Skip to content

feat(auto-review): add proof metrics to compact ledger #381

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
shiny-code-bot merged 1 commit into from
Jun 5, 2026
Merged

feat(auto-review): add proof metrics to compact ledger #381

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
186 changes: 186 additions & 0 deletions code-rs/core/src/review_store.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1187,6 +1187,12 @@ struct LedgerDiagnostics {
recent_runs: usize,
in_flight_runs: usize,
terminal_runs: usize,
skipped_runs: usize,
duplicate_skipped_runs: usize,
superseded_runs: usize,
failed_runs: usize,
cancelled_runs: usize,
lost_runs: usize,
token_count: u64,
token_runs: usize,
prompt_token_estimate: u64,
Expand Down Expand Up @@ -1219,6 +1225,18 @@ where
if suppresses_stale_findings {
diagnostics.suppressed_stale_runs += 1;
}
let has_terminal_proof = run.error_summary.is_some()
|| run.error_class.is_some()
|| run.cancel_reason.is_some();
let is_skipped = run.status == AutoReviewRunStatus::Skipped;
let is_duplicate_skipped = is_skipped
&& run.cancel_reason.as_deref() == Some("duplicate_auto_review_scope");
let is_superseded = run.status == AutoReviewRunStatus::Superseded;
let is_failed = run.status == AutoReviewRunStatus::Failed;
let is_cancelled = run.status == AutoReviewRunStatus::Cancelled;
let is_lost = run.status == AutoReviewRunStatus::Lost;
let is_proof_run = is_skipped || is_superseded;
let is_terminal_proof_run = has_terminal_proof && (is_failed || is_cancelled || is_lost);
let has_saved_signal = run.saved_token_estimate.is_some();
let has_cost_signal = run.token_count.is_some()
|| run.prompt_token_estimate.is_some()
Expand All @@ -1230,11 +1248,31 @@ where
&& !has_timing_signal
&& !high_burn
&& !suppresses_stale_findings
&& !is_proof_run
&& !is_terminal_proof_run
{
continue;
}

diagnostics.recent_runs += 1;
if is_skipped {
diagnostics.skipped_runs += 1;
}
if is_duplicate_skipped {
diagnostics.duplicate_skipped_runs += 1;
}
if is_superseded {
diagnostics.superseded_runs += 1;
}
if is_failed {
diagnostics.failed_runs += 1;
}
if is_cancelled {
diagnostics.cancelled_runs += 1;
}
if is_lost {
diagnostics.lost_runs += 1;
}
if run.status.is_in_flight() {
diagnostics.in_flight_runs += 1;
} else {
Expand Down Expand Up @@ -1305,6 +1343,27 @@ where
diagnostics.saved_token_estimate, diagnostics.saved_runs
));
}
if diagnostics.skipped_runs > 0 {
line.push_str(&format!(" skipped={}", diagnostics.skipped_runs));
}
if diagnostics.duplicate_skipped_runs > 0 {
line.push_str(&format!(
" duplicate_skipped={}",
diagnostics.duplicate_skipped_runs
));
}
if diagnostics.superseded_runs > 0 {
line.push_str(&format!(" superseded={}", diagnostics.superseded_runs));
}
if diagnostics.failed_runs > 0 {
line.push_str(&format!(" failed={}", diagnostics.failed_runs));
}
if diagnostics.cancelled_runs > 0 {
line.push_str(&format!(" cancelled={}", diagnostics.cancelled_runs));
}
if diagnostics.lost_runs > 0 {
line.push_str(&format!(" lost={}", diagnostics.lost_runs));
}
if let Some(longest_elapsed_bucket) = diagnostics.longest_elapsed_bucket {
line.push_str(&format!(" longest_elapsed={longest_elapsed_bucket}"));
}
Expand Down Expand Up @@ -2465,6 +2524,34 @@ mod tests {

assert!(ledger.contains("saved_estimate=84000t"));
assert!(ledger.contains("saved_runs=1"));
assert!(ledger.contains("skipped=1"));
assert!(ledger.contains("duplicate_skipped=1"));
assert!(!ledger.contains(&format!("run id={run_id}")));
}

#[test]
#[serial]
fn compact_ledger_counts_skipped_duplicates_without_saved_tokens() {
let code_home = TempDir::new().unwrap();
let repo = TempDir::new().unwrap();
set_code_home(code_home.path());
let run_id = Uuid::new_v4();
let duplicate_id = Uuid::new_v4();
let mut store = AutoReviewRunStore::open(repo.path()).unwrap();
store
.upsert(AutoReviewRun::new(run_id, AutoReviewRunSource::Tui, 10))
.unwrap();
store
.mark_skipped_duplicate(run_id, duplicate_id, Some("diff:abc"), None, 20)
.unwrap();

let ledger = store.compact_ledger(ledger_options(30)).expect("ledger");

assert!(ledger.contains("diagnostics recent_runs=1"));
assert!(ledger.contains("terminal=1"));
assert!(ledger.contains("skipped=1"));
assert!(ledger.contains("duplicate_skipped=1"));
assert!(!ledger.contains("saved_estimate="));
assert!(!ledger.contains(&format!("run id={run_id}")));
}

