@mailwoman/corpus 9.2.0 → 9.3.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +22 -22
- package/data/PROVENANCE.md +7 -7
- package/lib/adapters/ban/adapter.ts +190 -0
- package/lib/adapters/ban/street-decompose.ts +66 -0
- package/lib/adapters/fcc-bdc/adapter.ts +174 -0
- package/lib/adapters/geonames/adapter.ts +199 -0
- package/lib/adapters/geonames-postal/adapter.ts +122 -0
- package/lib/adapters/gnaf/adapter.ts +149 -0
- package/lib/adapters/gnaf/assemble.ts +225 -0
- package/lib/adapters/index.ts +153 -0
- package/lib/adapters/openaddresses/adapter.ts +251 -0
- package/lib/adapters/overture/adapter.ts +162 -0
- package/lib/adapters/state-hi-schools/adapter.ts +170 -0
- package/lib/adapters/state-hi-schools/workbook.ts +35 -0
- package/lib/adapters/synth-po-box/adapter.ts +204 -0
- package/lib/adapters/tiger/adapter.ts +252 -0
- package/lib/adapters/tiger/street-decompose.ts +83 -0
- package/lib/adapters/usgov-hrsa-fqhc/adapter.ts +167 -0
- package/lib/adapters/usgov-imls-pls/adapter.ts +143 -0
- package/lib/adapters/usgov-irs-bmf/adapter.ts +151 -0
- package/lib/adapters/usgov-nad/adapter.ts +334 -0
- package/lib/adapters/usgov-nppes/adapter.ts +151 -0
- package/lib/adapters/usgov-samhsa-treatment-locator/adapter.ts +193 -0
- package/lib/adapters/utils/index.ts +307 -0
- package/lib/adapters/wof-admin-jp/adapter.ts +214 -0
- package/lib/adapters/wof-admin-json/adapter.ts +231 -0
- package/lib/adapters/wof-json-rows.ts +129 -0
- package/lib/adapters/wof-postalcode-json/adapter.ts +176 -0
- package/lib/build.ts +378 -0
- package/lib/env.ts +50 -0
- package/lib/index.ts +13 -0
- package/lib/name-prone-us-suffixes.ts +12 -0
- package/lib/recipes/anchor-absorption.ts +101 -0
- package/lib/recipes/bare-country.ts +95 -0
- package/lib/recipes/boundary-stress.ts +100 -0
- package/lib/recipes/cn-organizational-units.ts +199 -0
- package/lib/recipes/country-balanced.ts +509 -0
- package/lib/recipes/cz-pcfirst-preposition.ts +111 -0
- package/lib/recipes/fr-admin-split.ts +242 -0
- package/lib/recipes/fr-bare-street.ts +170 -0
- package/lib/recipes/fr-fragment.ts +333 -0
- package/lib/recipes/fr-lieudit.ts +301 -0
- package/lib/recipes/fr-order.ts +245 -0
- package/lib/recipes/german.ts +359 -0
- package/lib/recipes/house-venue.ts +92 -0
- package/lib/recipes/index.ts +97 -0
- package/lib/recipes/intersection.ts +618 -0
- package/lib/recipes/locale.ts +599 -0
- package/lib/recipes/nl-postcode.ts +115 -0
- package/lib/recipes/no-fragment.ts +219 -0
- package/lib/recipes/no-street-led.ts +141 -0
- package/lib/recipes/no-street.ts +90 -0
- package/lib/recipes/po-box-cedex.ts +976 -0
- package/lib/recipes/po-box.ts +123 -0
- package/lib/recipes/reviewed-postcode-tail.ts +190 -0
- package/lib/recipes/scaffold.ts +469 -0
- package/lib/recipes/si-bare-village.ts +97 -0
- package/lib/recipes/street-affix.ts +804 -0
- package/lib/recipes/street-bare.ts +100 -0
- package/lib/recipes/street.ts +81 -0
- package/lib/recipes/sub-venue-sources.ts +877 -0
- package/lib/recipes/sub-venue.ts +997 -0
- package/lib/recipes/trailing-region.ts +183 -0
- package/lib/recipes/unit.ts +304 -0
- package/lib/runner.ts +312 -0
- package/lib/synthesizers/anchor-absorption.ts +288 -0
- package/lib/synthesizers/boundary-stress.ts +532 -0
- package/lib/synthesizers/german.ts +260 -0
- package/lib/synthesizers/house-venue.ts +308 -0
- package/lib/synthesizers/intersection.ts +233 -0
- package/lib/synthesizers/no-street.ts +327 -0
- package/lib/synthesizers/po-box.ts +323 -0
- package/lib/synthesizers/street.ts +269 -0
- package/lib/synthesizers/utils.ts +993 -0
- package/lib/test-kit/corpus-recipe.ts +86 -0
- package/lib/test-kit/index.ts +100 -0
- package/lib/tools/align-slice.ts +68 -0
- package/lib/tools/audit.ts +403 -0
- package/lib/tools/corpus-stats.ts +136 -0
- package/lib/tools/fetch/ban.ts +289 -0
- package/lib/tools/fetch/download.ts +262 -0
- package/lib/tools/fetch/geonames-dump.ts +257 -0
- package/lib/tools/fetch/geonames-postal.ts +180 -0
- package/lib/tools/fetch/hrsa.ts +59 -0
- package/lib/tools/fetch/imls-pls.ts +149 -0
- package/lib/tools/fetch/index.ts +147 -0
- package/lib/tools/fetch/nad.ts +341 -0
- package/lib/tools/fetch/nppes.ts +165 -0
- package/lib/tools/fetch/openaddresses.ts +347 -0
- package/lib/tools/fetch/ourairports.ts +170 -0
- package/lib/tools/fetch/ppd.ts +121 -0
- package/lib/tools/fetch/state-hi-schools.ts +161 -0
- package/lib/tools/fetch/state-sources.ts +146 -0
- package/lib/tools/fetch/tiger-full.ts +308 -0
- package/lib/tools/fetch/wikidata-subvenue.ts +384 -0
- package/lib/tools/golden-expand.ts +574 -0
- package/lib/tools/golden-promote.ts +337 -0
- package/lib/tools/golden-relabel-street.ts +683 -0
- package/lib/tools/index.ts +28 -0
- package/lib/tools/ingest-csv.ts +378 -0
- package/lib/tools/jsonl-to-parquet.ts +207 -0
- package/lib/tools/lint-slice-vocab.ts +388 -0
- package/lib/tools/lint-slice.ts +496 -0
- package/lib/tools/overlay-manifest.ts +148 -0
- package/lib/tools/overture-subvenue.ts +209 -0
- package/lib/tools/postcode-triples.ts +363 -0
- package/lib/tools/slice-kryptonite.ts +137 -0
- package/lib/tools/slice-translit.ts +218 -0
- package/lib/tools/sub-venue/harvest.ts +234 -0
- package/lib/tools/sub-venue/head-nouns.ts +245 -0
- package/lib/tools/sub-venue/surfaces.ts +95 -0
- package/lib/tools/sub-venue/table.ts +281 -0
- package/lib/tools/sub-venue/wikidata.ts +85 -0
- package/lib/tools/sub-venue-lexicon.ts +429 -0
- package/lib/tools/sub-venue-promotions.ts +330 -0
- package/lib/types.ts +247 -0
- package/lib/utils/align.ts +330 -0
- package/lib/utils/golden.ts +189 -0
- package/lib/utils/index.ts +14 -0
- package/lib/utils/parquet.ts +582 -0
- package/lib/utils/slice-stats.ts +95 -0
- package/lib/utils/split.ts +279 -0
- package/lib/utils/tokenize.ts +120 -0
- package/lib/utils/wof-json.ts +247 -0
- package/out/adapters/ban/adapter.d.ts.map +1 -0
- package/out/adapters/ban/adapter.js +149 -0
- package/out/adapters/ban/adapter.js.map +1 -0
- package/out/adapters/ban/street-decompose.d.ts.map +1 -0
- package/out/adapters/ban/street-decompose.js +51 -0
- package/out/adapters/ban/street-decompose.js.map +1 -0
- package/out/adapters/fcc-bdc/adapter.d.ts.map +1 -0
- package/out/adapters/fcc-bdc/adapter.js +144 -0
- package/out/adapters/fcc-bdc/adapter.js.map +1 -0
- package/out/adapters/geonames/adapter.d.ts.map +1 -0
- package/out/adapters/geonames/adapter.js +176 -0
- package/out/adapters/geonames/adapter.js.map +1 -0
- package/out/adapters/geonames-postal/adapter.d.ts +54 -0
- package/out/adapters/geonames-postal/adapter.d.ts.map +1 -0
- package/out/adapters/geonames-postal/adapter.js +107 -0
- package/out/adapters/geonames-postal/adapter.js.map +1 -0
- package/out/adapters/gnaf/adapter.d.ts +49 -0
- package/out/adapters/gnaf/adapter.d.ts.map +1 -0
- package/out/adapters/gnaf/adapter.js +124 -0
- package/out/adapters/gnaf/adapter.js.map +1 -0
- package/out/adapters/gnaf/assemble.d.ts +65 -0
- package/out/adapters/gnaf/assemble.d.ts.map +1 -0
- package/out/adapters/gnaf/assemble.js +154 -0
- package/out/adapters/gnaf/assemble.js.map +1 -0
- package/out/adapters/index.d.ts +51 -0
- package/out/adapters/index.d.ts.map +1 -0
- package/out/adapters/index.js +95 -0
- package/out/adapters/index.js.map +1 -0
- package/out/adapters/openaddresses/adapter.d.ts.map +1 -0
- package/out/adapters/openaddresses/adapter.js +186 -0
- package/out/adapters/openaddresses/adapter.js.map +1 -0
- package/out/adapters/overture/adapter.d.ts +48 -0
- package/out/adapters/overture/adapter.d.ts.map +1 -0
- package/out/adapters/overture/adapter.js +125 -0
- package/out/adapters/overture/adapter.js.map +1 -0
- package/out/adapters/state-hi-schools/adapter.d.ts +46 -0
- package/out/adapters/state-hi-schools/adapter.d.ts.map +1 -0
- package/out/adapters/state-hi-schools/adapter.js +145 -0
- package/out/adapters/state-hi-schools/adapter.js.map +1 -0
- package/out/adapters/state-hi-schools/workbook.d.ts +29 -0
- package/out/adapters/state-hi-schools/workbook.d.ts.map +1 -0
- package/out/adapters/state-hi-schools/workbook.js +20 -0
- package/out/adapters/state-hi-schools/workbook.js.map +1 -0
- package/out/adapters/state-ia-contractors/adapter.d.ts.map +1 -0
- package/out/adapters/state-ia-contractors/adapter.js.map +1 -0
- package/out/adapters/state-ny-notaries/adapter.d.ts.map +1 -0
- package/out/adapters/state-ny-notaries/adapter.js.map +1 -0
- package/out/adapters/state-tx-notaries/adapter.d.ts.map +1 -0
- package/out/adapters/state-tx-notaries/adapter.js.map +1 -0
- package/out/adapters/synth-po-box/adapter.d.ts.map +1 -0
- package/out/adapters/synth-po-box/adapter.js +140 -0
- package/out/adapters/synth-po-box/adapter.js.map +1 -0
- package/out/adapters/tiger/adapter.d.ts +55 -0
- package/out/adapters/tiger/adapter.d.ts.map +1 -0
- package/out/adapters/tiger/adapter.js +201 -0
- package/out/adapters/tiger/adapter.js.map +1 -0
- package/out/adapters/tiger/street-decompose.d.ts.map +1 -0
- package/out/adapters/tiger/street-decompose.js +68 -0
- package/out/adapters/tiger/street-decompose.js.map +1 -0
- package/out/adapters/usgov-hrsa-fqhc/adapter.d.ts +56 -0
- package/out/adapters/usgov-hrsa-fqhc/adapter.d.ts.map +1 -0
- package/out/adapters/usgov-hrsa-fqhc/adapter.js +132 -0
- package/out/adapters/usgov-hrsa-fqhc/adapter.js.map +1 -0
- package/out/adapters/usgov-imls-pls/adapter.d.ts +36 -0
- package/out/adapters/usgov-imls-pls/adapter.d.ts.map +1 -0
- package/out/adapters/usgov-imls-pls/adapter.js +114 -0
- package/out/adapters/usgov-imls-pls/adapter.js.map +1 -0
- package/out/adapters/usgov-irs-bmf/adapter.d.ts +37 -0
- package/out/adapters/usgov-irs-bmf/adapter.d.ts.map +1 -0
- package/out/adapters/usgov-irs-bmf/adapter.js +120 -0
- package/out/adapters/usgov-irs-bmf/adapter.js.map +1 -0
- package/out/adapters/usgov-nad/adapter.d.ts +48 -0
- package/out/adapters/usgov-nad/adapter.d.ts.map +1 -0
- package/out/adapters/usgov-nad/adapter.js +249 -0
- package/out/adapters/usgov-nad/adapter.js.map +1 -0
- package/out/adapters/usgov-nppes/adapter.d.ts.map +1 -0
- package/out/adapters/usgov-nppes/adapter.js +118 -0
- package/out/adapters/usgov-nppes/adapter.js.map +1 -0
- package/out/adapters/usgov-samhsa-treatment-locator/adapter.d.ts.map +1 -0
- package/out/adapters/usgov-samhsa-treatment-locator/adapter.js +154 -0
- package/out/adapters/usgov-samhsa-treatment-locator/adapter.js.map +1 -0
- package/out/adapters/utils/index.d.ts +191 -0
- package/out/adapters/utils/index.d.ts.map +1 -0
- package/out/adapters/utils/index.js +206 -0
- package/out/adapters/utils/index.js.map +1 -0
- package/out/adapters/wof-admin-jp/adapter.d.ts.map +1 -0
- package/out/adapters/wof-admin-jp/adapter.js +161 -0
- package/out/adapters/wof-admin-jp/adapter.js.map +1 -0
- package/out/adapters/wof-admin-json/adapter.d.ts +79 -0
- package/out/adapters/wof-admin-json/adapter.d.ts.map +1 -0
- package/out/adapters/wof-admin-json/adapter.js +197 -0
- package/out/adapters/wof-admin-json/adapter.js.map +1 -0
- package/out/adapters/wof-json-rows.d.ts +62 -0
- package/out/adapters/wof-json-rows.d.ts.map +1 -0
- package/out/adapters/wof-json-rows.js +83 -0
- package/out/adapters/wof-json-rows.js.map +1 -0
- package/out/adapters/wof-postalcode-json/adapter.d.ts +65 -0
- package/out/adapters/wof-postalcode-json/adapter.d.ts.map +1 -0
- package/out/adapters/wof-postalcode-json/adapter.js +151 -0
- package/out/adapters/wof-postalcode-json/adapter.js.map +1 -0
- package/out/build.d.ts +132 -0
- package/out/build.d.ts.map +1 -0
- package/out/build.js +245 -0
- package/out/build.js.map +1 -0
- package/out/codex/us-fips-state.d.ts.map +1 -0
- package/out/codex/us-fips-state.js.map +1 -0
- package/out/env.d.ts +46 -0
- package/out/env.d.ts.map +1 -0
- package/out/env.js +46 -0
- package/out/env.js.map +1 -0
