@mailwoman/corpus 9.1.0 → 9.3.0

package/package.json

@@ -1,12 +1,12 @@
 {
   "name": "@mailwoman/corpus",
-  "version": "9.1.0",
+  "version": "9.3.0",
   "description": "Mailwoman corpus pipeline: BIO-labeled dataset builder for the neural classifier.",
   "license": "AGPL-3.0-only OR LicenseRef-Commercial",
   "repository": {
     "type": "git",
     "url": "https://github.com/sister-software/mailwoman.git",
-    "directory": "corpus"
+    "directory": "packages/corpus"
   },
   "files": [
     "data/**",
@@ -22,62 +22,198 @@
     "!*.test.tsx",
     "!**/*.test.ts",
     "!**/*.test.tsx",
-    "!test-kit/**/*"
+    "!test/**"
   ],
   "type": "module",
   "imports": {
-    "#test-kit": "./test-kit/index.ts"
+    "#adapters/utils": {
+      "types": "./out/adapters/utils/index.d.ts",
+      "node": "./out/adapters/utils/index.js",
+      "default": "./out/adapters/utils/index.js"
+    },
+    "#runner": {
+      "types": "./out/runner.d.ts",
+      "node": "./out/runner.js",
+      "default": "./out/runner.js"
+    },
+    "#synthesizers/utils": {
+      "types": "./out/synthesizers/utils.d.ts",
+      "node": "./out/synthesizers/utils.js",
+      "default": "./out/synthesizers/utils.js"
+    },
+    "#types": {
+      "types": "./out/types.d.ts",
+      "node": "./out/types.js",
+      "default": "./out/types.js"
+    },
+    "#utils": {
+      "types": "./out/utils/index.d.ts",
+      "node": "./out/utils/index.js",
+      "default": "./out/utils/index.js"
+    },
+    "#tools": {
+      "types": "./out/tools/index.d.ts",
+      "node": "./out/tools/index.js",
+      "default": "./out/tools/index.js"
+    },
+    "#*": {
+      "types": "./out/*.d.ts",
+      "node": "./out/*.js",
+      "default": "./out/*.js"
+    }
   },
   "exports": {
     "./package.json": "./package.json",
     ".": {
-      "types": "./out/src/index.d.ts",
-      "default": "./out/src/index.js"
+      "types": "./out/index.d.ts",
+      "node": "./out/index.js",
+      "default": "./out/index.js"
+    },
+    "./env": {
+      "types": "./out/env.d.ts",
+      "node": "./out/env.js",
+      "default": "./out/env.js"
     },
     "./types": {
-      "types": "./out/src/types.d.ts",
-      "default": "./out/src/types.js"
+      "types": "./out/types.d.ts",
+      "node": "./out/types.js",
+      "default": "./out/types.js"
     },
     "./data/*": "./data/*",
     "./utils": {
-      "types": "./out/src/utils/index.d.ts",
-      "default": "./out/src/utils/index.js"
-    },
-    "./parquet-wrapper": {
-      "types": "./out/src/parquet-wrapper/index.d.ts",
-      "default": "./out/src/parquet-wrapper/index.js"
+      "types": "./out/utils/index.d.ts",
+      "node": "./out/utils/index.js",
+      "default": "./out/utils/index.js"
     },
     "./runner": {
-      "types": "./out/src/runner.d.ts",
-      "default": "./out/src/runner.js"
+      "types": "./out/runner.d.ts",
+      "node": "./out/runner.js",
+      "default": "./out/runner.js"
     },
     "./build": {
-      "types": "./out/src/build.d.ts",
-      "default": "./out/src/build.js"
+      "types": "./out/build.d.ts",
+      "node": "./out/build.js",
+      "default": "./out/build.js"
     },
     "./synthesizers/*": {
-      "types": "./out/src/synthesizers/*.d.ts",
-      "default": "./out/src/synthesizers/*.js"
+      "types": "./out/synthesizers/*.d.ts",
+      "node": "./out/synthesizers/*.js",
+      "default": "./out/synthesizers/*.js"
+    },
+    "./recipes/*": {
+      "types": "./out/recipes/*.d.ts",
+      "node": "./out/recipes/*.js",
+      "default": "./out/recipes/*.js"
     },
     "./adapters/utils": {
-      "types": "./out/src/adapters/utils/index.d.ts",
-      "default": "./out/src/adapters/utils/index.js"
+      "types": "./out/adapters/utils/index.d.ts",
+      "node": "./out/adapters/utils/index.js",
