bvandeusenandClaude Opus 5
799dab029a
The payoff slice. Until now a candidate's only claim on a slot was "some
artist you play is adjacent to it in a similarity graph" — a fact that says
nothing about whether the music sounds like anything you like. Now the
candidate's own folksonomy tags (cached by slice 5) are compared against
the user's taste-profile tags, so the deck ranks on taste and can say WHY.
The blend is MULTIPLICATIVE — score × (1 + weight × overlap) — and that
choice carries the whole safety argument:
- An untagged candidate has overlap 0, so its score is EXACTLY unchanged.
Tag coverage is permanently partial (#2376); it must cost a candidate
nothing, not sink it (rule #131).
- Nothing can leapfrog on tags alone. An additive term with a large
weight would let a near-zero-similarity artist outrank a strong match
for sharing one popular tag, which reads as noise.
- Weight 0 restores pure similarity order bit-for-bit, so the operator's
knob has a real off position.
overlap = Σ(shared) candWeight × normalizedTasteWeight ÷ Σ(all) candWeight.
Normalizing the taste side by the user's strongest tag makes the score
comparable across users (taste weights accumulate with listening, so a
heavy listener's raw numbers dwarf a new user's while meaning the same
thing). Dividing by the candidate's own mass makes it comparable across
candidates, so a densely-tagged artist can't win on tag count alone.
Applied to the whole over-fetched pool BEFORE selectSuggestions, so the
rotation and diversity rules operate on blended scores — boosting only the
twelve already chosen by similarity would leave the re-ranking undone.
A query failure is returned, NOT degraded past. Graceful degradation is
for expected absence (no taste profile, no cached tags) and both are
handled explicitly as empty inputs; swallowing a real error would hide a
broken DB behind a subtly worse ranking that nothing reports.
Migration 0051 adds a FOURTH tuning scope rather than columns on
taste_tuning, because snooze_days lives here too and a snooze must never
be read as taste signal (#2374) — filing it under 'taste' would put it one
careless join from the leak that design forbids. Expanding
recommendation_tuning_audit's CHECK is in the same migration per rule #36,
and a test asserts the audit row lands, which is what would catch its
absence.
snooze_days moves out of a Go constant onto the tuning card (rule #25),
closing the deferral from #2374.
Tag-overlap tests use deliberately SKEWED fixtures: an evenly-matching pool
cannot exercise a re-ranking, since every candidate gets the same
multiplier and the order is unchanged whether the blend works or not.
Admin UI + client attribution follow in this batch — rule #27.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
86 lines
2.8 KiB
SQL
86 lines
2.8 KiB
SQL
-- Recommendation tuning lab queries (#1250). Seeding happens via the
|
|
-- recsettings boot reconcile; shipped defaults live in Go only.
|
|
|
|
-- name: UpsertWeightProfileDefaults :exec
|
|
-- Boot reconcile: insert the shipped defaults for a profile if the row
|
|
-- doesn't exist yet. Never overwrites operator-tuned values.
|
|
INSERT INTO recommendation_weight_profiles (
|
|
profile, base_weight, like_boost, recency_weight, skip_penalty,
|
|
jitter_magnitude, context_weight, similarity_weight, taste_weight,
|
|
context_time_weight
|
|
) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10)
|
|
ON CONFLICT (profile) DO NOTHING;
|
|
|
|
-- name: ListWeightProfiles :many
|
|
SELECT * FROM recommendation_weight_profiles ORDER BY profile;
|
|
|
|
-- name: UpdateWeightProfile :one
|
|
UPDATE recommendation_weight_profiles
|
|
SET base_weight = $2,
|
|
like_boost = $3,
|
|
recency_weight = $4,
|
|
skip_penalty = $5,
|
|
jitter_magnitude = $6,
|
|
context_weight = $7,
|
|
similarity_weight = $8,
|
|
taste_weight = $9,
|
|
context_time_weight = $10,
|
|
updated_at = now()
|
|
WHERE profile = $1
|
|
RETURNING *;
|
|
|
|
-- name: UpsertTasteTuningDefaults :exec
|
|
INSERT INTO taste_tuning (
|
|
singleton, half_life_days, engagement_hard_skip,
|
|
engagement_neutral, engagement_full, enriched_tag_scale, era_scale,
|
|
mood_scale
|
|
) VALUES (true, $1, $2, $3, $4, $5, $6, $7)
|
|
ON CONFLICT (singleton) DO NOTHING;
|
|
|
|
-- name: GetTasteTuning :one
|
|
SELECT * FROM taste_tuning WHERE singleton = true;
|
|
|
|
-- name: UpdateTasteTuning :one
|
|
UPDATE taste_tuning
|
|
SET half_life_days = $1,
|
|
engagement_hard_skip = $2,
|
|
engagement_neutral = $3,
|
|
engagement_full = $4,
|
|
enriched_tag_scale = $5,
|
|
era_scale = $6,
|
|
mood_scale = $7,
|
|
updated_at = now()
|
|
WHERE singleton = true
|
|
RETURNING *;
|
|
|
|
-- name: UpsertDiscoverTuningDefaults :exec
|
|
-- Boot reconcile for the Discover scope (#2377). Never overwrites
|
|
-- operator-tuned values, same contract as the other two.
|
|
INSERT INTO discover_tuning (singleton, tag_overlap_weight, snooze_days)
|
|
VALUES (true, $1, $2)
|
|
ON CONFLICT (singleton) DO NOTHING;
|
|
|
|
-- name: GetDiscoverTuning :one
|
|
SELECT * FROM discover_tuning WHERE singleton = true;
|
|
|
|
-- name: UpdateDiscoverTuning :one
|
|
UPDATE discover_tuning
|
|
SET tag_overlap_weight = $1,
|
|
snooze_days = $2,
|
|
updated_at = now()
|
|
WHERE singleton = true
|
|
RETURNING *;
|
|
|
|
-- name: InsertTuningAudit :exec
|
|
-- changes is a jsonb array of {field, old, new} objects.
|
|
INSERT INTO recommendation_tuning_audit (scope, action, changes)
|
|
VALUES ($1, $2, $3);
|
|
|
|
-- name: ListTuningAudit :many
|
|
-- Newest first; consumed by the metrics trend view (#1251) to annotate
|
|
-- knob turns on the timeline.
|
|
SELECT id, changed_at, scope, action, changes
|
|
FROM recommendation_tuning_audit
|
|
ORDER BY changed_at DESC, id DESC
|
|
LIMIT $1;
|