- package/out/index.d.ts +13 -0
- package/out/index.d.ts.map +1 -0
- package/out/index.js +13 -0
- package/out/index.js.map +1 -0
- package/out/name-prone-us-suffixes.d.ts +12 -0
- package/out/name-prone-us-suffixes.d.ts.map +1 -0
- package/out/name-prone-us-suffixes.js +12 -0
- package/out/name-prone-us-suffixes.js.map +1 -0
- package/out/recipes/anchor-absorption.d.ts +20 -0
- package/out/recipes/anchor-absorption.d.ts.map +1 -0
- package/out/recipes/anchor-absorption.js +81 -0
- package/out/recipes/anchor-absorption.js.map +1 -0
- package/out/recipes/bare-country.d.ts +25 -0
- package/out/recipes/bare-country.d.ts.map +1 -0
- package/out/recipes/bare-country.js +87 -0
- package/out/recipes/bare-country.js.map +1 -0
- package/out/recipes/boundary-stress.d.ts +20 -0
- package/out/recipes/boundary-stress.d.ts.map +1 -0
- package/out/recipes/boundary-stress.js +83 -0
- package/out/recipes/boundary-stress.js.map +1 -0
- package/out/recipes/cn-organizational-units.d.ts +36 -0
- package/out/recipes/cn-organizational-units.d.ts.map +1 -0
- package/out/recipes/cn-organizational-units.js +159 -0
- package/out/recipes/cn-organizational-units.js.map +1 -0
- package/out/recipes/country-balanced.d.ts +32 -0
- package/out/recipes/country-balanced.d.ts.map +1 -0
- package/out/recipes/country-balanced.js +393 -0
- package/out/recipes/country-balanced.js.map +1 -0
- package/out/recipes/cz-pcfirst-preposition.d.ts +22 -0
- package/out/recipes/cz-pcfirst-preposition.d.ts.map +1 -0
- package/out/recipes/cz-pcfirst-preposition.js +100 -0
- package/out/recipes/cz-pcfirst-preposition.js.map +1 -0
- package/out/recipes/fr-admin-split.d.ts +38 -0
- package/out/recipes/fr-admin-split.d.ts.map +1 -0
- package/out/recipes/fr-admin-split.js +192 -0
- package/out/recipes/fr-admin-split.js.map +1 -0
- package/out/recipes/fr-bare-street.d.ts +28 -0
- package/out/recipes/fr-bare-street.d.ts.map +1 -0
- package/out/recipes/fr-bare-street.js +143 -0
- package/out/recipes/fr-bare-street.js.map +1 -0
- package/out/recipes/fr-fragment.d.ts +77 -0
- package/out/recipes/fr-fragment.d.ts.map +1 -0
- package/out/recipes/fr-fragment.js +276 -0
- package/out/recipes/fr-fragment.js.map +1 -0
- package/out/recipes/fr-lieudit.d.ts +32 -0
- package/out/recipes/fr-lieudit.d.ts.map +1 -0
- package/out/recipes/fr-lieudit.js +227 -0
- package/out/recipes/fr-lieudit.js.map +1 -0
- package/out/recipes/fr-order.d.ts +38 -0
- package/out/recipes/fr-order.d.ts.map +1 -0
- package/out/recipes/fr-order.js +197 -0
- package/out/recipes/fr-order.js.map +1 -0
- package/out/recipes/german.d.ts +44 -0
- package/out/recipes/german.d.ts.map +1 -0
- package/out/recipes/german.js +288 -0
- package/out/recipes/german.js.map +1 -0
- package/out/recipes/house-venue.d.ts +18 -0
- package/out/recipes/house-venue.d.ts.map +1 -0
- package/out/recipes/house-venue.js +73 -0
- package/out/recipes/house-venue.js.map +1 -0
- package/out/recipes/index.d.ts +24 -0
- package/out/recipes/index.d.ts.map +1 -0
- package/out/recipes/index.js +91 -0
- package/out/recipes/index.js.map +1 -0
- package/out/recipes/intersection.d.ts +42 -0
- package/out/recipes/intersection.d.ts.map +1 -0
- package/out/recipes/intersection.js +452 -0
- package/out/recipes/intersection.js.map +1 -0
- package/out/recipes/locale.d.ts +136 -0
- package/out/recipes/locale.d.ts.map +1 -0
- package/out/recipes/locale.js +460 -0
- package/out/recipes/locale.js.map +1 -0
- package/out/recipes/nl-postcode.d.ts +26 -0
- package/out/recipes/nl-postcode.d.ts.map +1 -0
- package/out/recipes/nl-postcode.js +102 -0
- package/out/recipes/nl-postcode.js.map +1 -0
- package/out/recipes/no-fragment.d.ts +43 -0
- package/out/recipes/no-fragment.d.ts.map +1 -0
- package/out/recipes/no-fragment.js +182 -0
- package/out/recipes/no-fragment.js.map +1 -0
- package/out/recipes/no-street-led.d.ts +24 -0
- package/out/recipes/no-street-led.d.ts.map +1 -0
- package/out/recipes/no-street-led.js +113 -0
- package/out/recipes/no-street-led.js.map +1 -0
- package/out/recipes/no-street.d.ts +18 -0
- package/out/recipes/no-street.d.ts.map +1 -0
- package/out/recipes/no-street.js +71 -0
- package/out/recipes/no-street.js.map +1 -0
- package/out/recipes/po-box-cedex.d.ts +38 -0
- package/out/recipes/po-box-cedex.d.ts.map +1 -0
- package/out/recipes/po-box-cedex.js +767 -0
- package/out/recipes/po-box-cedex.js.map +1 -0
- package/out/recipes/po-box.d.ts +17 -0
- package/out/recipes/po-box.d.ts.map +1 -0
- package/out/recipes/po-box.js +100 -0
- package/out/recipes/po-box.js.map +1 -0
- package/out/recipes/reviewed-postcode-tail.d.ts +47 -0
- package/out/recipes/reviewed-postcode-tail.d.ts.map +1 -0
- package/out/recipes/reviewed-postcode-tail.js +125 -0
- package/out/recipes/reviewed-postcode-tail.js.map +1 -0
- package/out/recipes/scaffold.d.ts +324 -0
- package/out/recipes/scaffold.d.ts.map +1 -0
- package/out/recipes/scaffold.js +186 -0
- package/out/recipes/scaffold.js.map +1 -0
- package/out/recipes/si-bare-village.d.ts +32 -0
- package/out/recipes/si-bare-village.d.ts.map +1 -0
- package/out/recipes/si-bare-village.js +88 -0
- package/out/recipes/si-bare-village.js.map +1 -0
- package/out/recipes/street-affix.d.ts +89 -0
- package/out/recipes/street-affix.d.ts.map +1 -0
- package/out/recipes/street-affix.js +563 -0
- package/out/recipes/street-affix.js.map +1 -0
- package/out/recipes/street-bare.d.ts +25 -0
- package/out/recipes/street-bare.d.ts.map +1 -0
- package/out/recipes/street-bare.js +86 -0
- package/out/recipes/street-bare.js.map +1 -0
- package/out/recipes/street.d.ts +15 -0
- package/out/recipes/street.d.ts.map +1 -0
- package/out/recipes/street.js +69 -0
- package/out/recipes/street.js.map +1 -0
- package/out/recipes/sub-venue-sources.d.ts +229 -0
- package/out/recipes/sub-venue-sources.d.ts.map +1 -0
- package/out/recipes/sub-venue-sources.js +653 -0
- package/out/recipes/sub-venue-sources.js.map +1 -0
- package/out/recipes/sub-venue.d.ts +269 -0
- package/out/recipes/sub-venue.d.ts.map +1 -0
- package/out/recipes/sub-venue.js +725 -0
- package/out/recipes/sub-venue.js.map +1 -0
- package/out/recipes/trailing-region.d.ts +74 -0
- package/out/recipes/trailing-region.d.ts.map +1 -0
- package/out/recipes/trailing-region.js +157 -0
- package/out/recipes/trailing-region.js.map +1 -0
- package/out/recipes/unit.d.ts +29 -0
- package/out/recipes/unit.d.ts.map +1 -0
- package/out/recipes/unit.js +230 -0
- package/out/recipes/unit.js.map +1 -0
- package/out/runner.d.ts +130 -0
- package/out/runner.d.ts.map +1 -0
- package/out/runner.js +185 -0
- package/out/runner.js.map +1 -0
- package/out/synthesizers/anchor-absorption.d.ts +63 -0
- package/out/synthesizers/anchor-absorption.d.ts.map +1 -0
- package/out/synthesizers/anchor-absorption.js +221 -0
- package/out/synthesizers/anchor-absorption.js.map +1 -0
- package/out/synthesizers/boundary-stress.d.ts +78 -0
- package/out/synthesizers/boundary-stress.d.ts.map +1 -0
- package/out/synthesizers/boundary-stress.js +455 -0
- package/out/synthesizers/boundary-stress.js.map +1 -0
- package/out/synthesizers/german.d.ts +115 -0
- package/out/synthesizers/german.d.ts.map +1 -0
- package/out/synthesizers/german.js +149 -0
- package/out/synthesizers/german.js.map +1 -0
- package/out/synthesizers/house-venue.d.ts +56 -0
- package/out/synthesizers/house-venue.d.ts.map +1 -0
- package/out/synthesizers/house-venue.js +247 -0
- package/out/synthesizers/house-venue.js.map +1 -0
- package/out/synthesizers/intersection.d.ts +52 -0
- package/out/synthesizers/intersection.d.ts.map +1 -0
- package/out/synthesizers/intersection.js +175 -0
- package/out/synthesizers/intersection.js.map +1 -0
- package/out/synthesizers/no-street.d.ts +72 -0
- package/out/synthesizers/no-street.d.ts.map +1 -0
- package/out/synthesizers/no-street.js +262 -0
- package/out/synthesizers/no-street.js.map +1 -0
- package/out/synthesizers/po-box.d.ts +108 -0
- package/out/synthesizers/po-box.d.ts.map +1 -0
- package/out/synthesizers/po-box.js +250 -0
- package/out/synthesizers/po-box.js.map +1 -0
- package/out/synthesizers/street.d.ts +54 -0
- package/out/synthesizers/street.d.ts.map +1 -0
- package/out/synthesizers/street.js +218 -0
- package/out/synthesizers/street.js.map +1 -0
- package/out/synthesizers/utils.d.ts +241 -0
- package/out/synthesizers/utils.d.ts.map +1 -0
- package/out/synthesizers/utils.js +776 -0
- package/out/synthesizers/utils.js.map +1 -0
- package/out/test-kit/corpus-recipe.d.ts +59 -0
- package/out/test-kit/corpus-recipe.d.ts.map +1 -0
- package/out/test-kit/corpus-recipe.js +39 -0
- package/out/test-kit/corpus-recipe.js.map +1 -0
- package/out/test-kit/index.d.ts +48 -0
- package/out/test-kit/index.d.ts.map +1 -0
- package/out/test-kit/index.js +78 -0
- package/out/test-kit/index.js.map +1 -0
- package/out/tools/align-slice.d.ts +7 -0
- package/out/tools/align-slice.d.ts.map +1 -0
- package/out/tools/align-slice.js +55 -0
- package/out/tools/align-slice.js.map +1 -0
- package/out/tools/audit.d.ts +29 -0
- package/out/tools/audit.d.ts.map +1 -0
- package/out/tools/audit.js +282 -0
- package/out/tools/audit.js.map +1 -0
- package/out/tools/corpus-stats.d.ts +44 -0
- package/out/tools/corpus-stats.d.ts.map +1 -0
- package/out/tools/corpus-stats.js +102 -0
- package/out/tools/corpus-stats.js.map +1 -0
- package/out/tools/fetch/ban.d.ts +31 -0
- package/out/tools/fetch/ban.d.ts.map +1 -0
- package/out/tools/fetch/ban.js +239 -0
- package/out/tools/fetch/ban.js.map +1 -0
- package/out/tools/fetch/download.d.ts +117 -0
- package/out/tools/fetch/download.d.ts.map +1 -0
- package/out/tools/fetch/download.js +160 -0
- package/out/tools/fetch/download.js.map +1 -0
- package/out/tools/fetch/geonames-dump.d.ts +91 -0
- package/out/tools/fetch/geonames-dump.d.ts.map +1 -0
- package/out/tools/fetch/geonames-dump.js +171 -0
- package/out/tools/fetch/geonames-dump.js.map +1 -0
- package/out/tools/fetch/geonames-postal.d.ts +76 -0
- package/out/tools/fetch/geonames-postal.d.ts.map +1 -0
- package/out/tools/fetch/geonames-postal.js +126 -0
- package/out/tools/fetch/geonames-postal.js.map +1 -0
- package/out/tools/fetch/hrsa.d.ts +18 -0
- package/out/tools/fetch/hrsa.d.ts.map +1 -0
- package/out/tools/fetch/hrsa.js +46 -0
- package/out/tools/fetch/hrsa.js.map +1 -0
- package/out/tools/fetch/imls-pls.d.ts +25 -0
- package/out/tools/fetch/imls-pls.d.ts.map +1 -0
- package/out/tools/fetch/imls-pls.js +108 -0
- package/out/tools/fetch/imls-pls.js.map +1 -0
- package/out/tools/fetch/index.d.ts +144 -0
- package/out/tools/fetch/index.d.ts.map +1 -0
- package/out/tools/fetch/index.js +143 -0
- package/out/tools/fetch/index.js.map +1 -0
- package/out/tools/fetch/nad.d.ts +70 -0
- package/out/tools/fetch/nad.d.ts.map +1 -0
- package/out/tools/fetch/nad.js +234 -0
- package/out/tools/fetch/nad.js.map +1 -0
- package/out/tools/fetch/nppes.d.ts +26 -0
- package/out/tools/fetch/nppes.d.ts.map +1 -0
- package/out/tools/fetch/nppes.js +126 -0
- package/out/tools/fetch/nppes.js.map +1 -0
- package/out/tools/fetch/openaddresses.d.ts +9 -0
- package/out/tools/fetch/openaddresses.d.ts.map +1 -0
- package/out/tools/fetch/openaddresses.js +279 -0
- package/out/tools/fetch/openaddresses.js.map +1 -0
- package/out/tools/fetch/ourairports.d.ts +45 -0
- package/out/tools/fetch/ourairports.d.ts.map +1 -0
- package/out/tools/fetch/ourairports.js +128 -0
- package/out/tools/fetch/ourairports.js.map +1 -0
- package/out/tools/fetch/ppd.d.ts.map +1 -0
- package/out/tools/fetch/ppd.js +85 -0
- package/out/tools/fetch/ppd.js.map +1 -0
- package/out/tools/fetch/state-hi-schools.d.ts +26 -0
- package/out/tools/fetch/state-hi-schools.d.ts.map +1 -0
- package/out/tools/fetch/state-hi-schools.js +112 -0
- package/out/tools/fetch/state-hi-schools.js.map +1 -0
- package/out/tools/fetch/state-sources.d.ts +26 -0