+      "default": "./out/adapters/utils/index.js"
     },
     "./adapters": {
-      "types": "./out/src/adapters/index.d.ts",
-      "default": "./out/src/adapters/index.js"
+      "types": "./out/adapters/index.d.ts",
+      "node": "./out/adapters/index.js",
+      "default": "./out/adapters/index.js"
     },
     "./adapters/*": {
-      "types": "./out/src/adapters/*.d.ts",
-      "default": "./out/src/adapters/*.js"
+      "types": "./out/adapters/*.d.ts",
+      "node": "./out/adapters/*.js",
+      "default": "./out/adapters/*.js"
     },
     "./tools": {
-      "types": "./out/src/tools/index.d.ts",
-      "default": "./out/src/tools/index.js"
+      "types": "./out/tools/index.d.ts",
+      "node": "./out/tools/index.js",
+      "default": "./out/tools/index.js"
     },
     "./codex/*": {
-      "types": "./out/src/codex/*.d.ts",
-      "default": "./out/src/codex/*.js"
+      "types": "./out/codex/*.d.ts",
+      "node": "./out/codex/*.js",
+      "default": "./out/codex/*.js"
+    },
+    "./utils/parquet": {
+      "types": "./out/utils/parquet.d.ts",
+      "node": "./out/utils/parquet.js",
+      "default": "./out/utils/parquet.js"
+    },
+    "./tools/postcode-triples": {
+      "types": "./out/tools/postcode-triples.d.ts",
+      "node": "./out/tools/postcode-triples.js",
+      "default": "./out/tools/postcode-triples.js"
+    },
+    "./tools/audit": {
+      "types": "./out/tools/audit.d.ts",
+      "node": "./out/tools/audit.js",
+      "default": "./out/tools/audit.js"
+    },
+    "./tools/gb-title-case": {
+      "types": "./out/tools/gb-title-case.d.ts",
+      "node": "./out/tools/gb-title-case.js",
+      "default": "./out/tools/gb-title-case.js"
+    },
+    "./tools/golden-relabel-street": {
+      "types": "./out/tools/golden-relabel-street.d.ts",
+      "node": "./out/tools/golden-relabel-street.js",
+      "default": "./out/tools/golden-relabel-street.js"
+    },
+    "./tools/overture-subvenue": {
+      "types": "./out/tools/overture-subvenue.d.ts",
+      "node": "./out/tools/overture-subvenue.js",
+      "default": "./out/tools/overture-subvenue.js"
+    },
+    "./tools/sub-venue-lexicon": {
+      "types": "./out/tools/sub-venue-lexicon.d.ts",
+      "node": "./out/tools/sub-venue-lexicon.js",
+      "default": "./out/tools/sub-venue-lexicon.js"
+    },
+    "./tools/sub-venue-promotions": {
+      "types": "./out/tools/sub-venue-promotions.d.ts",
+      "node": "./out/tools/sub-venue-promotions.js",
+      "default": "./out/tools/sub-venue-promotions.js"
+    },
+    "./tools/fetch/download": {
+      "types": "./out/tools/fetch/download.d.ts",
+      "node": "./out/tools/fetch/download.js",
+      "default": "./out/tools/fetch/download.js"
+    },
+    "./tools/fetch/ppd": {
+      "types": "./out/tools/fetch/ppd.d.ts",
+      "node": "./out/tools/fetch/ppd.js",
+      "default": "./out/tools/fetch/ppd.js"
+    },
+    "./utils/tokenize": {
+      "types": "./out/utils/tokenize.d.ts",
+      "node": "./out/utils/tokenize.js",
+      "default": "./out/utils/tokenize.js"
+    },
+    "./utils/golden": {
+      "types": "./out/utils/golden.d.ts",
+      "node": "./out/utils/golden.js",
+      "default": "./out/utils/golden.js"
+    },
+    "./utils/license": {
+      "types": "./out/utils/license.d.ts",
+      "node": "./out/utils/license.js",
+      "default": "./out/utils/license.js"
+    },
+    "./utils/split": {
+      "types": "./out/utils/split.d.ts",
+      "node": "./out/utils/split.js",
+      "default": "./out/utils/split.js"
+    },
+    "./utils/wof-json": {
+      "types": "./out/utils/wof-json.d.ts",
+      "node": "./out/utils/wof-json.js",
+      "default": "./out/utils/wof-json.js"
+    },
+    "./test-kit": {
+      "types": "./out/test-kit/index.d.ts",
+      "node": "./out/test-kit/index.js",
