Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 | 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2348x 2348x 2348x 2348x 2348x 2348x 2347x 2348x 2348x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 31852x 31852x 3x 3x 31852x 5x 5x 31852x 13x 13x 31831x 31831x 2x 2x 2x 2x 2x 2x 3x 3x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2370x 2370x 2370x 2370x 2370x 4x 4x 2366x 2366x 2370x 11x 11x 2355x 2355x 2355x 2355x 2355x 2355x 2355x 2348x 2347x 2370x 6x 6x 2349x 2349x 2349x 2349x 2349x 168x 165x 126x 59x 2370x 6x 6x 2343x 2343x 2343x 2x 2x 2x 2x 2x 2x 2x 4x 2x 2x 4x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 15x 15x 10x 10x 7x 6x 15x 15x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 121x 121x 121x 121x 121x 1016x 4x 4x 1016x 121x 121x | /**
* Effective response-time resolution and Tier-1 timing classification.
*
* This is the single place every timing consumer resolves "how long did this
* attempt actually take?" (`getEffectiveResponseTimeMs`) and "is this sample
* usable?" (`classifyAttemptTiming`). Keeping the logic here guarantees the
* session planner, the Start-Practice modals, and the review tool all agree.
*
* Client-safe: imports ONLY types from the schema — no db/server runtime.
*/
import type { SlotResult } from '@/db/schema/session-plans'
import { MAX_RESPONSE_TIME_CAP_MS } from './constants'
/**
* The subset of a {@link SlotResult} needed to reason about its timing.
*
* A full `SlotResult` is structurally assignable, so real callers just pass
* the whole record; the narrow shape keeps callers (and tests) from having to
* build an entire result object.
*/
export type TimingSample = Pick<
SlotResult,
| 'responseTimeMs'
| 'source'
| 'wasIdleCapped'
| 'responseTimeMsRaw'
| 'capReason'
| 'capSource'
| 'capThresholdMs'
| 'timingReview'
>
export type AttemptTimingTier = 'ok' | 'tier1' | 'tier2'
export type AttemptTimingReason = 'idle-capped' | 'legacy-implausible' | 'unusual-for-child'
export interface AttemptTimingClassification {
tier: AttemptTimingTier
reason?: AttemptTimingReason
}
/**
* A child's own windowed timing distribution, used to detect Tier-2
* (genuine-but-unusual, per-child) samples. Produced by
* `computeChildTimingStats` in `pace-estimation.ts` (#157) and passed into
* {@link classifyAttemptTiming}. `spread` is precomputed there so the
* classifier stays arithmetic-free and both consumers agree on the formula.
*/
export interface ChildTimingStats {
/** Median of the child's non-Tier-1 effective response times (ms). */
median: number
/** Median absolute deviation of those samples (ms). */
mad: number
/**
* Robust spread for the Tier-2 cutoff:
* `max(1.4826·MAD, 0.25·median, 5_000ms)` (guards a degenerate MAD of 0).
*/
spread: number
/** How many non-Tier-1 samples the stats were computed from. */
cleanSampleCount: number
}
/**
* Tier-2 thresholds (per the shared contract). A non-Tier-1 sample is Tier-2
* when `x > median + TIER2_SPREAD_MULTIPLIER·spread`, AND
* `x >= TIER2_ABSOLUTE_FLOOR_MS`, AND the child has at least
* `TIER2_MIN_CLEAN_SAMPLES` clean samples. Only the slow tail is flagged.
*/
export const TIER2_SPREAD_MULTIPLIER = 3.5
export const TIER2_ABSOLUTE_FLOOR_MS = 60_000
export const TIER2_MIN_CLEAN_SAMPLES = 10
/**
* Whether the #156 guard fired on this attempt (i.e. any capture flag is set).
* Used to keep the legacy heuristic from double-handling guarded records.
*/
function hasTimingGuardFlags(result: TimingSample): boolean {
return (
result.wasIdleCapped != null ||
result.responseTimeMsRaw != null ||
result.capReason != null ||
result.capSource != null ||
result.capThresholdMs != null
)
}
/**
* Resolve the response time an attempt should contribute to timing estimates,
* or `null` if it contributes no timing sample.
*
* Resolution order (per the shared contract):
* 1. `timingReview.omitFromTiming` → null (adult removed the sample)
* 2. `timingReview.adjustedResponseTimeMs != null` → that value
* (adult-vouched replacement; never re-classified)
* 3. `source === 'recency-refresh'` or `responseTimeMs <= 0` → null
* (sentinel record / no real measurement)
* 4. else `responseTimeMs`
*/
export function getEffectiveResponseTimeMs(result: TimingSample): number | null {
const review = result.timingReview
if (review?.omitFromTiming) {
return null
}
if (review?.adjustedResponseTimeMs != null) {
return review.adjustedResponseTimeMs
}
if (result.source === 'recency-refresh' || result.responseTimeMs <= 0) {
return null
}
return result.responseTimeMs
}
/**
* Whether the #156 guard auto-quarantined this attempt as idle-capped, i.e.