- package/out/tools/fetch/state-sources.d.ts.map +1 -0
- package/out/tools/fetch/state-sources.js +111 -0
- package/out/tools/fetch/state-sources.js.map +1 -0
- package/out/tools/fetch/tiger-full.d.ts +22 -0
- package/out/tools/fetch/tiger-full.d.ts.map +1 -0
- package/out/tools/fetch/tiger-full.js +217 -0
- package/out/tools/fetch/tiger-full.js.map +1 -0
- package/out/tools/fetch/wikidata-subvenue.d.ts +171 -0
- package/out/tools/fetch/wikidata-subvenue.d.ts.map +1 -0
- package/out/tools/fetch/wikidata-subvenue.js +274 -0
- package/out/tools/fetch/wikidata-subvenue.js.map +1 -0
- package/out/tools/gb-title-case.d.ts.map +1 -0
- package/out/tools/gb-title-case.js.map +1 -0
- package/out/tools/golden-expand.d.ts +90 -0
- package/out/tools/golden-expand.d.ts.map +1 -0
- package/out/tools/golden-expand.js +406 -0
- package/out/tools/golden-expand.js.map +1 -0
- package/out/tools/golden-promote.d.ts.map +1 -0
- package/out/tools/golden-promote.js +240 -0
- package/out/tools/golden-promote.js.map +1 -0
- package/out/tools/golden-relabel-street.d.ts +196 -0
- package/out/tools/golden-relabel-street.d.ts.map +1 -0
- package/out/tools/golden-relabel-street.js +440 -0
- package/out/tools/golden-relabel-street.js.map +1 -0
- package/out/tools/index.d.ts +28 -0
- package/out/tools/index.d.ts.map +1 -0
- package/out/tools/index.js +28 -0
- package/out/tools/index.js.map +1 -0
- package/out/tools/ingest-csv.d.ts.map +1 -0
- package/out/tools/ingest-csv.js +265 -0
- package/out/tools/ingest-csv.js.map +1 -0
- package/out/tools/jsonl-to-parquet.d.ts +61 -0
- package/out/tools/jsonl-to-parquet.d.ts.map +1 -0
- package/out/tools/jsonl-to-parquet.js +150 -0
- package/out/tools/jsonl-to-parquet.js.map +1 -0
- package/out/tools/lint-slice-vocab.d.ts +90 -0
- package/out/tools/lint-slice-vocab.d.ts.map +1 -0
- package/out/tools/lint-slice-vocab.js +267 -0
- package/out/tools/lint-slice-vocab.js.map +1 -0
- package/out/tools/lint-slice.d.ts +93 -0
- package/out/tools/lint-slice.d.ts.map +1 -0
- package/out/tools/lint-slice.js +322 -0
- package/out/tools/lint-slice.js.map +1 -0
- package/out/tools/overlay-manifest.d.ts +39 -0
- package/out/tools/overlay-manifest.d.ts.map +1 -0
- package/out/tools/overlay-manifest.js +90 -0
- package/out/tools/overlay-manifest.js.map +1 -0
- package/out/tools/overture-subvenue.d.ts +112 -0
- package/out/tools/overture-subvenue.d.ts.map +1 -0
- package/out/tools/overture-subvenue.js +139 -0
- package/out/tools/overture-subvenue.js.map +1 -0
- package/out/tools/postcode-triples.d.ts +174 -0
- package/out/tools/postcode-triples.d.ts.map +1 -0
- package/out/tools/postcode-triples.js +292 -0
- package/out/tools/postcode-triples.js.map +1 -0
- package/out/tools/slice-kryptonite.d.ts +32 -0
- package/out/tools/slice-kryptonite.d.ts.map +1 -0
- package/out/tools/slice-kryptonite.js +99 -0
- package/out/tools/slice-kryptonite.js.map +1 -0
- package/out/tools/slice-translit.d.ts +51 -0
- package/out/tools/slice-translit.d.ts.map +1 -0
- package/out/tools/slice-translit.js +166 -0
- package/out/tools/slice-translit.js.map +1 -0
- package/out/tools/sub-venue/harvest.d.ts +100 -0
- package/out/tools/sub-venue/harvest.d.ts.map +1 -0
- package/out/tools/sub-venue/harvest.js +168 -0
- package/out/tools/sub-venue/harvest.js.map +1 -0
- package/out/tools/sub-venue/head-nouns.d.ts +50 -0
- package/out/tools/sub-venue/head-nouns.d.ts.map +1 -0
- package/out/tools/sub-venue/head-nouns.js +213 -0
- package/out/tools/sub-venue/head-nouns.js.map +1 -0
- package/out/tools/sub-venue/surfaces.d.ts +45 -0
- package/out/tools/sub-venue/surfaces.d.ts.map +1 -0
- package/out/tools/sub-venue/surfaces.js +77 -0
- package/out/tools/sub-venue/surfaces.js.map +1 -0
- package/out/tools/sub-venue/table.d.ts +229 -0
- package/out/tools/sub-venue/table.d.ts.map +1 -0
- package/out/tools/sub-venue/table.js +108 -0
- package/out/tools/sub-venue/table.js.map +1 -0
- package/out/tools/sub-venue/wikidata.d.ts +26 -0
- package/out/tools/sub-venue/wikidata.d.ts.map +1 -0
- package/out/tools/sub-venue/wikidata.js +62 -0
- package/out/tools/sub-venue/wikidata.js.map +1 -0
- package/out/tools/sub-venue-lexicon.d.ts +156 -0
- package/out/tools/sub-venue-lexicon.d.ts.map +1 -0
- package/out/tools/sub-venue-lexicon.js +283 -0
- package/out/tools/sub-venue-lexicon.js.map +1 -0
- package/out/tools/sub-venue-promotions.d.ts +94 -0
- package/out/tools/sub-venue-promotions.d.ts.map +1 -0
- package/out/tools/sub-venue-promotions.js +266 -0
- package/out/tools/sub-venue-promotions.js.map +1 -0
- package/out/types.d.ts +221 -0
- package/out/types.d.ts.map +1 -0
- package/out/types.js.map +1 -0
- package/out/utils/align.d.ts +93 -0
- package/out/utils/align.d.ts.map +1 -0
- package/out/utils/align.js +234 -0
- package/out/utils/align.js.map +1 -0
- package/out/utils/golden.d.ts +85 -0
- package/out/utils/golden.d.ts.map +1 -0
- package/out/utils/golden.js +127 -0
- package/out/utils/golden.js.map +1 -0
- package/out/utils/index.d.ts +14 -0
- package/out/utils/index.d.ts.map +1 -0
- package/out/utils/index.js +14 -0
- package/out/utils/index.js.map +1 -0
- package/out/utils/license.d.ts.map +1 -0
- package/out/utils/license.js.map +1 -0
- package/out/utils/parquet.d.ts +178 -0
- package/out/utils/parquet.d.ts.map +1 -0
- package/out/utils/parquet.js +371 -0
- package/out/utils/parquet.js.map +1 -0
- package/out/utils/slice-stats.d.ts +52 -0
- package/out/utils/slice-stats.d.ts.map +1 -0
- package/out/utils/slice-stats.js +58 -0
- package/out/utils/slice-stats.js.map +1 -0
- package/out/utils/split.d.ts +128 -0
- package/out/utils/split.d.ts.map +1 -0
- package/out/utils/split.js +204 -0
- package/out/utils/split.js.map +1 -0
- package/out/utils/tokenize.d.ts +64 -0
- package/out/utils/tokenize.d.ts.map +1 -0
- package/out/utils/tokenize.js +82 -0
- package/out/utils/tokenize.js.map +1 -0
- package/out/utils/wof-json.d.ts +113 -0
- package/out/utils/wof-json.d.ts.map +1 -0
- package/out/utils/wof-json.js +173 -0
- package/out/utils/wof-json.js.map +1 -0
- package/package.json +259 -194
- package/out/src/adapters/ban/adapter.d.ts.map +0 -1
- package/out/src/adapters/ban/adapter.js +0 -149
- package/out/src/adapters/ban/adapter.js.map +0 -1
- package/out/src/adapters/ban/street-decompose.d.ts.map +0 -1
- package/out/src/adapters/ban/street-decompose.js +0 -75
- package/out/src/adapters/ban/street-decompose.js.map +0 -1
- package/out/src/adapters/fcc-bdc/adapter.d.ts.map +0 -1
- package/out/src/adapters/fcc-bdc/adapter.js +0 -144
- package/out/src/adapters/fcc-bdc/adapter.js.map +0 -1
- package/out/src/adapters/geonames/adapter.d.ts.map +0 -1
- package/out/src/adapters/geonames/adapter.js +0 -176
- package/out/src/adapters/geonames/adapter.js.map +0 -1
- package/out/src/adapters/geonames-postal/adapter.d.ts +0 -41
- package/out/src/adapters/geonames-postal/adapter.d.ts.map +0 -1
- package/out/src/adapters/geonames-postal/adapter.js +0 -104
- package/out/src/adapters/geonames-postal/adapter.js.map +0 -1
- package/out/src/adapters/gnaf/adapter.d.ts +0 -49
- package/out/src/adapters/gnaf/adapter.d.ts.map +0 -1
- package/out/src/adapters/gnaf/adapter.js +0 -124
- package/out/src/adapters/gnaf/adapter.js.map +0 -1
- package/out/src/adapters/gnaf/assemble.d.ts +0 -65
- package/out/src/adapters/gnaf/assemble.d.ts.map +0 -1
- package/out/src/adapters/gnaf/assemble.js +0 -154
- package/out/src/adapters/gnaf/assemble.js.map +0 -1
- package/out/src/adapters/index.d.ts +0 -51
- package/out/src/adapters/index.d.ts.map +0 -1
- package/out/src/adapters/index.js +0 -95
- package/out/src/adapters/index.js.map +0 -1
- package/out/src/adapters/openaddresses/adapter.d.ts.map +0 -1
- package/out/src/adapters/openaddresses/adapter.js +0 -186
- package/out/src/adapters/openaddresses/adapter.js.map +0 -1
- package/out/src/adapters/overture/adapter.d.ts +0 -48
- package/out/src/adapters/overture/adapter.d.ts.map +0 -1
- package/out/src/adapters/overture/adapter.js +0 -125
- package/out/src/adapters/overture/adapter.js.map +0 -1
- package/out/src/adapters/state-hi-schools/adapter.d.ts +0 -47
- package/out/src/adapters/state-hi-schools/adapter.d.ts.map +0 -1
- package/out/src/adapters/state-hi-schools/adapter.js +0 -133
- package/out/src/adapters/state-hi-schools/adapter.js.map +0 -1
- package/out/src/adapters/state-ia-contractors/adapter.d.ts.map +0 -1
- package/out/src/adapters/state-ia-contractors/adapter.js.map +0 -1
- package/out/src/adapters/state-ny-notaries/adapter.d.ts.map +0 -1
- package/out/src/adapters/state-ny-notaries/adapter.js.map +0 -1
- package/out/src/adapters/state-tx-notaries/adapter.d.ts.map +0 -1
- package/out/src/adapters/state-tx-notaries/adapter.js.map +0 -1
- package/out/src/adapters/synth-po-box/adapter.d.ts.map +0 -1
- package/out/src/adapters/synth-po-box/adapter.js +0 -140
- package/out/src/adapters/synth-po-box/adapter.js.map +0 -1
- package/out/src/adapters/tiger/adapter.d.ts +0 -55
- package/out/src/adapters/tiger/adapter.d.ts.map +0 -1
- package/out/src/adapters/tiger/adapter.js +0 -201
- package/out/src/adapters/tiger/adapter.js.map +0 -1
- package/out/src/adapters/tiger/street-decompose.d.ts.map +0 -1
- package/out/src/adapters/tiger/street-decompose.js +0 -90
- package/out/src/adapters/tiger/street-decompose.js.map +0 -1
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.d.ts +0 -47
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.d.ts.map +0 -1
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.js +0 -139
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.js.map +0 -1
- package/out/src/adapters/usgov-imls-pls/adapter.d.ts +0 -36
- package/out/src/adapters/usgov-imls-pls/adapter.d.ts.map +0 -1
- package/out/src/adapters/usgov-imls-pls/adapter.js +0 -114
- package/out/src/adapters/usgov-imls-pls/adapter.js.map +0 -1
- package/out/src/adapters/usgov-irs-bmf/adapter.d.ts +0 -37
- package/out/src/adapters/usgov-irs-bmf/adapter.d.ts.map +0 -1
- package/out/src/adapters/usgov-irs-bmf/adapter.js +0 -124
- package/out/src/adapters/usgov-irs-bmf/adapter.js.map +0 -1
- package/out/src/adapters/usgov-nad/adapter.d.ts +0 -48
- package/out/src/adapters/usgov-nad/adapter.d.ts.map +0 -1
- package/out/src/adapters/usgov-nad/adapter.js +0 -252
- package/out/src/adapters/usgov-nad/adapter.js.map +0 -1
- package/out/src/adapters/usgov-nppes/adapter.d.ts.map +0 -1
- package/out/src/adapters/usgov-nppes/adapter.js +0 -116
- package/out/src/adapters/usgov-nppes/adapter.js.map +0 -1
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.d.ts.map +0 -1
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.js +0 -156
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.js.map +0 -1
- package/out/src/adapters/utils/index.d.ts +0 -150
- package/out/src/adapters/utils/index.d.ts.map +0 -1
- package/out/src/adapters/utils/index.js +0 -159
- package/out/src/adapters/utils/index.js.map +0 -1
- package/out/src/adapters/wof-admin-jp/adapter.d.ts.map +0 -1
- package/out/src/adapters/wof-admin-jp/adapter.js +0 -164
- package/out/src/adapters/wof-admin-jp/adapter.js.map +0 -1
- package/out/src/adapters/wof-admin-json/adapter.d.ts +0 -82
- package/out/src/adapters/wof-admin-json/adapter.d.ts.map +0 -1
- package/out/src/adapters/wof-admin-json/adapter.js +0 -224
- package/out/src/adapters/wof-admin-json/adapter.js.map +0 -1
- package/out/src/adapters/wof-json-rows.d.ts +0 -32
- package/out/src/adapters/wof-json-rows.d.ts.map +0 -1
- package/out/src/adapters/wof-json-rows.js +0 -45
- package/out/src/adapters/wof-json-rows.js.map +0 -1
- package/out/src/adapters/wof-postalcode-json/adapter.d.ts +0 -65
- package/out/src/adapters/wof-postalcode-json/adapter.d.ts.map +0 -1
- package/out/src/adapters/wof-postalcode-json/adapter.js +0 -160
- package/out/src/adapters/wof-postalcode-json/adapter.js.map +0 -1
- package/out/src/build.d.ts +0 -132