+      "default": "./out/test-kit/index.js"
+    },
+    "./test-kit/corpus-recipe": {
+      "types": "./out/test-kit/corpus-recipe.d.ts",
+      "node": "./out/test-kit/corpus-recipe.js",
+      "default": "./out/test-kit/corpus-recipe.js"
     }
   },
   "publishConfig": {
@@ -85,69 +221,218 @@
     "exports": {
       "./package.json": "./package.json",
       ".": {
-        "types": "./out/src/index.d.ts",
-        "default": "./out/src/index.js"
+        "types": "./out/index.d.ts",
+        "node": "./out/index.js",
+        "default": "./out/index.js"
+      },
+      "./env": {
+        "types": "./out/env.d.ts",
+        "node": "./out/env.js",
+        "default": "./out/env.js"
       },
       "./types": {
-        "types": "./out/src/types.d.ts",
-        "default": "./out/src/types.js"
+        "types": "./out/types.d.ts",
+        "node": "./out/types.js",
+        "default": "./out/types.js"
       },
       "./data/*": "./data/*",
       "./utils": {
-        "types": "./out/src/utils/index.d.ts",
-        "default": "./out/src/utils/index.js"
-      },
-      "./parquet-wrapper": {
-        "types": "./out/src/parquet-wrapper/index.d.ts",
-        "default": "./out/src/parquet-wrapper/index.js"
+        "types": "./out/utils/index.d.ts",
+        "node": "./out/utils/index.js",
+        "default": "./out/utils/index.js"
       },
       "./runner": {
-        "types": "./out/src/runner.d.ts",
-        "default": "./out/src/runner.js"
+        "types": "./out/runner.d.ts",
+        "node": "./out/runner.js",
+        "default": "./out/runner.js"
       },
       "./build": {
-        "types": "./out/src/build.d.ts",
-        "default": "./out/src/build.js"
+        "types": "./out/build.d.ts",
+        "node": "./out/build.js",
+        "default": "./out/build.js"
       },
       "./synthesizers/*": {
-        "types": "./out/src/synthesizers/*.d.ts",
-        "default": "./out/src/synthesizers/*.js"
+        "types": "./out/synthesizers/*.d.ts",
+        "node": "./out/synthesizers/*.js",
+        "default": "./out/synthesizers/*.js"
+      },
+      "./recipes/*": {
+        "types": "./out/recipes/*.d.ts",
+        "node": "./out/recipes/*.js",
+        "default": "./out/recipes/*.js"
       },
       "./adapters/utils": {
-        "types": "./out/src/adapters/utils/index.d.ts",
-        "default": "./out/src/adapters/utils/index.js"
+        "types": "./out/adapters/utils/index.d.ts",
+        "node": "./out/adapters/utils/index.js",
+        "default": "./out/adapters/utils/index.js"
       },
       "./adapters": {
-        "types": "./out/src/adapters/index.d.ts",
-        "default": "./out/src/adapters/index.js"
+        "types": "./out/adapters/index.d.ts",
+        "node": "./out/adapters/index.js",
+        "default": "./out/adapters/index.js"
       },
       "./adapters/*": {
-        "types": "./out/src/adapters/*.d.ts",
-        "default": "./out/src/adapters/*.js"
+        "types": "./out/adapters/*.d.ts",
+        "node": "./out/adapters/*.js",
+        "default": "./out/adapters/*.js"
       },
       "./tools": {
-        "types": "./out/src/tools/index.d.ts",
-        "default": "./out/src/tools/index.js"
+        "types": "./out/tools/index.d.ts",
+        "node": "./out/tools/index.js",
+        "default": "./out/tools/index.js"
       },
       "./codex/*": {
-        "types": "./out/src/codex/*.d.ts",
-        "default": "./out/src/codex/*.js"
+        "types": "./out/codex/*.d.ts",
+        "node": "./out/codex/*.js",
+        "default": "./out/codex/*.js"
+      },
+      "./utils/parquet": {
+        "types": "./out/utils/parquet.d.ts",
+        "node": "./out/utils/parquet.js",
+        "default": "./out/utils/parquet.js"
+      },
+      "./tools/postcode-triples": {
+        "types": "./out/tools/postcode-triples.d.ts",
