ollama_corpus_rerank

Rerank hits. Post-retrieval re-sort of hits from a prior ollama_corpus_search. No Ollama call (no embed, no generate). Three modes: ‘recency’ (newer file mtime wins; stats the file), ‘path_specificity’ (deeper paths win), ‘lexical_boost’ (boosts hits whose preview/heading_path/title contains any lexical_terms — case-insensitive, word-boundary; lexical_terms REQUIRED for this mode). Preserves each hit’s original score + original_rank; appends rerank_score + rank. Use after corpus_search when you need a different ordering heuristic than semantic similarity.

Parameters

Parameter	Type	Required	Default	Description
hits	object[]	yes	—	Hits from a prior ollama_corpus_search (or equivalent). Required fields: id, path, score, chunk_index. Optional: preview, heading_path, title.
rerank_by	"recency" \	"path_specificity" \	"lexical_boost"	yes
lexical_terms	string[]	no	—	Required when rerank_by=‘lexical_boost’. Each term is matched case-insensitively with word-boundaries against preview + heading_path + title.

Full input schema

The JSON Schema below is generated from the same zod schema the server validates against at the wire — it cannot drift from runtime behavior.

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "type": "object",
  "properties": {
    "hits": {
      "minItems": 1,
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "id": {
            "type": "string",
            "minLength": 1
          },
          "path": {
            "type": "string",
            "minLength": 1
          },
          "score": {
            "type": "number"
          },
          "chunk_index": {
            "type": "integer",
            "minimum": -9007199254740991,
            "maximum": 9007199254740991
          },
          "preview": {
            "type": "string"
          },
          "heading_path": {
            "type": "array",
            "items": {
              "type": "string"
            }
          },
          "title": {
            "anyOf": [
              {
                "type": "string"
              },
              {
                "type": "null"
              }
            ]
          }
        },
        "required": [
          "id",
          "path",
          "score",
          "chunk_index"
        ]
      },
      "description": "Hits from a prior ollama_corpus_search (or equivalent). Required fields: id, path, score, chunk_index. Optional: preview, heading_path, title."
    },
    "rerank_by": {
      "type": "string",
      "enum": [
        "recency",
        "path_specificity",
        "lexical_boost"
      ],
      "description": "'recency' stats file mtime (newer wins). 'path_specificity' prefers deeper paths. 'lexical_boost' requires lexical_terms."
    },
    "lexical_terms": {
      "description": "Required when rerank_by='lexical_boost'. Each term is matched case-insensitively with word-boundaries against preview + heading_path + title.",
      "type": "array",
      "items": {
        "type": "string",
        "minLength": 1
      }
    }
  },
  "required": [
    "hits",
    "rerank_by"
  ]
}

Envelope

Every tool returns the standard envelope — tier_used, model, tokens_in/tokens_out, elapsed_ms, backend provenance, warnings. See Envelope & tiers.

Source

Schema + handler: src/tools/corpusRerank.ts · registration: src/index.ts

---

Back to the Tool Reference.