- package/out/src/build.d.ts.map +0 -1
- package/out/src/build.js +0 -251
- package/out/src/build.js.map +0 -1
- package/out/src/codex/us-fips-state.d.ts.map +0 -1
- package/out/src/codex/us-fips-state.js.map +0 -1
- package/out/src/index.d.ts +0 -14
- package/out/src/index.d.ts.map +0 -1
- package/out/src/index.js +0 -14
- package/out/src/index.js.map +0 -1
- package/out/src/name-prone-us-suffixes.d.ts +0 -12
- package/out/src/name-prone-us-suffixes.d.ts.map +0 -1
- package/out/src/name-prone-us-suffixes.js +0 -12
- package/out/src/name-prone-us-suffixes.js.map +0 -1
- package/out/src/parquet-wrapper/index.d.ts +0 -9
- package/out/src/parquet-wrapper/index.d.ts.map +0 -1
- package/out/src/parquet-wrapper/index.js +0 -9
- package/out/src/parquet-wrapper/index.js.map +0 -1
- package/out/src/parquet-wrapper/reader.d.ts +0 -34
- package/out/src/parquet-wrapper/reader.d.ts.map +0 -1
- package/out/src/parquet-wrapper/reader.js +0 -69
- package/out/src/parquet-wrapper/reader.js.map +0 -1
- package/out/src/parquet-wrapper/schema.d.ts +0 -49
- package/out/src/parquet-wrapper/schema.d.ts.map +0 -1
- package/out/src/parquet-wrapper/schema.js +0 -55
- package/out/src/parquet-wrapper/schema.js.map +0 -1
- package/out/src/parquet-wrapper/writer.d.ts +0 -46
- package/out/src/parquet-wrapper/writer.d.ts.map +0 -1
- package/out/src/parquet-wrapper/writer.js +0 -76
- package/out/src/parquet-wrapper/writer.js.map +0 -1
- package/out/src/runner.d.ts +0 -126
- package/out/src/runner.d.ts.map +0 -1
- package/out/src/runner.js +0 -189
- package/out/src/runner.js.map +0 -1
- package/out/src/shard-recipes/anchor-absorption.d.ts +0 -20
- package/out/src/shard-recipes/anchor-absorption.d.ts.map +0 -1
- package/out/src/shard-recipes/anchor-absorption.js +0 -82
- package/out/src/shard-recipes/anchor-absorption.js.map +0 -1
- package/out/src/shard-recipes/bare-country.d.ts +0 -25
- package/out/src/shard-recipes/bare-country.d.ts.map +0 -1
- package/out/src/shard-recipes/bare-country.js +0 -87
- package/out/src/shard-recipes/bare-country.js.map +0 -1
- package/out/src/shard-recipes/boundary-stress.d.ts +0 -20
- package/out/src/shard-recipes/boundary-stress.d.ts.map +0 -1
- package/out/src/shard-recipes/boundary-stress.js +0 -83
- package/out/src/shard-recipes/boundary-stress.js.map +0 -1
- package/out/src/shard-recipes/country-balanced.d.ts +0 -32
- package/out/src/shard-recipes/country-balanced.d.ts.map +0 -1
- package/out/src/shard-recipes/country-balanced.js +0 -406
- package/out/src/shard-recipes/country-balanced.js.map +0 -1
- package/out/src/shard-recipes/cz-pcfirst-preposition.d.ts +0 -22
- package/out/src/shard-recipes/cz-pcfirst-preposition.d.ts.map +0 -1
- package/out/src/shard-recipes/cz-pcfirst-preposition.js +0 -100
- package/out/src/shard-recipes/cz-pcfirst-preposition.js.map +0 -1
- package/out/src/shard-recipes/fr-admin-split.d.ts +0 -38
- package/out/src/shard-recipes/fr-admin-split.d.ts.map +0 -1
- package/out/src/shard-recipes/fr-admin-split.js +0 -188
- package/out/src/shard-recipes/fr-admin-split.js.map +0 -1
- package/out/src/shard-recipes/fr-bare-street.d.ts +0 -28
- package/out/src/shard-recipes/fr-bare-street.d.ts.map +0 -1
- package/out/src/shard-recipes/fr-bare-street.js +0 -143
- package/out/src/shard-recipes/fr-bare-street.js.map +0 -1
- package/out/src/shard-recipes/fr-fragment.d.ts +0 -77
- package/out/src/shard-recipes/fr-fragment.d.ts.map +0 -1
- package/out/src/shard-recipes/fr-fragment.js +0 -276
- package/out/src/shard-recipes/fr-fragment.js.map +0 -1
- package/out/src/shard-recipes/fr-lieudit.d.ts +0 -32
- package/out/src/shard-recipes/fr-lieudit.d.ts.map +0 -1
- package/out/src/shard-recipes/fr-lieudit.js +0 -226
- package/out/src/shard-recipes/fr-lieudit.js.map +0 -1
- package/out/src/shard-recipes/fr-order.d.ts +0 -38
- package/out/src/shard-recipes/fr-order.d.ts.map +0 -1
- package/out/src/shard-recipes/fr-order.js +0 -208
- package/out/src/shard-recipes/fr-order.js.map +0 -1
- package/out/src/shard-recipes/german.d.ts +0 -23
- package/out/src/shard-recipes/german.d.ts.map +0 -1
- package/out/src/shard-recipes/german.js +0 -132
- package/out/src/shard-recipes/german.js.map +0 -1
- package/out/src/shard-recipes/house-venue.d.ts +0 -18
- package/out/src/shard-recipes/house-venue.d.ts.map +0 -1
- package/out/src/shard-recipes/house-venue.js +0 -74
- package/out/src/shard-recipes/house-venue.js.map +0 -1
- package/out/src/shard-recipes/index.d.ts +0 -24
- package/out/src/shard-recipes/index.d.ts.map +0 -1
- package/out/src/shard-recipes/index.js +0 -89
- package/out/src/shard-recipes/index.js.map +0 -1
- package/out/src/shard-recipes/intersection.d.ts +0 -42
- package/out/src/shard-recipes/intersection.d.ts.map +0 -1
- package/out/src/shard-recipes/intersection.js +0 -460
- package/out/src/shard-recipes/intersection.js.map +0 -1
- package/out/src/shard-recipes/locale.d.ts +0 -135
- package/out/src/shard-recipes/locale.d.ts.map +0 -1
- package/out/src/shard-recipes/locale.js +0 -459
- package/out/src/shard-recipes/locale.js.map +0 -1
- package/out/src/shard-recipes/nl-postcode.d.ts +0 -26
- package/out/src/shard-recipes/nl-postcode.d.ts.map +0 -1
- package/out/src/shard-recipes/nl-postcode.js +0 -102
- package/out/src/shard-recipes/nl-postcode.js.map +0 -1
- package/out/src/shard-recipes/no-fragment.d.ts +0 -43
- package/out/src/shard-recipes/no-fragment.d.ts.map +0 -1
- package/out/src/shard-recipes/no-fragment.js +0 -188
- package/out/src/shard-recipes/no-fragment.js.map +0 -1
- package/out/src/shard-recipes/no-street-led.d.ts +0 -24
- package/out/src/shard-recipes/no-street-led.d.ts.map +0 -1
- package/out/src/shard-recipes/no-street-led.js +0 -121
- package/out/src/shard-recipes/no-street-led.js.map +0 -1
- package/out/src/shard-recipes/no-street.d.ts +0 -18
- package/out/src/shard-recipes/no-street.d.ts.map +0 -1
- package/out/src/shard-recipes/no-street.js +0 -72
- package/out/src/shard-recipes/no-street.js.map +0 -1
- package/out/src/shard-recipes/po-box-cedex.d.ts +0 -38
- package/out/src/shard-recipes/po-box-cedex.d.ts.map +0 -1
- package/out/src/shard-recipes/po-box-cedex.js +0 -726
- package/out/src/shard-recipes/po-box-cedex.js.map +0 -1
- package/out/src/shard-recipes/po-box.d.ts +0 -17
- package/out/src/shard-recipes/po-box.d.ts.map +0 -1
- package/out/src/shard-recipes/po-box.js +0 -101
- package/out/src/shard-recipes/po-box.js.map +0 -1
- package/out/src/shard-recipes/reviewed-postcode-tail.d.ts +0 -47
- package/out/src/shard-recipes/reviewed-postcode-tail.d.ts.map +0 -1
- package/out/src/shard-recipes/reviewed-postcode-tail.js +0 -125
- package/out/src/shard-recipes/reviewed-postcode-tail.js.map +0 -1
- package/out/src/shard-recipes/scaffold.d.ts +0 -246
- package/out/src/shard-recipes/scaffold.d.ts.map +0 -1
- package/out/src/shard-recipes/scaffold.js +0 -137
- package/out/src/shard-recipes/scaffold.js.map +0 -1
- package/out/src/shard-recipes/si-bare-village.d.ts +0 -32
- package/out/src/shard-recipes/si-bare-village.d.ts.map +0 -1
- package/out/src/shard-recipes/si-bare-village.js +0 -88
- package/out/src/shard-recipes/si-bare-village.js.map +0 -1
- package/out/src/shard-recipes/street-affix.d.ts +0 -89
- package/out/src/shard-recipes/street-affix.d.ts.map +0 -1
- package/out/src/shard-recipes/street-affix.js +0 -578
- package/out/src/shard-recipes/street-affix.js.map +0 -1
- package/out/src/shard-recipes/street-bare.d.ts +0 -25
- package/out/src/shard-recipes/street-bare.d.ts.map +0 -1
- package/out/src/shard-recipes/street-bare.js +0 -86
- package/out/src/shard-recipes/street-bare.js.map +0 -1
- package/out/src/shard-recipes/street.d.ts +0 -15
- package/out/src/shard-recipes/street.d.ts.map +0 -1
- package/out/src/shard-recipes/street.js +0 -69
- package/out/src/shard-recipes/street.js.map +0 -1
- package/out/src/shard-recipes/sub-venue-sources.d.ts +0 -228
- package/out/src/shard-recipes/sub-venue-sources.d.ts.map +0 -1
- package/out/src/shard-recipes/sub-venue-sources.js +0 -664
- package/out/src/shard-recipes/sub-venue-sources.js.map +0 -1
- package/out/src/shard-recipes/sub-venue.d.ts +0 -269
- package/out/src/shard-recipes/sub-venue.d.ts.map +0 -1
- package/out/src/shard-recipes/sub-venue.js +0 -710
- package/out/src/shard-recipes/sub-venue.js.map +0 -1
- package/out/src/shard-recipes/trailing-region.d.ts +0 -74
- package/out/src/shard-recipes/trailing-region.d.ts.map +0 -1
- package/out/src/shard-recipes/trailing-region.js +0 -157
- package/out/src/shard-recipes/trailing-region.js.map +0 -1
- package/out/src/shard-recipes/unit.d.ts +0 -29
- package/out/src/shard-recipes/unit.d.ts.map +0 -1
- package/out/src/shard-recipes/unit.js +0 -243
- package/out/src/shard-recipes/unit.js.map +0 -1
- package/out/src/synthesizers/anchor-absorption.d.ts +0 -63
- package/out/src/synthesizers/anchor-absorption.d.ts.map +0 -1
- package/out/src/synthesizers/anchor-absorption.js +0 -223
- package/out/src/synthesizers/anchor-absorption.js.map +0 -1
- package/out/src/synthesizers/boundary-stress.d.ts +0 -78
- package/out/src/synthesizers/boundary-stress.d.ts.map +0 -1
- package/out/src/synthesizers/boundary-stress.js +0 -455
- package/out/src/synthesizers/boundary-stress.js.map +0 -1
- package/out/src/synthesizers/german.d.ts +0 -107
- package/out/src/synthesizers/german.d.ts.map +0 -1
- package/out/src/synthesizers/german.js +0 -148
- package/out/src/synthesizers/german.js.map +0 -1
- package/out/src/synthesizers/house-venue.d.ts +0 -56
- package/out/src/synthesizers/house-venue.d.ts.map +0 -1
- package/out/src/synthesizers/house-venue.js +0 -267
- package/out/src/synthesizers/house-venue.js.map +0 -1
- package/out/src/synthesizers/intersection.d.ts +0 -52
- package/out/src/synthesizers/intersection.d.ts.map +0 -1
- package/out/src/synthesizers/intersection.js +0 -175
- package/out/src/synthesizers/intersection.js.map +0 -1
- package/out/src/synthesizers/no-street.d.ts +0 -72
- package/out/src/synthesizers/no-street.d.ts.map +0 -1
- package/out/src/synthesizers/no-street.js +0 -282
- package/out/src/synthesizers/no-street.js.map +0 -1
- package/out/src/synthesizers/po-box.d.ts +0 -104
- package/out/src/synthesizers/po-box.d.ts.map +0 -1
- package/out/src/synthesizers/po-box.js +0 -266
- package/out/src/synthesizers/po-box.js.map +0 -1
- package/out/src/synthesizers/street.d.ts +0 -54
- package/out/src/synthesizers/street.d.ts.map +0 -1
- package/out/src/synthesizers/street.js +0 -221
- package/out/src/synthesizers/street.js.map +0 -1
- package/out/src/synthesizers/utils.d.ts +0 -203
- package/out/src/synthesizers/utils.d.ts.map +0 -1
- package/out/src/synthesizers/utils.js +0 -708
- package/out/src/synthesizers/utils.js.map +0 -1
- package/out/src/tools/align-shard.d.ts +0 -8
- package/out/src/tools/align-shard.d.ts.map +0 -1
- package/out/src/tools/align-shard.js +0 -56
- package/out/src/tools/align-shard.js.map +0 -1
- package/out/src/tools/audit.d.ts +0 -28
- package/out/src/tools/audit.d.ts.map +0 -1
- package/out/src/tools/audit.js +0 -285
- package/out/src/tools/audit.js.map +0 -1
- package/out/src/tools/corpus-stats.d.ts +0 -44
- package/out/src/tools/corpus-stats.d.ts.map +0 -1
- package/out/src/tools/corpus-stats.js +0 -138
- package/out/src/tools/corpus-stats.js.map +0 -1
- package/out/src/tools/fetch/ban.d.ts +0 -31
- package/out/src/tools/fetch/ban.d.ts.map +0 -1
- package/out/src/tools/fetch/ban.js +0 -239
- package/out/src/tools/fetch/ban.js.map +0 -1
- package/out/src/tools/fetch/download.d.ts +0 -78
- package/out/src/tools/fetch/download.d.ts.map +0 -1
- package/out/src/tools/fetch/download.js +0 -116
- package/out/src/tools/fetch/download.js.map +0 -1
- package/out/src/tools/fetch/geonames-dump.d.ts +0 -87
- package/out/src/tools/fetch/geonames-dump.d.ts.map +0 -1
- package/out/src/tools/fetch/geonames-dump.js +0 -178
