Recipe catalog / continuation-compare

Compare two musical continuations

Which of firstContinuation and secondContinuation better follows the musical context in motif, contour, harmony, and any style supplied?

A generation loop has produced two candidate continuations for a passage and wants a head-to-head preference before committing to one.

Explore this recipe interactively · Source and implementation guide

Use continuation-compare in TypeScript

Install with npm install jev-recipes. Requires Node.js 22.9 or newer and ES modules. Set TYPESAFE_API_KEY in your server environment for live calls, which send input to TypeSafe and use API quota. See the installation guide.

import { continuationCompare } from 'jev-recipes/continuation-compare';

const result = await continuationCompare({
  "context": "G major, 4/4. Opening motif: G4 A4 B4 D5 | B4 A4 G4 (half). A rising line to the fifth answered by a stepwise descent home.",
  "firstContinuation": "D5 E5 F#5 A5 | F#5 E5 D5 (half). The motif transposed up a fifth, same rhythm, same rise-and-fall contour.",
  "secondContinuation": "C#4 (eighth) G#5 (eighth) Bb3 (quarter) rest | Eb5 (dotted half). Wide chromatic leaps in a new, jagged rhythm.",
  "style": "Celtic air, lyrical and unhurried",
  "minConfidence": 0.8
});
console.log(result);

Input contract

{
  "input": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
      "context": {
        "type": "string"
      },
      "firstContinuation": {
        "type": "string"
      },
      "secondContinuation": {
        "type": "string"
      },
      "style": {
        "type": "string"
      },
      "minConfidence": {
        "type": "number",
        "minimum": 0,
        "maximum": 1
      }
    },
    "required": [
      "context",
      "firstContinuation",
      "secondContinuation"
    ]
  },
  "result": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
      "model": {
        "type": "string"
      },
      "usage": {
        "type": "object",
        "properties": {
          "input_tokens": {
            "type": "integer",
            "minimum": 0,
            "maximum": 9007199254740991
          },
          "output_tokens": {
            "type": "integer",
            "minimum": 0,
            "maximum": 9007199254740991
          }
        },
        "required": [
          "input_tokens",
          "output_tokens"
        ],
        "additionalProperties": false
      },
      "status": {
        "type": "string",
        "enum": [
          "ready",
          "review"
        ]
      },
      "verdict": {
        "type": "string",
        "enum": [
          "first",
          "second",
          "tie",
          "neither",
          "unclear"
        ]
      },
      "confidence": {
        "type": "number",
        "minimum": 0,
        "maximum": 1
      },
      "probabilities": {
        "type": "object",
        "propertyNames": {
          "type": "string",
          "enum": [
            "first",
            "second",
            "tie",
            "neither",
            "unclear"
          ]
        },
        "additionalProperties": {
          "type": "number",
          "minimum": 0,
          "maximum": 1
        },
        "required": [
          "first",
          "second",
          "tie",
          "neither",
          "unclear"
        ]
      }
    },
    "required": [
      "model",
      "usage",
      "status",
      "verdict",
      "confidence",
      "probabilities"
    ],
    "additionalProperties": false
  }
}

Saved example result

This hand-authored response demonstrates the contract. It is not a model accuracy measurement. Run it without an API key: npx jev-recipes demo continuation-compare.

{
  "model": "demo-fixture",
  "usage": {
    "input_tokens": 0,
    "output_tokens": 0
  },
  "status": "ready",
  "verdict": "first",
  "confidence": 0.91,
  "probabilities": {
    "first": 0.91,
    "second": 0.02,
    "tie": 0.02,
    "neither": 0.03,
    "unclear": 0.02
  }
}

Evaluation evidence

Use the evaluation guide to measure this decision on your own labeled cases.

Limitations

• Compares the two continuations as written. It does not verify key membership, count beats, or check voice leading; do those exactly in application code and describe the results in the text.
• A preferred continuation is only better than the other one supplied; neither may be good.
• Style preference is a judgment, so continuations that fit similarly well can tie or swap on different runs.

Related recipes

• draft-compare: Use draft-compare to choose between two text drafts under a rubric, not between two musical continuations of a passage.
• action-compare: Use action-compare to compare two next steps against a goal and constraints, rather than two continuations against a musical context.