this repo first, then run |
2 participants
Summary
Introduces the Necromancer Score metric for user behavior analysis.
Motivation
Closes #45.
Changes
termstory/insights.py:calculate_necromancer_scoreto tally meaningful sessions resurrecting 180+ day dormant projects.termstory/tui.py:Acceptance Criteria
Impact & Side Effects
No breaking changes. Adds visual metrics to the TUI.
How to Test
termstory guiand open any dashboard timeframe view to verify the headers.NECROMANCERmetric correctly.Quality Checklist