- package/out/src/tools/fetch/geonames-dump.js.map +0 -1
- package/out/src/tools/fetch/geonames-postal.d.ts +0 -76
- package/out/src/tools/fetch/geonames-postal.d.ts.map +0 -1
- package/out/src/tools/fetch/geonames-postal.js +0 -125
- package/out/src/tools/fetch/geonames-postal.js.map +0 -1
- package/out/src/tools/fetch/hrsa.d.ts +0 -18
- package/out/src/tools/fetch/hrsa.d.ts.map +0 -1
- package/out/src/tools/fetch/hrsa.js +0 -46
- package/out/src/tools/fetch/hrsa.js.map +0 -1
- package/out/src/tools/fetch/imls-pls.d.ts +0 -25
- package/out/src/tools/fetch/imls-pls.d.ts.map +0 -1
- package/out/src/tools/fetch/imls-pls.js +0 -108
- package/out/src/tools/fetch/imls-pls.js.map +0 -1
- package/out/src/tools/fetch/index.d.ts +0 -144
- package/out/src/tools/fetch/index.d.ts.map +0 -1
- package/out/src/tools/fetch/index.js +0 -143
- package/out/src/tools/fetch/index.js.map +0 -1
- package/out/src/tools/fetch/nad.d.ts +0 -70
- package/out/src/tools/fetch/nad.d.ts.map +0 -1
- package/out/src/tools/fetch/nad.js +0 -234
- package/out/src/tools/fetch/nad.js.map +0 -1
- package/out/src/tools/fetch/nppes.d.ts +0 -26
- package/out/src/tools/fetch/nppes.d.ts.map +0 -1
- package/out/src/tools/fetch/nppes.js +0 -126
- package/out/src/tools/fetch/nppes.js.map +0 -1
- package/out/src/tools/fetch/openaddresses.d.ts +0 -61
- package/out/src/tools/fetch/openaddresses.d.ts.map +0 -1
- package/out/src/tools/fetch/openaddresses.js +0 -317
- package/out/src/tools/fetch/openaddresses.js.map +0 -1
- package/out/src/tools/fetch/ourairports.d.ts +0 -45
- package/out/src/tools/fetch/ourairports.d.ts.map +0 -1
- package/out/src/tools/fetch/ourairports.js +0 -128
- package/out/src/tools/fetch/ourairports.js.map +0 -1
- package/out/src/tools/fetch/ppd.d.ts.map +0 -1
- package/out/src/tools/fetch/ppd.js +0 -100
- package/out/src/tools/fetch/ppd.js.map +0 -1
- package/out/src/tools/fetch/state-hi-schools.d.ts +0 -25
- package/out/src/tools/fetch/state-hi-schools.d.ts.map +0 -1
- package/out/src/tools/fetch/state-hi-schools.js +0 -162
- package/out/src/tools/fetch/state-hi-schools.js.map +0 -1
- package/out/src/tools/fetch/state-sources.d.ts +0 -26
- package/out/src/tools/fetch/state-sources.d.ts.map +0 -1
- package/out/src/tools/fetch/state-sources.js +0 -111
- package/out/src/tools/fetch/state-sources.js.map +0 -1
- package/out/src/tools/fetch/tiger-full.d.ts +0 -42
- package/out/src/tools/fetch/tiger-full.d.ts.map +0 -1
- package/out/src/tools/fetch/tiger-full.js +0 -248
- package/out/src/tools/fetch/tiger-full.js.map +0 -1
- package/out/src/tools/fetch/wikidata-subvenue.d.ts +0 -171
- package/out/src/tools/fetch/wikidata-subvenue.d.ts.map +0 -1
- package/out/src/tools/fetch/wikidata-subvenue.js +0 -275
- package/out/src/tools/fetch/wikidata-subvenue.js.map +0 -1
- package/out/src/tools/gb-title-case.d.ts.map +0 -1
- package/out/src/tools/gb-title-case.js.map +0 -1
- package/out/src/tools/golden-expand.d.ts +0 -90
- package/out/src/tools/golden-expand.d.ts.map +0 -1
- package/out/src/tools/golden-expand.js +0 -413
- package/out/src/tools/golden-expand.js.map +0 -1
- package/out/src/tools/golden-promote.d.ts.map +0 -1
- package/out/src/tools/golden-promote.js +0 -244
- package/out/src/tools/golden-promote.js.map +0 -1
- package/out/src/tools/golden-relabel-street.d.ts +0 -196
- package/out/src/tools/golden-relabel-street.d.ts.map +0 -1
- package/out/src/tools/golden-relabel-street.js +0 -440
- package/out/src/tools/golden-relabel-street.js.map +0 -1
- package/out/src/tools/index.d.ts +0 -28
- package/out/src/tools/index.d.ts.map +0 -1
- package/out/src/tools/index.js +0 -28
- package/out/src/tools/index.js.map +0 -1
- package/out/src/tools/ingest-csv.d.ts.map +0 -1
- package/out/src/tools/ingest-csv.js +0 -296
- package/out/src/tools/ingest-csv.js.map +0 -1
- package/out/src/tools/jsonl-to-parquet.d.ts +0 -61
- package/out/src/tools/jsonl-to-parquet.d.ts.map +0 -1
- package/out/src/tools/jsonl-to-parquet.js +0 -164
- package/out/src/tools/jsonl-to-parquet.js.map +0 -1
- package/out/src/tools/lint-shard-vocab.d.ts +0 -90
- package/out/src/tools/lint-shard-vocab.d.ts.map +0 -1
- package/out/src/tools/lint-shard-vocab.js +0 -281
- package/out/src/tools/lint-shard-vocab.js.map +0 -1
- package/out/src/tools/lint-shard.d.ts +0 -93
- package/out/src/tools/lint-shard.d.ts.map +0 -1
- package/out/src/tools/lint-shard.js +0 -354
- package/out/src/tools/lint-shard.js.map +0 -1
- package/out/src/tools/overlay-manifest.d.ts +0 -39
- package/out/src/tools/overlay-manifest.d.ts.map +0 -1
- package/out/src/tools/overlay-manifest.js +0 -98
- package/out/src/tools/overlay-manifest.js.map +0 -1
- package/out/src/tools/overture-subvenue.d.ts +0 -112
- package/out/src/tools/overture-subvenue.d.ts.map +0 -1
- package/out/src/tools/overture-subvenue.js +0 -144
- package/out/src/tools/overture-subvenue.js.map +0 -1
- package/out/src/tools/postcode-triples.d.ts +0 -175
- package/out/src/tools/postcode-triples.d.ts.map +0 -1
- package/out/src/tools/postcode-triples.js +0 -304
- package/out/src/tools/postcode-triples.js.map +0 -1
- package/out/src/tools/shard-kryptonite.d.ts +0 -32
- package/out/src/tools/shard-kryptonite.d.ts.map +0 -1
- package/out/src/tools/shard-kryptonite.js +0 -100
- package/out/src/tools/shard-kryptonite.js.map +0 -1
- package/out/src/tools/shard-translit.d.ts +0 -51
- package/out/src/tools/shard-translit.d.ts.map +0 -1
- package/out/src/tools/shard-translit.js +0 -175
- package/out/src/tools/shard-translit.js.map +0 -1
- package/out/src/tools/sub-venue/harvest.d.ts +0 -100
- package/out/src/tools/sub-venue/harvest.d.ts.map +0 -1
- package/out/src/tools/sub-venue/harvest.js +0 -168
- package/out/src/tools/sub-venue/harvest.js.map +0 -1
- package/out/src/tools/sub-venue/head-nouns.d.ts +0 -50
- package/out/src/tools/sub-venue/head-nouns.d.ts.map +0 -1
- package/out/src/tools/sub-venue/head-nouns.js +0 -210
- package/out/src/tools/sub-venue/head-nouns.js.map +0 -1
- package/out/src/tools/sub-venue/surfaces.d.ts +0 -45
- package/out/src/tools/sub-venue/surfaces.d.ts.map +0 -1
- package/out/src/tools/sub-venue/surfaces.js +0 -77
- package/out/src/tools/sub-venue/surfaces.js.map +0 -1
- package/out/src/tools/sub-venue/table.d.ts +0 -229
- package/out/src/tools/sub-venue/table.d.ts.map +0 -1
- package/out/src/tools/sub-venue/table.js +0 -108
- package/out/src/tools/sub-venue/table.js.map +0 -1
- package/out/src/tools/sub-venue/wikidata.d.ts +0 -26
- package/out/src/tools/sub-venue/wikidata.d.ts.map +0 -1
- package/out/src/tools/sub-venue/wikidata.js +0 -62
- package/out/src/tools/sub-venue/wikidata.js.map +0 -1
- package/out/src/tools/sub-venue-lexicon.d.ts +0 -156
- package/out/src/tools/sub-venue-lexicon.d.ts.map +0 -1
- package/out/src/tools/sub-venue-lexicon.js +0 -282
- package/out/src/tools/sub-venue-lexicon.js.map +0 -1
- package/out/src/tools/sub-venue-promotions.d.ts +0 -94
- package/out/src/tools/sub-venue-promotions.d.ts.map +0 -1
- package/out/src/tools/sub-venue-promotions.js +0 -266
- package/out/src/tools/sub-venue-promotions.js.map +0 -1
- package/out/src/types.d.ts +0 -221
- package/out/src/types.d.ts.map +0 -1
- package/out/src/types.js.map +0 -1
- package/out/src/utils/align.d.ts +0 -93
- package/out/src/utils/align.d.ts.map +0 -1
- package/out/src/utils/align.js +0 -234
- package/out/src/utils/align.js.map +0 -1
- package/out/src/utils/golden.d.ts +0 -69
- package/out/src/utils/golden.d.ts.map +0 -1
- package/out/src/utils/golden.js +0 -127
- package/out/src/utils/golden.js.map +0 -1
- package/out/src/utils/index.d.ts +0 -14
- package/out/src/utils/index.d.ts.map +0 -1
- package/out/src/utils/index.js +0 -14
- package/out/src/utils/index.js.map +0 -1
- package/out/src/utils/license.d.ts.map +0 -1
- package/out/src/utils/license.js.map +0 -1
- package/out/src/utils/parquet.d.ts +0 -162
- package/out/src/utils/parquet.d.ts.map +0 -1
- package/out/src/utils/parquet.js +0 -262
- package/out/src/utils/parquet.js.map +0 -1
- package/out/src/utils/split.d.ts +0 -127
- package/out/src/utils/split.d.ts.map +0 -1
- package/out/src/utils/split.js +0 -204
- package/out/src/utils/split.js.map +0 -1
- package/out/src/utils/tokenize.d.ts +0 -56
- package/out/src/utils/tokenize.d.ts.map +0 -1
- package/out/src/utils/tokenize.js +0 -48
- package/out/src/utils/tokenize.js.map +0 -1
- package/out/src/utils/wof-json.d.ts +0 -109
- package/out/src/utils/wof-json.d.ts.map +0 -1
- package/out/src/utils/wof-json.js +0 -173
- package/out/src/utils/wof-json.js.map +0 -1
- package/src/adapters/ban/adapter.ts +0 -191
- package/src/adapters/ban/street-decompose.ts +0 -96
- package/src/adapters/fcc-bdc/adapter.ts +0 -174
- package/src/adapters/geonames/adapter.ts +0 -200
- package/src/adapters/geonames-postal/adapter.ts +0 -119
- package/src/adapters/gnaf/adapter.ts +0 -149
- package/src/adapters/gnaf/assemble.ts +0 -226
- package/src/adapters/index.ts +0 -149
- package/src/adapters/openaddresses/adapter.ts +0 -251
- package/src/adapters/overture/adapter.ts +0 -162
- package/src/adapters/state-hi-schools/adapter.ts +0 -164
- package/src/adapters/synth-po-box/adapter.ts +0 -204
- package/src/adapters/tiger/adapter.ts +0 -253
- package/src/adapters/tiger/street-decompose.ts +0 -111
- package/src/adapters/usgov-hrsa-fqhc/adapter.ts +0 -183
- package/src/adapters/usgov-imls-pls/adapter.ts +0 -143
- package/src/adapters/usgov-irs-bmf/adapter.ts +0 -163
- package/src/adapters/usgov-nad/adapter.ts +0 -348
- package/src/adapters/usgov-nppes/adapter.ts +0 -159
- package/src/adapters/usgov-samhsa-treatment-locator/adapter.ts +0 -204
- package/src/adapters/utils/index.ts +0 -233
- package/src/adapters/wof-admin-jp/adapter.ts +0 -216
- package/src/adapters/wof-admin-json/adapter.ts +0 -256
- package/src/adapters/wof-json-rows.ts +0 -79
- package/src/adapters/wof-postalcode-json/adapter.ts +0 -183
- package/src/build.ts +0 -386
- package/src/index.ts +0 -14
- package/src/name-prone-us-suffixes.ts +0 -12
- package/src/parquet-wrapper/index.ts +0 -9
- package/src/parquet-wrapper/reader.ts +0 -97
- package/src/parquet-wrapper/schema.ts +0 -88
- package/src/parquet-wrapper/writer.ts +0 -110
- package/src/runner.ts +0 -317
- package/src/shard-recipes/anchor-absorption.ts +0 -103
- package/src/shard-recipes/bare-country.ts +0 -95
- package/src/shard-recipes/boundary-stress.ts +0 -101
- package/src/shard-recipes/country-balanced.ts +0 -524
- package/src/shard-recipes/cz-pcfirst-preposition.ts +0 -111
- package/src/shard-recipes/fr-admin-split.ts +0 -239
- package/src/shard-recipes/fr-bare-street.ts +0 -170
- package/src/shard-recipes/fr-fragment.ts +0 -332
- package/src/shard-recipes/fr-lieudit.ts +0 -302
- package/src/shard-recipes/fr-order.ts +0 -258
- package/src/shard-recipes/german.ts +0 -170
- package/src/shard-recipes/house-venue.ts +0 -88
- package/src/shard-recipes/index.ts +0 -95
- package/src/shard-recipes/intersection.ts +0 -630
- package/src/shard-recipes/locale.ts +0 -599
- package/src/shard-recipes/nl-postcode.ts +0 -115
- package/src/shard-recipes/no-fragment.ts +0 -219
- package/src/shard-recipes/no-street-led.ts +0 -143
- package/src/shard-recipes/no-street.ts +0 -86
- package/src/shard-recipes/po-box-cedex.ts +0 -920
- package/src/shard-recipes/po-box.ts +0 -119
- package/src/shard-recipes/reviewed-postcode-tail.ts +0 -189
- package/src/shard-recipes/scaffold.ts +0 -360
- package/src/shard-recipes/si-bare-village.ts +0 -97
- package/src/shard-recipes/street-affix.ts +0 -826
- package/src/shard-recipes/street-bare.ts +0 -101
- package/src/shard-recipes/street.ts +0 -82
- package/src/shard-recipes/sub-venue-sources.ts +0 -885
- package/src/shard-recipes/sub-venue.ts +0 -984
- package/src/shard-recipes/trailing-region.ts +0 -183
- package/src/shard-recipes/unit.ts +0 -321
- package/src/synthesizers/anchor-absorption.ts +0 -290