+        "node": "./out/tools/postcode-triples.js",
+        "default": "./out/tools/postcode-triples.js"
+      },
+      "./tools/audit": {
+        "types": "./out/tools/audit.d.ts",
+        "node": "./out/tools/audit.js",
+        "default": "./out/tools/audit.js"
+      },
+      "./tools/gb-title-case": {
+        "types": "./out/tools/gb-title-case.d.ts",
+        "node": "./out/tools/gb-title-case.js",
+        "default": "./out/tools/gb-title-case.js"
+      },
+      "./tools/golden-relabel-street": {
+        "types": "./out/tools/golden-relabel-street.d.ts",
+        "node": "./out/tools/golden-relabel-street.js",
+        "default": "./out/tools/golden-relabel-street.js"
+      },
+      "./tools/overture-subvenue": {
+        "types": "./out/tools/overture-subvenue.d.ts",
+        "node": "./out/tools/overture-subvenue.js",
+        "default": "./out/tools/overture-subvenue.js"
+      },
+      "./tools/sub-venue-lexicon": {
+        "types": "./out/tools/sub-venue-lexicon.d.ts",
+        "node": "./out/tools/sub-venue-lexicon.js",
+        "default": "./out/tools/sub-venue-lexicon.js"
+      },
+      "./tools/sub-venue-promotions": {
+        "types": "./out/tools/sub-venue-promotions.d.ts",
+        "node": "./out/tools/sub-venue-promotions.js",
+        "default": "./out/tools/sub-venue-promotions.js"
+      },
+      "./tools/fetch/download": {
+        "types": "./out/tools/fetch/download.d.ts",
+        "node": "./out/tools/fetch/download.js",
+        "default": "./out/tools/fetch/download.js"
+      },
+      "./tools/fetch/ppd": {
+        "types": "./out/tools/fetch/ppd.d.ts",
+        "node": "./out/tools/fetch/ppd.js",
+        "default": "./out/tools/fetch/ppd.js"
+      },
+      "./utils/tokenize": {
+        "types": "./out/utils/tokenize.d.ts",
+        "node": "./out/utils/tokenize.js",
+        "default": "./out/utils/tokenize.js"
+      },
+      "./utils/golden": {
+        "types": "./out/utils/golden.d.ts",
+        "node": "./out/utils/golden.js",
+        "default": "./out/utils/golden.js"
+      },
+      "./utils/license": {
+        "types": "./out/utils/license.d.ts",
+        "node": "./out/utils/license.js",
+        "default": "./out/utils/license.js"
+      },
+      "./utils/split": {
+        "types": "./out/utils/split.d.ts",
+        "node": "./out/utils/split.js",
+        "default": "./out/utils/split.js"
+      },
+      "./utils/wof-json": {
+        "types": "./out/utils/wof-json.d.ts",
+        "node": "./out/utils/wof-json.js",
+        "default": "./out/utils/wof-json.js"
+      },
+      "./test-kit": {
+        "types": "./out/test-kit/index.d.ts",
+        "node": "./out/test-kit/index.js",
+        "default": "./out/test-kit/index.js"
+      },
+      "./test-kit/corpus-recipe": {
+        "types": "./out/test-kit/corpus-recipe.d.ts",
+        "node": "./out/test-kit/corpus-recipe.js",
+        "default": "./out/test-kit/corpus-recipe.js"
+      }
+    },
+    "imports": {
+      "#adapters/utils": {
+        "types": "./out/adapters/utils/index.d.ts",
+        "node": "./out/adapters/utils/index.js",
+        "default": "./out/adapters/utils/index.js"
+      },
+      "#runner": {
+        "types": "./out/runner.d.ts",
+        "node": "./out/runner.js",
+        "default": "./out/runner.js"
+      },
+      "#synthesizers/utils": {
+        "types": "./out/synthesizers/utils.d.ts",
+        "node": "./out/synthesizers/utils.js",
+        "default": "./out/synthesizers/utils.js"
+      },
+      "#types": {
+        "types": "./out/types.d.ts",
+        "node": "./out/types.js",
+        "default": "./out/types.js"
+      },
+      "#utils": {
+        "types": "./out/utils/index.d.ts",
+        "node": "./out/utils/index.js",
+        "default": "./out/utils/index.js"
+      },
+      "#tools": {