Expand All @@ -2491,11 +2578,110 @@ mod tests {
let ledger = store.compact_ledger(ledger_options(60)).expect("ledger");

assert!(ledger.contains("prompt_estimate=42000t"));
assert!(ledger.contains("superseded=1"));
assert!(!ledger.contains("saved_estimate="));
assert!(!ledger.contains("saved_runs="));
assert!(!ledger.contains(&format!("run id={clean_id}")));
assert_eq!(store.get(clean_id).unwrap().saved_token_estimate, None);
}

#[test]
#[serial]
fn compact_ledger_omits_old_skipped_and_superseded_proof_runs() {
let code_home = TempDir::new().unwrap();
let repo = TempDir::new().unwrap();
set_code_home(code_home.path());
let skipped_id = Uuid::new_v4();
let superseded_id = Uuid::new_v4();
let mut skipped = AutoReviewRun::new(skipped_id, AutoReviewRunSource::Tui, 1);
skipped.cancel_reason = Some("duplicate_auto_review_scope".to_string());
skipped.mark_status(AutoReviewRunStatus::Skipped, 2);
let mut superseded = AutoReviewRun::new(superseded_id, AutoReviewRunSource::Tui, 3);
superseded.mark_status(AutoReviewRunStatus::Superseded, 4);
let mut store = AutoReviewRunStore::open(repo.path()).unwrap();
store.upsert(skipped).unwrap();
store.upsert(superseded).unwrap();

assert!(
store
.compact_ledger(ledger_options(LEDGER_DIAGNOSTIC_RECENT_SECS + 20))
.is_none()
);
}

#[test]
#[serial]
fn compact_ledger_combines_active_run_with_dedup_proof() {
let code_home = TempDir::new().unwrap();
let repo = TempDir::new().unwrap();
set_code_home(code_home.path());
let skipped_id = Uuid::new_v4();
let active_id = Uuid::new_v4();
let duplicate_id = Uuid::new_v4();
let mut active = AutoReviewRun::new(active_id, AutoReviewRunSource::Tui, 30);
active.agent_id = Some("agent-active".to_string());
active.freshness = AutoReviewFreshness::Current;
active.mark_status(AutoReviewRunStatus::Reviewing, 40);
active.mark_activity(45);
let mut store = AutoReviewRunStore::open(repo.path()).unwrap();
store.upsert(active).unwrap();
store
.upsert(AutoReviewRun::new(skipped_id, AutoReviewRunSource::Tui, 10))
.unwrap();
store
.mark_skipped_duplicate(skipped_id, duplicate_id, Some("diff:abc"), Some(84_000), 20)
.unwrap();

let ledger = store.compact_ledger(ledger_options(60)).expect("ledger");

assert!(ledger.contains("diagnostics recent_runs=1"));
assert!(ledger.contains("skipped=1"));
assert!(ledger.contains("duplicate_skipped=1"));
assert!(ledger.contains("saved_estimate=84000t"));
assert!(ledger.contains(&format!("run id={active_id}")));
assert!(!ledger.contains(&format!("run id={skipped_id}")));
}

#[test]
#[serial]
fn compact_ledger_counts_terminal_outcomes_without_listing_clean_runs() {
let code_home = TempDir::new().unwrap();
let repo = TempDir::new().unwrap();
set_code_home(code_home.path());
let mut store = AutoReviewRunStore::open(repo.path()).unwrap();

let failed_id = Uuid::new_v4();
let cancelled_id = Uuid::new_v4();
let lost_id = Uuid::new_v4();

let mut failed = AutoReviewRun::new(failed_id, AutoReviewRunSource::Tui, 10);
failed.error_class = Some("context_window".to_string());
failed.error_summary = Some("context window exceeded".to_string());
failed.mark_status(AutoReviewRunStatus::Failed, 20);
store.upsert(failed).unwrap();

let mut cancelled = AutoReviewRun::new(cancelled_id, AutoReviewRunSource::Tui, 30);
cancelled.cancel_reason = Some("agent_cancelled".to_string());
cancelled.mark_status(AutoReviewRunStatus::Cancelled, 40);
store.upsert(cancelled).unwrap();

let mut lost = AutoReviewRun::new(lost_id, AutoReviewRunSource::Tui, 50);
lost.cancel_reason = Some("agent_missing_after_restart".to_string());
lost.mark_status(AutoReviewRunStatus::Lost, 60);
store.upsert(lost).unwrap();

let ledger = store.compact_ledger(ledger_options(70)).expect("ledger");

assert!(ledger.contains("diagnostics recent_runs=3"));
assert!(ledger.contains("terminal=3"));
assert!(ledger.contains("failed=1"));
assert!(ledger.contains("cancelled=1"));
assert!(ledger.contains("lost=1"));
assert!(ledger.contains(&format!("run id={failed_id}")));
assert!(!ledger.contains(&format!("run id={cancelled_id}")));
assert!(!ledger.contains(&format!("run id={lost_id}")));
}

#[test]
#[serial]
fn compact_ledger_includes_recent_burn_diagnostics() {
Expand Down