- package/src/synthesizers/boundary-stress.ts +0 -531
- package/src/synthesizers/german.ts +0 -251
- package/src/synthesizers/house-venue.ts +0 -330
- package/src/synthesizers/intersection.ts +0 -232
- package/src/synthesizers/no-street.ts +0 -349
- package/src/synthesizers/po-box.ts +0 -339
- package/src/synthesizers/street.ts +0 -274
- package/src/synthesizers/utils.ts +0 -903
- package/src/tools/align-shard.ts +0 -70
- package/src/tools/audit.ts +0 -399
- package/src/tools/corpus-stats.ts +0 -184
- package/src/tools/fetch/ban.ts +0 -290
- package/src/tools/fetch/download.ts +0 -176
- package/src/tools/fetch/geonames-dump.ts +0 -267
- package/src/tools/fetch/geonames-postal.ts +0 -180
- package/src/tools/fetch/hrsa.ts +0 -68
- package/src/tools/fetch/imls-pls.ts +0 -158
- package/src/tools/fetch/index.ts +0 -147
- package/src/tools/fetch/nad.ts +0 -342
- package/src/tools/fetch/nppes.ts +0 -174
- package/src/tools/fetch/openaddresses.ts +0 -404
- package/src/tools/fetch/ourairports.ts +0 -171
- package/src/tools/fetch/ppd.ts +0 -137
- package/src/tools/fetch/state-hi-schools.ts +0 -208
- package/src/tools/fetch/state-sources.ts +0 -155
- package/src/tools/fetch/tiger-full.ts +0 -354
- package/src/tools/fetch/wikidata-subvenue.ts +0 -386
- package/src/tools/golden-expand.ts +0 -584
- package/src/tools/golden-promote.ts +0 -353
- package/src/tools/golden-relabel-street.ts +0 -685
- package/src/tools/index.ts +0 -28
- package/src/tools/ingest-csv.ts +0 -422
- package/src/tools/jsonl-to-parquet.ts +0 -221
- package/src/tools/lint-shard-vocab.ts +0 -409
- package/src/tools/lint-shard.ts +0 -543
- package/src/tools/overlay-manifest.ts +0 -157
- package/src/tools/overture-subvenue.ts +0 -215
- package/src/tools/postcode-triples.ts +0 -375
- package/src/tools/shard-kryptonite.ts +0 -139
- package/src/tools/shard-translit.ts +0 -245
- package/src/tools/sub-venue/harvest.ts +0 -235
- package/src/tools/sub-venue/head-nouns.ts +0 -242
- package/src/tools/sub-venue/surfaces.ts +0 -95
- package/src/tools/sub-venue/table.ts +0 -281
- package/src/tools/sub-venue/wikidata.ts +0 -85
- package/src/tools/sub-venue-lexicon.ts +0 -431
- package/src/tools/sub-venue-promotions.ts +0 -330
- package/src/types.ts +0 -247
- package/src/utils/align.ts +0 -331
- package/src/utils/golden.ts +0 -176
- package/src/utils/index.ts +0 -14
- package/src/utils/parquet.ts +0 -391
- package/src/utils/split.ts +0 -279
- package/src/utils/tokenize.ts +0 -78
- package/src/utils/wof-json.ts +0 -244
- /package/{src → lib}/adapters/state-ia-contractors/adapter.ts +0 -0
- /package/{src → lib}/adapters/state-ny-notaries/adapter.ts +0 -0
- /package/{src → lib}/adapters/state-tx-notaries/adapter.ts +0 -0
- /package/{src → lib}/codex/us-fips-state.ts +0 -0
- /package/{src → lib}/tools/gb-title-case.ts +0 -0
- /package/{src → lib}/utils/license.ts +0 -0
- /package/out/{src/adapters → adapters}/ban/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/ban/street-decompose.d.ts +0 -0
- /package/out/{src/adapters → adapters}/fcc-bdc/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/geonames/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/openaddresses/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/state-ia-contractors/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/state-ia-contractors/adapter.js +0 -0
- /package/out/{src/adapters → adapters}/state-ny-notaries/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/state-ny-notaries/adapter.js +0 -0
- /package/out/{src/adapters → adapters}/state-tx-notaries/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/state-tx-notaries/adapter.js +0 -0
- /package/out/{src/adapters → adapters}/synth-po-box/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/tiger/street-decompose.d.ts +0 -0
- /package/out/{src/adapters → adapters}/usgov-nppes/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/usgov-samhsa-treatment-locator/adapter.d.ts +0 -0
- /package/out/{src/adapters → adapters}/wof-admin-jp/adapter.d.ts +0 -0
- /package/out/{src/codex → codex}/us-fips-state.d.ts +0 -0
- /package/out/{src/codex → codex}/us-fips-state.js +0 -0
- /package/out/{src/tools → tools}/fetch/ppd.d.ts +0 -0
- /package/out/{src/tools → tools}/gb-title-case.d.ts +0 -0
- /package/out/{src/tools → tools}/gb-title-case.js +0 -0
- /package/out/{src/tools → tools}/golden-promote.d.ts +0 -0
- /package/out/{src/tools → tools}/ingest-csv.d.ts +0 -0
- /package/out/{src/types.js → types.js} +0 -0
- /package/out/{src/utils → utils}/license.d.ts +0 -0
- /package/out/{src/utils → utils}/license.js +0 -0
|
@@ -0,0 +1,406 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* @copyright Sister Software
|
|
3
|
+
* @license AGPL-3.0
|
|
4
|
+
* @author Teffen Ellis, et al.
|
|
5
|
+
*
|
|
6
|
+
* Generate golden-set candidate entries by LLM-driven surface-form synthesis from a verified-label
|
|
7
|
+
* seed pulled out of the corpus's labeled test split.
|
|
8
|
+
*
|
|
9
|
+
* ## Why this approach
|
|
10
|
+
*
|
|
11
|
+
* The Phase 2 golden set has 74 entries; session-notes.md called for ≥500/locale. Manual curation
|
|
12
|
+
* doesn't scale. Pure-LLM generation (invent raw + labels from scratch) is too noisy — labels
|
|
13
|
+
* would be unverified.
|
|
14
|
+
*
|
|
15
|
+
* This module takes the middle path:
|
|
16
|
+
*
|
|
17
|
+
* 1. **Seeds come from corpus-v0.2.0 test slice** — already through the alignment pipeline, so labels
|
|
18
|
+
* are pipeline-verified.
|
|
19
|
+
* 2. **LLM only varies the surface form** — case, abbreviations, reordering, dropped components. The
|
|
20
|
+
* component VALUES (locality string, postcode digits, etc.) are preserved verbatim.
|
|
21
|
+
* 3. **Programmatic validator drops hallucinations** — every component value must appear as a
|
|
22
|
+
* substring (case-insensitive, whitespace-normalized) of the variant's raw. Failures dropped
|
|
23
|
+
* silently; cost is wasted tokens, never bad-labeled golden entries.
|
|
24
|
+
*
|
|
25
|
+
* ## Usage
|
|
26
|
+
*
|
|
27
|
+
* ```sh
|
|
28
|
+
* DEEPSEEK_API_KEY=sk-... \
|
|
29
|
+
* mailwoman corpus golden expand \
|
|
30
|
+
* --count 1000 \
|
|
31
|
+
* --variants 5 \
|
|
32
|
+
* --output data/eval/golden/candidates/expand-$(date +%Y%m%d-%H%M%S).jsonl
|
|
33
|
+
* ```
|
|
34
|
+
*
|
|
35
|
+
* ## Env
|
|
36
|
+
*
|
|
37
|
+
* - `DEEPSEEK_API_KEY` — required for provider `deepseek`
|
|
38
|
+
* - `ANTHROPIC_API_KEY` — required for provider `anthropic`
|
|
39
|
+
*
|
|
40
|
+
* ## What this module does NOT do
|
|
41
|
+
*
|
|
42
|
+
* - Does not commit anything or modify the versioned golden dir. Candidates land in
|
|
43
|
+
* `data/eval/golden/candidates/` for operator review (skim, prune, then run
|
|
44
|
+
* `mailwoman corpus golden promote`).
|
|
45
|
+
* - Does not score the LLM's quality — that's an eyeball job after pilot lands.
|
|
46
|
+
* - Does not retry hallucinated candidates. Cost of wasted tokens is trivial (~$0.0006/each).
|
|
47
|
+
*/
|
|
48
|
+
import { APIClient, pluckResponseData } from "@mailwoman/core/api";
|
|
49
|
+
import { dataRootPath } from "@mailwoman/core/data-root";
|
|
50
|
+
import { makeDirectories } from "@mailwoman/core/fs/writers";
|
|
51
|
+
import { tryParsingJSON } from "@mailwoman/core/json";
|
|
52
|
+
import { isPresent } from "@mailwoman/core/objects";
|
|
53
|
+
import { foldCaseWhitespace } from "@mailwoman/normalize/fold";
|
|
54
|
+
import { dirname } from "path-ts";
|
|
55
|
+
import { createNewlineWriter } from "spliterator";
|
|
56
|
+
import { $private } from "#env";
|
|
57
|
+
import { streamParquetRows } from "#utils/parquet";
|
|
58
|
+
// ── Types ─────────────────────────────────────────────────────────────────
|
|
59
|
+
/**
|
|
60
|
+
* Longest raw line still plausibly a single address rather than a concatenated record.
|
|
61
|
+
*/
|
|
62
|
+
const MAX_CANDIDATE_LENGTH = 500;
|
|
63
|
+
// ── Seed loading ──────────────────────────────────────────────────────────
|
|
64
|
+
/**
|
|
65
|
+
* Decode BIO labels + tokens into a verified components map. Mirrors the Python `decode_components` in
|
|
66
|
+
* mailwoman_train/eval.py — first-occurrence-wins per tag, contiguous B-X/I-X runs concatenated with a single space
|
|
67
|
+
* (the canonical separator used by corpus alignment).
|
|
68
|
+
*/
|
|
69
|
+
function decodeComponents(tokens, labels) {
|
|
70
|
+
const out = {};
|
|
71
|
+
let currentTag = null;
|
|
72
|
+
let currentTokens = [];
|
|
73
|
+
const flush = () => {
|
|
74
|
+
if (currentTag && currentTokens.length && !(currentTag in out)) {
|
|
75
|
+
out[currentTag] = currentTokens.join(" ").trim();
|
|
76
|
+
}
|
|
77
|
+
currentTag = null;
|
|
78
|
+
currentTokens = [];
|
|
79
|
+
};
|
|
80
|
+
for (let i = 0; i < labels.length; i++) {
|
|
81
|
+
const label = labels[i];
|
|
82
|
+
const tok = tokens[i] ?? "";
|
|
83
|
+
if (label === "O") {
|
|
84
|
+
flush();
|
|
85
|
+
continue;
|
|
86
|
+
}
|
|
87
|
+
const [prefix, tag] = label.split("-", 2);
|
|
88
|
+
if (prefix === "B" || currentTag !== tag) {
|
|
89
|
+
flush();
|
|
90
|
+
currentTag = tag;
|
|
91
|
+
currentTokens = [tok];
|
|
92
|
+
}
|
|
93
|
+
else {
|
|
94
|
+
currentTokens.push(tok);
|
|
95
|
+
}
|
|
96
|
+
}
|
|
97
|
+
flush();
|
|
98
|
+
return out;
|
|
99
|
+
}
|
|
100
|
+
async function loadSeeds(corpusPath, count, includeSources, report) {
|
|
101
|
+
const paths = corpusPath
|
|
102
|
+
.split(",")
|
|
103
|
+
.map((p) => p.trim())
|
|
104
|
+
.filter(isPresent);
|
|
105
|
+
report?.(`reading seeds from ${paths.length} slice(s) (target: ${count}, stratified)`);
|
|
106
|
+
if (includeSources) {
|
|
107
|
+
report?.(` include-sources filter: ${Array.from(includeSources).join(", ")}`);
|
|
108
|
+
}
|
|
109
|
+
// Stratified sampling: read all rows from all slices, group by source. Bounded by per-source
|
|
110
|
+
// reservoir: keep at most max(2*count, 5000) rows per source so we don't blow memory on train
|
|
111
|
+
// slices (1M rows × many slices). Sampling later is uniform within each pool.
|
|
112
|
+
const bySource = new Map();
|
|
113
|
+
const PER_SOURCE_CAP = Math.max(2 * count, 5000);
|
|
114
|
+
let scanned = 0;
|
|
115
|
+
let skippedThinComponents = 0;
|
|
116
|
+
for (const path of paths) {
|
|
117
|
+
for await (const row of streamParquetRows(path)) {
|
|
118
|
+
scanned++;
|
|
119
|
+
// Source allow-list (--include-sources) — applied early to skip parsing rows we won't use
|
|
120
|
+
if (includeSources && !includeSources.has(row.source))
|
|
121
|
+
continue;
|
|
122
|
+
const components = decodeComponents(row.tokens ?? [], row.labels ?? []);
|
|
123
|
+
// Skip rows with too few components — single-name wof-admin entries don't make useful seeds
|
|
124
|
+
if (Object.keys(components).length < 2) {
|
|
125
|
+
skippedThinComponents++;
|
|
126
|
+
continue;
|
|
127
|
+
}
|
|
128
|
+
const seed = {
|
|
129
|
+
raw: row.raw,
|
|
130
|
+
components,
|
|
131
|
+
country: row.country,
|
|
132
|
+
source: row.source,
|
|
133
|
+
source_id: row.source_id,
|
|
134
|
+
};
|
|
135
|
+
let bucket = bySource.get(row.source);
|
|
136
|
+
if (!bucket) {
|
|
137
|
+
bucket = [];
|
|
138
|
+
bySource.set(row.source, bucket);
|
|
139
|
+
}
|
|
140
|
+
if (bucket.length < PER_SOURCE_CAP) {
|
|
141
|
+
bucket.push(seed);
|
|
142
|
+
}
|
|
143
|
+
}
|
|
144
|
+
}
|
|
145
|
+
report?.(` scanned ${scanned} rows across ${paths.length} slice(s); thin-components dropped: ${skippedThinComponents}`);
|
|
146
|
+
report?.(` per-source pool sizes:`);
|
|
147
|
+
for (const [src, pool] of bySource) {
|
|
148
|
+
report?.(` ${src}: ${pool.length}`);
|
|
149
|
+
}
|
|
150
|
+
// Round-robin sample. Each source gives floor(count / nSources) seeds; rounding goes
|
|
151
|
+
// to sources in alphabetical order. If a pool is smaller than its target, take all of it.