* `responseTimeMs` is NOT the raw measurement.
*/
export function isIdleCapped(result: TimingSample): boolean {
return result.wasIdleCapped === true
}
/**
* Classify an attempt's timing sample.
*
* Tier-1 (provably broken → auto-quarantined):
* - `wasIdleCapped === true` (any stored value) → `tier1` / `idle-capped`
* - legacy rule: no #156 guard flags and effective time
* `> MAX_RESPONSE_TIME_CAP_MS` → `tier1` / `legacy-implausible`
*
* Tier-2 (genuine-but-unusual, per-child): only when `childStats` is supplied
* (#157). Among non-Tier-1 samples, a value is Tier-2 when it exceeds
* `median + TIER2_SPREAD_MULTIPLIER·spread` AND `TIER2_ABSOLUTE_FLOOR_MS`, and
* the child has at least `TIER2_MIN_CLEAN_SAMPLES` clean samples. Tier-2 samples
* are kept in the estimate (winsorized), never auto-excluded.
*
* An adult-adjusted value (`timingReview.adjustedResponseTimeMs`) re-enters as
* `ok` and is never re-classified.
*/
export function classifyAttemptTiming(
result: TimingSample,
childStats?: ChildTimingStats
): AttemptTimingClassification {
// Adult-adjusted values re-enter the estimate and are never re-classified.
if (result.timingReview?.adjustedResponseTimeMs != null) {
return { tier: 'ok' }
}
// Tier-1a: explicitly idle-capped by the #156 guard.
if (result.wasIdleCapped === true) {
return { tier: 'tier1', reason: 'idle-capped' }
}
const effective = getEffectiveResponseTimeMs(result)
// Tier-1b: legacy pre-#156 records with no guard flags whose effective time
// provably outran the maximum possible auto-pause clamp.
if (
effective != null &&
!hasTimingGuardFlags(result) &&
effective > MAX_RESPONSE_TIME_CAP_MS
) {
return { tier: 'tier1', reason: 'legacy-implausible' }
}
// Tier-2: unusually slow for THIS child, judged against their own windowed
// distribution. Requires enough clean samples for the stats to be meaningful.
if (
childStats != null &&
effective != null &&
childStats.cleanSampleCount >= TIER2_MIN_CLEAN_SAMPLES &&
effective >= TIER2_ABSOLUTE_FLOOR_MS &&
effective > childStats.median + TIER2_SPREAD_MULTIPLIER * childStats.spread
) {
return { tier: 'tier2', reason: 'unusual-for-child' }
}
return { tier: 'ok' }
}
/**
* Count how many results are Tier-1 (auto-quarantined) samples.
*/
export function countQuarantined(results: readonly TimingSample[]): number {
let count = 0
for (const result of results) {
if (classifyAttemptTiming(result).tier === 'tier1') {
count++
}
}
return count
}
/**
* Whether an adult has *resolved the timing flag* on this attempt — i.e. taken
* an action that settles how the sample should count. A bare review stamp (e.g.
* a mastery-only exclusion, which stamps `reviewedBy`/`reviewedAt` but touches
* no timing field) does NOT resolve the timing flag.
*
* Resolved iff a `timingReview` exists AND any of:
* - `omitFromTiming === true` (sample dropped from timing),
* - `adjustedResponseTimeMs != null` (adult-vouched replacement value),
* - `timingConfirmed === true` (adult vouched the raw value is genuine).
*
* Note: neither `omitFromTiming` nor `timingConfirmed` changes the classifier's
* tier, so a raw `classifyAttemptTiming(...).tier` count is NOT resolution-aware
* on its own — pair it with this helper via {@link countUnresolvedFlagged}.
*/
export function isFlagResolved(result: TimingSample): boolean {
const review = result.timingReview
if (review == null) return false
return (
review.omitFromTiming === true ||
review.adjustedResponseTimeMs != null ||
review.timingConfirmed === true
)
}
/**
* Count attempts still *awaiting adult review*: flagged by the classifier
* (`tier !== 'ok'`, i.e. Tier-1 or Tier-2) AND not yet resolved via
* {@link isFlagResolved}. This is the resolution-aware "needs review" count the
* summary banner and the session-list badge surface, so an omitted / adjusted /
* confirmed attempt drops out of the count and the badge disappears at zero.
*
* `childStats` is optional and only affects Tier-2 detection; per-session,
* context-free callers omit it (Tier-1 needs no stats), matching how
* `countQuarantined` classifies.
*/
export function countUnresolvedFlagged(
results: readonly TimingSample[],
childStats?: ChildTimingStats
): number {
let count = 0
for (const result of results) {
if (classifyAttemptTiming(result, childStats).tier !== 'ok' && !isFlagResolved(result)) {
count++
}
}
return count
}
|