+        "types": "./out/tools/index.d.ts",
+        "node": "./out/tools/index.js",
+        "default": "./out/tools/index.js"
+      },
+      "#*": {
+        "types": "./out/*.d.ts",
+        "node": "./out/*.js",
+        "default": "./out/*.js"
       }
     }
   },
   "dependencies": {
-    "@dsnp/parquetjs": "1.8.9",
-    "@mailwoman/ban": "9.1.0",
-    "@mailwoman/codex": "9.1.0",
-    "@mailwoman/core": "9.1.0",
-    "@mailwoman/formatter": "9.1.0",
+    "@mailwoman/ban": "9.3.0",
+    "@mailwoman/bdc": "9.3.0",
+    "@mailwoman/codex": "9.3.0",
+    "@mailwoman/core": "9.3.0",
+    "@mailwoman/formatter": "9.3.0",
+    "@mailwoman/normalize": "9.3.0",
+    "@mailwoman/resolver-wof-sqlite": "9.3.0",
+    "@mailwoman/sqlite": "9.3.0",
+    "@mailwoman/tiger": "9.3.0",
+    "apache-arrow": "^21.2.0",
     "fast-glob": "^3.3.3",
     "fastest-levenshtein": "^1.0.16",
-    "lru-cache": "^11.5.2",
-    "path-ts": "^2.1.0",
-    "spliterator": "^6.1.0"
+    "parquet-wasm": "0.7.2",
+    "path-ts": "^2.3.0",
+    "read-excel-file": "^9.3.10",
+    "spliterator": "^6.4.1",
+    "zod": "^4.5.4"
   },
   "devDependencies": {
-    "@types/node": ">=26.1.2"
+    "@types/node": "^26.4.1",
+    "adm-zip": "^0.6.0",
+    "kysely": "^0.29.5",
+    "write-excel-file": "^4.1.1"
   }
 }

package/out/src/adapters/ban/adapter.d.ts

@@ -1,44 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   `ban`: Base Adresse Nationale CSV adapter (FR street-level).
- *
- *   Input: a CSV dump from `adresse.data.gouv.fr` (semicolon-separated, ~25M rows nationally). The
- *   adapter only reads the small set of columns needed for the corpus:
- *
- *   - `numero` → `house_number`
- *   - `rep` → repetition index ("bis", "ter") appended to house_number
- *   - `nom_voie` → `street` (full road name; includes the prefix "Rue", "Avenue", etc.)
- *   - `code_postal` → `postcode`
- *   - `nom_commune` → `locality`
- *
- *   `region` and `country` are not in BAN. The adapter stamps `country: "FR"` on every row; region is
- *   left for the wof-postalcode + wof-admin cross-reference at corpus build time (a future pass;
- *   for Phase 1 the row's region is simply absent).
- *
- *   License: the official BAN (adresse.data.gouv.fr) is DUAL-licensed — Licence Ouverte 2.0 (Etalab,
- *   attribution-only) OR ODbL (share-alike). We ELECT Licence Ouverte 2.0 (issue #26 Tier B:
- *   allowed for training with attribution; the ODbL option's share-alike obligation would defeat
- *   the proprietary-weights goal). Stamped onto every row as `Licence Ouverte 2.0` — NOT the older
- *   conservative `ODbL-1.0` label, which wrongly read as Tier-C-denied in the corpus license audit.
- *   The model card MUST carry the BAN attribution (Tier B obligation).
- *
- *   The adapter is streaming-aware: `CSVSpliterator.fromAsync` reads the `;`-delimited dump row by
- *   row, so a 25M-row file never sits in memory. Honors `opts.limit` for fixture / smoke runs,
- *   `opts.signal` for cancellation, and `opts.country` for a self-consistency check (errors if
- *   country !== FR).
- */
-import type { CorpusAdapter } from "@mailwoman/corpus/types";
-/**
- * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
- * it came from.
- */
-export declare const BAN_ADAPTER_ID = "ban";
-export declare function createBanAdapter(): CorpusAdapter;
-/**
- * The configured adapter instance registered with the corpus builder.
- */
-export declare const banAdapter: CorpusAdapter;
-//# sourceMappingURL=adapter.d.ts.map

package/out/src/adapters/ban/adapter.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/ban/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AAGH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,yBAAyB,CAAA;AAM1F;;;GAGG;AACH,eAAO,MAAM,cAAc,QAAQ,CAAA;AAsDnC,wBAAgB,gBAAgB,IAAI,aAAa,CAqFhD;AAED;;GAEG;AACH,eAAO,MAAM,UAAU,eAAqB,CAAA"}