|
|
152
|
+
const sources = Array.from(bySource.keys()).toSorted();
|
|
153
|
+
const perSource = Math.floor(count / sources.length);
|
|
154
|
+
const remainder = count - perSource * sources.length;
|
|
155
|
+
const picked = [];
|
|
156
|
+
for (let i = 0; i < sources.length; i++) {
|
|
157
|
+
const src = sources[i];
|
|
158
|
+
const pool = bySource.get(src);
|
|
159
|
+
const target = perSource + (i < remainder ? 1 : 0);
|
|
160
|
+
// Random subsample without replacement — deterministic via shuffle then slice
|
|
161
|
+
for (let j = pool.length - 1; j > 0; j--) {
|
|
162
|
+
const k = Math.floor(Math.random() * (j + 1));
|
|
163
|
+
[pool[j], pool[k]] = [pool[k], pool[j]];
|
|
164
|
+
}
|
|
165
|
+
const take = Math.min(target, pool.length);
|
|
166
|
+
picked.push(...pool.slice(0, take));
|
|
167
|
+
if (take < target) {
|
|
168
|
+
report?.(` ⚠ ${src}: requested ${target}, pool had ${pool.length}`);
|
|
169
|
+
}
|
|
170
|
+
}
|
|
171
|
+
report?.(` → loaded ${picked.length} seeds across ${sources.length} sources`);
|
|
172
|
+
return picked;
|
|
173
|
+
}
|
|
174
|
+
const SYSTEM_PROMPT = `You are a postal-address surface-form generator. Given a structured address, produce realistic variants a human might type into a geocoder.
|
|
175
|
+
|
|
176
|
+
CONSTRAINT — you MUST preserve every kept component value verbatim in the output. You may:
|
|
177
|
+
- vary case (UPPER, lower, Title Case)
|
|
178
|
+
- abbreviate (Saint → St, Avenue → Ave, Boulevard → Blvd, North → N, etc.)
|
|
179
|
+
- vary punctuation (commas, dashes, spaces, line breaks)
|
|
180
|
+
- reorder components (postcode-first, country-first, address-only)
|
|
181
|
+
- drop OPTIONAL components — list them in "dropped"
|
|
182
|
+
|
|
183
|
+
OPTIONAL components (allowed to drop): country, postcode, dependent_locality, subregion, cedex.
|
|
184
|
+
REQUIRED components (must keep, even if input has them): locality, region (when present),
|
|
185
|
+
street, house_number, venue (when present).
|
|
186
|
+
|
|
187
|
+
ALWAYS keep AT LEAST 2 components in the final raw text. Single-component variants like
|
|
188
|
+
"VT" or "Paris" are USELESS as eval entries — do not produce them.
|
|
189
|
+
|
|
190
|
+
You MUST NOT:
|
|
191
|
+
- introduce typos (validator drops these silently — wasted tokens)
|
|
192
|
+
- invent new component values
|
|
193
|
+
- produce text longer than 500 characters
|
|
194
|
+
- output a degenerate single-token answer
|
|
195
|
+
|
|
196
|
+
OUTPUT a JSON array of N objects, each shaped {"raw": "...", "dropped": ["..."]}.`;
|
|
197
|
+
function buildUserPrompt(seed, n) {
|
|
198
|
+
return `INPUT:
|
|
199
|
+
${JSON.stringify({ raw: seed.raw, components: seed.components, country: seed.country }, null, 2)}
|
|
200
|
+
|
|
201
|
+
N: ${n}`;
|
|
202
|
+
}
|
|
203
|
+
function makeDeepseekProvider(model) {
|
|
204
|
+
const apiKey = $private.DEEPSEEK_API_KEY;
|
|
205
|
+
if (!apiKey)
|
|
206
|
+
throw new Error("DEEPSEEK_API_KEY env var is required for provider deepseek");
|
|
207
|
+
return {
|
|
208
|
+
name: "deepseek",
|
|
209
|
+
model,
|
|
210
|
+
async generateVariants(seed, n) {
|
|
211
|
+
const data = await new APIClient({
|
|
212
|
+
displayName: "deepseek",
|
|
213
|
+
retry: true,
|
|
214
|
+
axios: { headers: { Authorization: `Bearer ${apiKey}`, "Content-Type": "application/json" } },
|
|
215
|
+
})
|
|
216
|
+
.fetch({
|
|
217
|
+
url: "https://api.deepseek.com/chat/completions",
|
|
218
|
+
method: "post",
|
|
219
|
+
timeout: 60_000,
|
|
220
|
+
data: {
|
|
221
|
+
model,
|
|
222
|
+
messages: [
|
|
223
|
+
{ role: "system", content: SYSTEM_PROMPT },
|
|
224
|
+
{ role: "user", content: buildUserPrompt(seed, n) },
|
|
225
|
+
],
|
|
226
|
+
response_format: { type: "json_object" },
|
|
227
|
+
max_tokens: 800,
|
|
228
|
+
temperature: 0.7,
|
|
229
|
+
},
|
|
230
|
+
})
|
|
231
|
+
.then(pluckResponseData);
|
|
232
|
+
const content = data.choices?.[0]?.message.content ?? "{}";
|
|
233
|
+
return parseCandidates(content);
|
|
234
|
+
},
|
|
235
|
+
};
|
|
236
|
+
}
|
|
237
|
+
function makeAnthropicProvider(model) {
|
|
238
|
+
const apiKey = $private.ANTHROPIC_API_KEY;
|
|
239
|
+
if (!apiKey)
|
|
240
|
+
throw new Error("ANTHROPIC_API_KEY env var is required for provider anthropic");
|
|
241
|
+
return {
|
|
242
|
+
name: "anthropic",
|
|
243
|
+
model,
|
|
244
|
+
async generateVariants(seed, n) {
|
|
245
|
+
const data = await new APIClient({
|
|
246
|
+
displayName: "anthropic",
|
|
247
|
+
retry: true,
|
|
248
|
+
axios: {
|
|
249
|
+
headers: {
|
|
250
|
+
"x-api-key": apiKey,
|
|
251
|
+
"anthropic-version": "2023-06-01",
|
|
252
|
+
"Content-Type": "application/json",
|
|
253
|
+
},
|
|
254
|
+
},
|
|
255
|
+
})
|
|
256
|
+
.fetch({
|
|
257
|
+
url: "https://api.anthropic.com/v1/messages",
|
|
258
|
+
method: "post",
|
|
259
|
+
timeout: 60_000,
|
|
260
|
+
data: {
|
|
261
|
+
model,
|
|
262
|
+
max_tokens: 800,
|
|
263
|
+
system: SYSTEM_PROMPT,
|
|
264
|
+
messages: [{ role: "user", content: buildUserPrompt(seed, n) }],
|
|
265
|
+
},
|
|
266
|
+
})
|
|
267
|
+
.then(pluckResponseData);
|
|
268
|
+
const text = data.content?.find((c) => c.type === "text")?.text ?? "[]";
|
|
269
|
+
return parseCandidates(text);
|
|
270
|
+
},
|
|
271
|
+
};
|
|
272
|
+
}
|
|
273
|
+
function parseCandidates(text) {
|
|
274
|
+
// Strip markdown fences the model sometimes wraps around JSON
|
|
275
|
+
const cleaned = text.replaceAll(/^```(?:json)?\n?|\n?```$/g, "").trim();
|
|
276
|
+
const parsed = tryParsingJSON(cleaned);
|
|
277
|
+
if (Array.isArray(parsed))
|
|
278
|
+
return parsed;
|
|
279
|
+
// Some providers wrap in {"variants": [...]} or {"candidates": [...]}
|
|
280
|
+
if (typeof parsed === "object" && parsed !== null) {
|
|
281
|
+
for (const key of ["variants", "candidates", "results"]) {
|
|
282
|
+
const v = parsed[key];
|
|
283
|
+
if (Array.isArray(v))
|
|
284
|
+
return v;
|
|
285
|
+
}
|
|
286
|
+
}
|
|
287
|
+
return [];
|
|
288
|
+
}
|
|
289
|
+
// ── Validator ─────────────────────────────────────────────────────────────
|
|
290
|
+
/**
|
|
291
|
+
* Components that are NEVER allowed to be dropped — keeps degenerate single-token candidates out.
|
|
292
|
+
*/
|
|
293
|
+
const REQUIRED_COMPONENT_TAGS = new Set(["locality", "region", "street", "house_number", "venue"]);
|
|
294
|
+
function validate(seed, candidate) {
|
|
295
|
+
if (!candidate.raw || typeof candidate.raw !== "string")
|
|
296
|
+
return false;
|
|
297
|
+
if (candidate.raw.length > MAX_CANDIDATE_LENGTH)
|
|
298
|
+
return false;
|
|
299
|
+
if (/```|<\/?\w+>|^\s*\{/.test(candidate.raw))
|
|
300
|
+
return false;
|
|
301
|
+
const normRaw = foldCaseWhitespace(candidate.raw);
|
|
302
|
+
const dropped = new Set(candidate.dropped);
|
|
303
|
+
// LLM cannot drop required components present in the seed.
|
|
304
|
+
for (const tag of dropped) {
|
|
305
|
+
if (REQUIRED_COMPONENT_TAGS.has(tag) && seed.components[tag])
|
|
306
|
+
return false;
|
|
307
|
+
}
|
|
308
|
+
// Every kept component value must appear verbatim (post-normalization) in the candidate raw.
|
|
309
|
+
let keptCount = 0;
|
|
310
|
+
for (const [tag, value] of Object.entries(seed.components)) {
|
|
311
|
+
if (dropped.has(tag))
|
|
312
|
+
continue;
|
|
313
|
+
if (!value)
|
|
314
|
+
continue;
|
|
315
|
+
if (!normRaw.includes(foldCaseWhitespace(value)))
|
|
316
|
+
return false;
|
|
317
|
+
keptCount++;
|
|
318
|
+
}
|
|
319
|
+
// Reject degenerate single-component candidates ("VT", "Paris" alone).
|
|
320
|
+
if (keptCount < 2)
|
|
321
|
+
return false;
|
|
322
|
+
return true;
|
|
323
|
+
}
|
|
324
|
+
// ── Main pipeline ─────────────────────────────────────────────────────────
|
|
325
|
+
export async function expandGolden(options = {}, report) {
|
|
326
|
+
const corpusPath = (options.corpus ?? dataRootPath("corpus", "versioned", "v0.2.0", "corpus-v0.2.0", "test", "part-0000.parquet")).toString();
|
|
327
|
+
const count = options.count ?? 100;
|
|
328
|
+
const variants = options.variants ?? 5;
|
|
329
|
+
const providerName = options.provider ?? "deepseek";
|
|
330
|
+
const model = options.model ?? (providerName === "anthropic" ? "claude-haiku-4-5-20251001" : "deepseek-chat");
|
|
331
|
+
const concurrencyLimit = options.concurrency ?? 4;
|
|
332
|
+
const includeSources = options.includeSources ? new Set(options.includeSources.split(",").map((s) => s.trim())) : null;
|
|
333
|
+
const ts = new Date().toISOString().replaceAll(/[:.]/g, "-").slice(0, 19);
|
|
334
|
+
const outputPath = options.output ?? `data/eval/golden/candidates/expand-${ts}.jsonl`;
|
|
335
|
+
const provider = providerName === "anthropic" ? makeAnthropicProvider(model) : makeDeepseekProvider(model);
|
|
336
|
+
report?.(`provider: ${provider.name} model: ${provider.model}`);
|
|
337
|
+
const seeds = await loadSeeds(corpusPath, count, includeSources, report);
|
|
338
|
+
if (!seeds.length) {
|
|
339
|
+
throw new Error("no seeds loaded — corpus path or filter is wrong");
|
|
340
|
+
}
|
|
341
|
+
await makeDirectories(dirname(outputPath));
|
|
342
|
+
const outRows = [];
|
|
343
|
+
let kept = 0;
|
|
344
|
+
let dropped = 0;
|
|
345
|
+
let errored = 0;
|
|
346
|
+
// Bounded-concurrency worker pool
|
|
347
|
+
let cursor = 0;
|
|
348
|
+
const workers = Array.from({ length: Math.min(concurrencyLimit, seeds.length) }, async () => {
|
|
349
|
+
while (true) {
|
|
350
|
+
const i = cursor++;
|
|
351
|
+
if (i >= seeds.length)
|
|
352
|
+
return;
|
|
353
|
+
const seed = seeds[i];
|
|
354
|
+
try {
|
|
355
|
+
const candidates = await provider.generateVariants(seed, variants);
|
|
356
|
+
for (const cand of candidates) {
|
|
357
|
+
if (validate(seed, cand)) {
|
|
358
|
+
const goldenCandidate = {
|
|
359
|
+
raw: cand.raw,
|
|
360
|
+
components: { ...seed.components },
|
|
361
|
+
country: seed.country,
|
|
362
|
+
source: `expand-golden:${provider.name}`,
|
|
363
|
+
seed_source_id: seed.source_id,
|
|
364
|
+
seed_source_adapter: seed.source,
|
|
365
|
+
dropped_components: cand.dropped ?? [],
|
|
366
|
+
provenance: { provider: provider.name, model: provider.model },
|
|
367
|
+
};
|
|
368
|
+
// Remove dropped components from the components map
|
|
369
|
+
for (const tag of goldenCandidate.dropped_components) {
|
|
370
|
+
// oxlint-disable-next-line typescript/no-dynamic-delete -- removing one key from a plain record; the object is not on a hot path
|
|
371
|
+
delete goldenCandidate.components[tag];
|
|
372
|
+
}
|
|
373
|
+
outRows.push(goldenCandidate);
|
|
374
|
+
kept++;
|
|
375
|
+
}
|
|
376
|
+
else {
|
|
377
|
+
dropped++;
|
|
378
|
+
}
|
|
379
|
+
}
|
|
380
|
+
}
|
|
381
|
+
catch (error) {
|
|
382
|
+
errored++;
|
|
383
|
+
report?.(` ✗ seed ${seed.source_id}: ${error.message}`);
|
|
384
|
+
}
|
|
385
|
+
if ((i + 1) % 10 === 0) {
|
|
386
|
+
report?.(` progress: ${i + 1}/${seeds.length} kept=${kept} dropped=${dropped} errored=${errored}`);
|
|
387
|
+
}
|
|
388
|
+
}
|
|
389
|
+
});
|
|
390
|
+
await Promise.all(workers);
