khodo/product-vamp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Cleanup

Browse Source
This commit is contained in:
khodowany
2026-10-04 18:48:02 -05:00
parent 68df7292ee
commit 0e68f5364b
58 changed files with 1416 additions and 7567 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/composer/evaluation_contract.md
+13 -16
View File
@@ -1,23 +1,20 @@
# Evaluation and validity
# One musical evaluator
The active final selector remains `scoring.musical` / `MusicalEvaluation`, version `joint-musical-1`. Each phrase score is its evaluation total; the complete solo uses duration-weighted completed phrase evaluations. This revision changes construction coverage, not the final weights.
`scoring.evaluate` returns `Evaluation(version='phrase-musical-1')`: validity, raw values, enabled weights, weighted contributions, total, and exact-event normalization counts. The selected phrase score is exactly its reported total. Selection additionally reports the bounded next-phrase total; it is not disguised as the local score.
| Component | Weight |
|---|---:|
| Rhythmic grouping/interruptions | 3 |
| Destination preparation | 2 |
| Harmonic exposure | 2 |
| Soft register preference | 1 |
| Contextual repetition | 1 |
| Activity | 1 |
| Approximate playing difficulty | 0.5 |
Lookahead averages the current normalized phrase score and the best feasible following phrase score. Candidate phrases can have different durations. This is a bounded, variable-horizon heuristic, not equal-duration whole-chorus optimization or a proven remaining-cost bound. Both local and lookahead totals appear in the decision trace so this limitation is inspectable.
Reports expose raw values, denominators, weights, contributions, total and validity. Identity, expression completion, physical domain and protected holds are hard contracts, not duplicate resemblance costs. Early retrieval estimates remain separate.
| Contribution | Weight | Interpretation |
|---|---:|---|
| Purpose | 2 | Unprepared large motion, repeated oscillation and preparation of the chosen destination |
| Continuity | 2 | Incoming musical connection beyond a comfortable interval |
| Repetition | 1 | Recent complete rhythmic/interval figure overuse; deliberate return receives bounded allowance |
| Register | 1 | Duration-normalized excursions beyond the role's broad comfortable region |
Planned arrival spans have arrival activity semantics; they need no attack on every held beat. Silence cannot exploit a missing register value. Ties merge before harmonic duration/attack evaluation; repeated attacks cannot fulfill a hold.
These are contextual preferences, not a perceptual memorability model. Broad acceptability can produce equal musical totals. A stable role-specific choice key breaks those ties. Guitar travel is absent from this total: it chooses a fingering for the selected music rather than giving an easier but otherwise unrelated tune an automatic advantage.
Joint arrival/color/approach candidates exist before route search. An approach into future harmony requires the exact shared anticipation target. Mapped upper neighbors must actually belong to the collection. The stable minor-sixth policy is bounded by harmony/function/duration and does not legalize unrelated notes.
Harmonic validity, complete typed resolutions, exact duration, promised held destinations, rhythmic-group integrity, physical range and arrangement allocation are checked independently. Approved expression has no generic chromatic penalty or reward. Active empty candidates cannot pass; explicit no-harmony space can.
Selected ordinary arrival pitches and tails are also protected through refinement; ordinary computational endpoints remain provisional. Disabling expressions removes joint expression proposals and local ornaments and may choose another valid complete phrase. It does not mean deleting exceptional notes from a winner.
Rests are constructed within groups, at entrances, or after destinations. There is no global attack-density/occupancy optimization. Holds count as one sounding event, regardless of tied notation fragments. A repeated attack does not fulfill a single committed sustain.
Reference recognition, proposed candidate, validity, retention, selected use and exact event matching are reported separately. The optional comparison has no scoring-return channel. Exact agreement is not a production loss function.
Ordinary phrase endpoints may be mapped non-chord tones when the phrase remains open. Local settling, completion and handoff have actual timed destinations. Expression disabling reconstructs valid strict phrases; it does not delete notes from a chromatic winner.