Reciprocal Rank Fusion

knowledge.reciprocal_rank_fusion

ProductionBounded custom (Class B)Deterministicstrictfastv1.0.0

No network

Overview

You send two or more named rankings, each a list of unique document ids in rank order. Optional weights (strictly positive, one per ranking) default to 1. k defaults to 60 and limit defaults to 1000. The fused score of a document is the sum of weight / (k + rank) over rankings that contain it. Ties break by best individual rank, then lexical id. This does not retrieve documents or learn weights.

Two rankers that both saw a document should lift it without a learned fusion model.

Reliability

Scores round to nine decimals. Ties break by best individual rank, then lexical id. RRF_RANKINGS_INVALID — trigger: empty rankings, missing ranking id, or empty documents; retryable: no; next: send named non-empty lists. RRF_DOCUMENT_ID_DUPLICATE — trigger: a ranking repeats an id; retryable: no; next: deduplicate that ranking. RRF_PARAMETER_INVALID — trigger: weights length mismatch or a non-positive weight; retryable: no; next: send one positive weight per ranking. RRF_LIMIT_EXCEEDED — trigger: more than 100 rankings, k/limit above 10000, or more than 100000 ids; retryable: no; next: split the request.

Context footprint

Ids, ranks and scores only; document text is never returned. A successful call costs 1 execution credit.

When to use it

When two or more already-ranked id lists must be fused with a published RRF formula.

When documents still need retrieving, or when fusion weights must be learned from labels.

Scope

Note

Fuses named ranked document-id lists with reciprocal-rank fusion and published tie-breaks.

Non-goals

Does not retrieve documents
Does not learn weights
Does not compare document text

Properties

IdempotentRetry safeStreaming supportedFile input supportedRequires external service

Capabilities

analysissearch

Cost & billing

Flat rate per invocation

MetercallsBillable onsuccessCost per call1 credit

File handling

Accepts file_refProduces file_refMax input2 MB

How we handle your data

Sensitive processing

What this means

This skill is designed to detect, mask, redact, validate, or otherwise process sensitive data.

Squillify guarantees

Runs inside Squillify
This skill runs inside Squillify.
Payloads are not sent to an external service by this skill.
Squillify does not permanently store submitted payloads.
Telemetry records metadata only, not submitted content.
Submitted payloads are not used for training.
Submitted payloads are not reviewed by humans.
Targets sensitive data
The response does not contain submitted content.

Your responsibility

Squillify does not determine whether submitted content is sensitive. Your application is responsible for deciding what data it is allowed to submit. Your application is responsible for securing the response.

Read the Microskill Data Handling Policy

Input schema

Field	Type	Required	Description
rankings	array	Yes	From 1 through 101 closed rankings {id, documents}; runtime accepts at most 100 and 100000 ids in aggregate.
weights	array	No	Optional weights aligned with rankings; omit it for weight 1 on every ranking.
k	integer	No	RRF constant from 1 through 10001; omit it for 60.
limit	integer	No	Fused-list length from 1 through 10001; omit it for 1000.

Output schema

Field	Type	Required	Description
fused	array	Yes	Fused documents best first, each a closed {rank, id, score, bestRank, contributions} object.
k	integer	Yes	RRF constant k that was actually applied to every ranking.
limit	integer	Yes	Fused-list length that was actually applied after clamping to the ranked set.
rankingCount	integer	Yes	How many rankings were fused.
documentCount	integer	Yes	How many distinct document ids scored.
truncated	boolean	Yes	True when limit cut the fused ranking short.
formulaVersion	string	Yes	Pinned reciprocal-rank-fusion formula identifier.

API contract

POST /v1/execute

curl

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"knowledge.reciprocal_rank_fusion","input":{"rankings":[{"id":"lexical","documents":["D1","D2"]},{"id":"vector","documents":["D2","D3"]}]}}'

Integrations

TypeScript SDK

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("knowledge.reciprocal_rank_fusion", {"rankings":[{"id":"lexical","documents":["D1","D2"]},{"id":"vector","documents":["D2","D3"]}]});

Errors

Code	HTTP	Retryable	Description
VALIDATION_ERROR	400	No	The request payload failed schema validation.
UNAUTHORIZED	401	No	Missing or invalid API key.
SKILL_NOT_ACCESSIBLE	403	No	The API key is not granted access to this skill.
SKILL_NOT_AVAILABLE_FOR_PLAN	403	No	This skill is not included in your current plan.
PLAN_ELIGIBILITY_UNAVAILABLE	503	Yes	Plan eligibility could not be decided. Retry after backoff.
SKILL_NOT_FOUND	404	No	The requested skill or version does not exist.
TENANT_RATE_LIMITED	429	Yes	Request rate limit exceeded. Retry after backoff.
QUOTA_EXCEEDED	429	Yes	Plan quota exhausted. Retryable once the quota period resets.
EXECUTION_TIMEOUT	504	Yes	Execution exceeded the allowed time budget.
RRF_RANKINGS_INVALID	400	No	The request payload failed this skill's input validation.
RRF_DOCUMENT_ID_DUPLICATE	400	No	The request payload failed this skill's input validation.
RRF_PARAMETER_INVALID	400	No	The request payload failed this skill's input validation.
RRF_LIMIT_EXCEEDED	429	No	A size, quota, rate, or time limit was exceeded.

Other limits

Limit	Value
Recommended inline bytes	32 KiB

Limitations

Does not retrieve documents
Does not learn weights
Does not compare document text
Searching a corpus you have not ranked
Learning fusion weights from labels

How to call

POST /v1/execute (requires authentication)

Request bodyjson

{
  "skillKey": "knowledge.reciprocal_rank_fusion",
  "input": {
    "rankings": [
      {
        "id": "lexical",
        "documents": [
          "D1",
          "D2"
        ]
      },
      {
        "id": "vector",
        "documents": [
          "D2",
          "D3"
        ]
      }
    ]
  }
}

Examples

D2 outranks D1 when it appears in both lists

Inputjson

{
  "rankings": [
    {
      "id": "lexical",
      "documents": [
        "D1",
        "D2"
      ]
    },
    {
      "id": "vector",
      "documents": [
        "D2",
        "D3"
      ]
    }
  ]
}

Expected outputjson

{
  "fused": [
    {
      "rank": 1,
      "id": "D2",
      "score": 0.032522475,
      "bestRank": 1,
      "contributions": [
        {
          "rankingId": "lexical",
          "rank": 2,
          "contribution": 0.016129032
        },
        {
          "rankingId": "vector",
          "rank": 1,
          "contribution": 0.016393443
        }
      ]
    },
    {
      "rank": 2,
      "id": "D1",
      "score": 0.016393443,
      "bestRank": 1,
      "contributions": [
        {
          "rankingId": "lexical",
          "rank": 1,
          "contribution": 0.016393443
        }
      ]
    },
    {
      "rank": 3,
      "id": "D3",
      "score": 0.016129032,
      "bestRank": 2,
      "contributions": [
        {
          "rankingId": "vector",
          "rank": 2,
          "contribution": 0.016129032
        }
      ]
    }
  ],
  "k": 60,
  "documentCount": 3,
  "rankingCount": 2,
  "truncated": false,
  "limit": 1000,
  "formulaVersion": "knowledge.reciprocal_rank_fusion.v1"
}

Tip

D2 has the highest fused score because both rankings list it.

Equal fused scores break on best rank then lexical id