|
|
391
|
+
{
|
|
392
|
+
await using out = createNewlineWriter(outputPath);
|
|
393
|
+
for (const row of outRows) {
|
|
394
|
+
await out.write(JSON.stringify(row));
|
|
395
|
+
}
|
|
396
|
+
}
|
|
397
|
+
report?.(`=== summary ===`);
|
|
398
|
+
report?.(`seeds processed: ${seeds.length}`);
|
|
399
|
+
report?.(`candidates kept: ${kept}`);
|
|
400
|
+
report?.(`candidates dropped (validator): ${dropped}`);
|
|
401
|
+
report?.(`seeds with errors: ${errored}`);
|
|
402
|
+
report?.(`yield: ${seeds.length ? ((kept / (seeds.length * variants)) * 100).toFixed(1) : "0"}%`);
|
|
403
|
+
report?.(`output: ${outputPath}`);
|
|
404
|
+
return { seedsProcessed: seeds.length, kept, dropped, errored, outputPath };
|
|
405
|
+
}
|
|
406
|
+
//# sourceMappingURL=golden-expand.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"golden-expand.js","sourceRoot":"","sources":["../../lib/tools/golden-expand.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8CG;AAEH,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAA;AAClE,OAAO,EAAE,YAAY,EAAE,MAAM,2BAA2B,CAAA;AACxD,OAAO,EAAE,eAAe,EAAE,MAAM,4BAA4B,CAAA;AAC5D,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAA;AACrD,OAAO,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAA;AACnD,OAAO,EAAE,kBAAkB,EAAE,MAAM,2BAA2B,CAAA;AAC9D,OAAO,EAAE,OAAO,EAAE,MAAM,SAAS,CAAA;AACjC,OAAO,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAA;AAEjD,OAAO,EAAE,QAAQ,EAAE,MAAM,MAAM,CAAA;AAC/B,OAAO,EAAE,iBAAiB,EAAE,MAAM,gBAAgB,CAAA;AAElD,6EAA6E;AAE7E;;GAEG;AACH,MAAM,oBAAoB,GAAG,GAAG,CAAA;AA+EhC,6EAA6E;AAE7E;;;;GAIG;AACH,SAAS,gBAAgB,CAAC,MAAgB,EAAE,MAAgB;IAC3D,MAAM,GAAG,GAA2B,EAAE,CAAA;IACtC,IAAI,UAAU,GAAkB,IAAI,CAAA;IACpC,IAAI,aAAa,GAAa,EAAE,CAAA;IAEhC,MAAM,KAAK,GAAG,GAAG,EAAE;QAClB,IAAI,UAAU,IAAI,aAAa,CAAC,MAAM,IAAI,CAAC,CAAC,UAAU,IAAI,GAAG,CAAC,EAAE,CAAC;YAChE,GAAG,CAAC,UAAU,CAAC,GAAG,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;QACjD,CAAC;QAED,UAAU,GAAG,IAAI,CAAA;QACjB,aAAa,GAAG,EAAE,CAAA;IACnB,CAAC,CAAA;IAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;QACxB,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QAE3B,IAAI,KAAK,KAAK,GAAG,EAAE,CAAC;YACnB,KAAK,EAAE,CAAA;YAEP,SAAQ;QACT,CAAC;QAED,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;QAEzC,IAAI,MAAM,KAAK,GAAG,IAAI,UAAU,KAAK,GAAG,EAAE,CAAC;YAC1C,KAAK,EAAE,CAAA;YACP,UAAU,GAAG,GAAI,CAAA;YACjB,aAAa,GAAG,CAAC,GAAG,CAAC,CAAA;QACtB,CAAC;aAAM,CAAC;YACP,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACxB,CAAC;IACF,CAAC;IAED,KAAK,EAAE,CAAA;IAEP,OAAO,GAAG,CAAA;AACX,CAAC;AAED,KAAK,UAAU,SAAS,CACvB,UAAkB,EAClB,KAAa,EACb,cAAkC,EAClC,MAA+B;IAE/B,MAAM,KAAK,GAAG,UAAU;SACtB,KAAK,CAAC,GAAG,CAAC;SACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,SAAS,CAAC,CAAA;IAEnB,MAAM,EAAE,CAAC,sBAAsB,KAAK,CAAC,MAAM,sBAAsB,KAAK,eAAe,CAAC,CAAA;IAEtF,IAAI,cAAc,EAAE,CAAC;QACpB,MAAM,EAAE,CAAC,6BAA6B,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IAC/E,CAAC;IAED,6FAA6F;IAC7F,8FAA8F;IAC9F,8EAA8E;IAC9E,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAA;IAC1C,MAAM,cAAc,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,KAAK,EAAE,IAAI,CAAC,CAAA;IAChD,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,IAAI,qBAAqB,GAAG,CAAC,CAAA;IAE7B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,iBAAiB,CAAY,IAAI,CAAC,EAAE,CAAC;YAC5D,OAAO,EAAE,CAAA;YAET,0FAA0F;YAC1F,IAAI,cAAc,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC;gBAAE,SAAQ;YAC/D,MAAM,UAAU,GAAG,gBAAgB,CAAC,GAAG,CAAC,MAAM,IAAI,EAAE,EAAE,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC,CAAA;YAEvE,4FAA4F;YAC5F,IAAI,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACxC,qBAAqB,EAAE,CAAA;gBAEvB,SAAQ;YACT,CAAC;YAED,MAAM,IAAI,GAAS;gBAClB,GAAG,EAAE,GAAG,CAAC,GAAG;gBACZ,UAAU;gBACV,OAAO,EAAE,GAAG,CAAC,OAAO;gBACpB,MAAM,EAAE,GAAG,CAAC,MAAM;gBAClB,SAAS,EAAE,GAAG,CAAC,SAAS;aACxB,CAAA;YAED,IAAI,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;YAErC,IAAI,CAAC,MAAM,EAAE,CAAC;gBACb,MAAM,GAAG,EAAE,CAAA;gBACX,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;YACjC,CAAC;YAED,IAAI,MAAM,CAAC,MAAM,GAAG,cAAc,EAAE,CAAC;gBACpC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAClB,CAAC;QACF,CAAC;IACF,CAAC;IAED,MAAM,EAAE,CACP,aAAa,OAAO,gBAAgB,KAAK,CAAC,MAAM,uCAAuC,qBAAqB,EAAE,CAC9G,CAAA;IAED,MAAM,EAAE,CAAC,0BAA0B,CAAC,CAAA;IAEpC,KAAK,MAAM,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,QAAQ,EAAE,CAAC;QACpC,MAAM,EAAE,CAAC,OAAO,GAAG,KAAK,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;IACvC,CAAC;IAED,qFAAqF;IACrF,0FAA0F;IAC1F,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAA;IACtD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAA;IACpD,MAAM,SAAS,GAAG,KAAK,GAAG,SAAS,GAAG,OAAO,CAAC,MAAM,CAAA;IACpD,MAAM,MAAM,GAAW,EAAE,CAAA;IAEzB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACzC,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAE,CAAA;QACvB,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAE,CAAA;QAC/B,MAAM,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;QAElD,8EAA8E;QAC9E,KAAK,IAAI,CAAC,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC1C,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAC5C;YAAA,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAE,EAAE,IAAI,CAAC,CAAC,CAAE,CAAC,CAAA;QAC3C,CAAC;QAED,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;QAC1C,MAAM,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAA;QAEnC,IAAI,IAAI,GAAG,MAAM,EAAE,CAAC;YACnB,MAAM,EAAE,CAAC,SAAS,GAAG,eAAe,MAAM,cAAc,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;QACvE,CAAC;IACF,CAAC;IAED,MAAM,EAAE,CAAC,cAAc,MAAM,CAAC,MAAM,iBAAiB,OAAO,CAAC,MAAM,UAAU,CAAC,CAAA;IAE9E,OAAO,MAAM,CAAA;AACd,CAAC;AAUD,MAAM,aAAa,GAAG;;;;;;;;;;;;;;;;;;;;;;kFAsB4D,CAAA;AAElF,SAAS,eAAe,CAAC,IAAU,EAAE,CAAS;IAC7C,OAAO;EACN,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,UAAU,EAAE,IAAI,CAAC,UAAU,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;;KAE3F,CAAC,EAAE,CAAA;AACR,CAAC;AAED,SAAS,oBAAoB,CAAC,KAAa;IAC1C,MAAM,MAAM,GAAG,QAAQ,CAAC,gBAAgB,CAAA;IAExC,IAAI,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,4DAA4D,CAAC,CAAA;IAE1F,OAAO;QACN,IAAI,EAAE,UAAU;QAChB,KAAK;QACL,KAAK,CAAC,gBAAgB,CAAC,IAAI,EAAE,CAAC;YAC7B,MAAM,IAAI,GAAG,MAAM,IAAI,SAAS,CAAC;gBAChC,WAAW,EAAE,UAAU;gBACvB,KAAK,EAAE,IAAI;gBACX,KAAK,EAAE,EAAE,OAAO,EAAE,EAAE,aAAa,EAAE,UAAU,MAAM,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,EAAE;aAC7F,CAAC;iBACA,KAAK,CAAwD;gBAC7D,GAAG,EAAE,2CAA2C;gBAChD,MAAM,EAAE,MAAM;gBACd,OAAO,EAAE,MAAM;gBACf,IAAI,EAAE;oBACL,KAAK;oBACL,QAAQ,EAAE;wBACT,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,aAAa,EAAE;wBAC1C,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,eAAe,CAAC,IAAI,EAAE,CAAC,CAAC,EAAE;qBACnD;oBACD,eAAe,EAAE,EAAE,IAAI,EAAE,aAAa,EAAE;oBACxC,UAAU,EAAE,GAAG;oBACf,WAAW,EAAE,GAAG;iBAChB;aACD,CAAC;iBACD,IAAI,CAAC,iBAAiB,CAAC,CAAA;YAEzB,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,OAAO,IAAI,IAAI,CAAA;YAE1D,OAAO,eAAe,CAAC,OAAO,CAAC,CAAA;QAChC,CAAC;KACD,CAAA;AACF,CAAC;AAED,SAAS,qBAAqB,CAAC,KAAa;IAC3C,MAAM,MAAM,GAAG,QAAQ,CAAC,iBAAiB,CAAA;IAEzC,IAAI,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,8DAA8D,CAAC,CAAA;IAE5F,OAAO;QACN,IAAI,EAAE,WAAW;QACjB,KAAK;QACL,KAAK,CAAC,gBAAgB,CAAC,IAAI,EAAE,CAAC;YAC7B,MAAM,IAAI,GAAG,MAAM,IAAI,SAAS,CAAC;gBAChC,WAAW,EAAE,WAAW;gBACxB,KAAK,EAAE,IAAI;gBACX,KAAK,EAAE;oBACN,OAAO,EAAE;wBACR,WAAW,EAAE,MAAM;wBACnB,mBAAmB,EAAE,YAAY;wBACjC,cAAc,EAAE,kBAAkB;qBAClC;iBACD;aACD,CAAC;iBACA,KAAK,CAAsD;gBAC3D,GAAG,EAAE,uCAAuC;gBAC5C,MAAM,EAAE,MAAM;gBACd,OAAO,EAAE,MAAM;gBACf,IAAI,EAAE;oBACL,KAAK;oBACL,UAAU,EAAE,GAAG;oBACf,MAAM,EAAE,aAAa;oBACrB,QAAQ,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,eAAe,CAAC,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC;iBAC/D;aACD,CAAC;iBACD,IAAI,CAAC,iBAAiB,CAAC,CAAA;YAEzB,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC,EAAE,IAAI,IAAI,IAAI,CAAA;YAEvE,OAAO,eAAe,CAAC,IAAI,CAAC,CAAA;QAC7B,CAAC;KACD,CAAA;AACF,CAAC;AAED,SAAS,eAAe,CAAC,IAAY;IACpC,8DAA8D;IAC9D,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,2BAA2B,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;IAEvE,MAAM,MAAM,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IAEtC,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;QAAE,OAAO,MAAqB,CAAA;IAEvD,sEAAsE;IACtE,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;QACnD,KAAK,MAAM,GAAG,IAAI,CAAC,UAAU,EAAE,YAAY,EAAE,SAAS,CAAC,EAAE,CAAC;YACzD,MAAM,CAAC,GAAI,MAAkC,CAAC,GAAG,CAAC,CAAA;YAElD,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;gBAAE,OAAO,CAAgB,CAAA;QAC9C,CAAC;IACF,CAAC;IAED,OAAO,EAAE,CAAA;AACV,CAAC;AAED,6EAA6E;AAE7E;;GAEG;AACH,MAAM,uBAAuB,GAAG,IAAI,GAAG,CAAC,CAAC,UAAU,EAAE,QAAQ,EAAE,QAAQ,EAAE,cAAc,EAAE,OAAO,CAAC,CAAC,CAAA;AAElG,SAAS,QAAQ,CAAC,IAAU,EAAE,SAAoB;IACjD,IAAI,CAAC,SAAS,CAAC,GAAG,IAAI,OAAO,SAAS,CAAC,GAAG,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAA;IAErE,IAAI,SAAS,CAAC,GAAG,CAAC,MAAM,GAAG,oBAAoB;QAAE,OAAO,KAAK,CAAA;IAE7D,IAAI,qBAAqB,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAA;IAC3D,MAAM,OAAO,GAAG,kBAAkB,CAAC,SAAS,CAAC,GAAG,CAAC,CAAA;IACjD,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAA;IAE1C,2DAA2D;IAC3D,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;QAC3B,IAAI,uBAAuB,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,OAAO,KAAK,CAAA;IAC3E,CAAC;IAED,6FAA6F;IAC7F,IAAI,SAAS,GAAG,CAAC,CAAA;IAEjB,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,EAAE,CAAC;QAC5D,IAAI,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAQ;QAE9B,IAAI,CAAC,KAAK;YAAE,SAAQ;QAEpB,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAC;YAAE,OAAO,KAAK,CAAA;QAE9D,SAAS,EAAE,CAAA;IACZ,CAAC;IAED,uEAAuE;IACvE,IAAI,SAAS,GAAG,CAAC;QAAE,OAAO,KAAK,CAAA;IAE/B,OAAO,IAAI,CAAA;AACZ,CAAC;AAED,6EAA6E;AAE7E,MAAM,CAAC,KAAK,UAAU,YAAY,CACjC,UAA+B,EAAE,EACjC,MAA+B;IAE/B,MAAM,UAAU,GAAG,CAClB,OAAO,CAAC,MAAM,IAAI,YAAY,CAAC,QAAQ,EAAE,WAAW,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,EAAE,mBAAmB,CAAC,CAC7G,CAAC,QAAQ,EAAE,CAAA;IAEZ,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,IAAI,GAAG,CAAA;IAClC,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,CAAC,CAAA;IACtC,MAAM,YAAY,GAAG,OAAO,CAAC,QAAQ,IAAI,UAAU,CAAA;IACnD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,IAAI,CAAC,YAAY,KAAK,WAAW,CAAC,CAAC,CAAC,2BAA2B,CAAC,CAAC,CAAC,eAAe,CAAC,CAAA;IAC7G,MAAM,gBAAgB,GAAG,OAAO,CAAC,WAAW,IAAI,CAAC,CAAA;IACjD,MAAM,cAAc,GAAG,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC,cAAc,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IACtH,MAAM,EAAE,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;IACzE,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,IAAI,sCAAsC,EAAE,QAAQ,CAAA;IAErF,MAAM,QAAQ,GAAG,YAAY,KAAK,WAAW,CAAC,CAAC,CAAC,qBAAqB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,oBAAoB,CAAC,KAAK,CAAC,CAAA;IAC1G,MAAM,EAAE,CAAC,aAAa,QAAQ,CAAC,IAAI,YAAY,QAAQ,CAAC,KAAK,EAAE,CAAC,CAAA;IAEhE,MAAM,KAAK,GAAG,MAAM,SAAS,CAAC,UAAU,EAAE,KAAK,EAAE,cAAc,EAAE,MAAM,CAAC,CAAA;IAExE,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;QACnB,MAAM,IAAI,KAAK,CAAC,kDAAkD,CAAC,CAAA;IACpE,CAAC;IAED,MAAM,eAAe,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAA;IAC1C,MAAM,OAAO,GAAsB,EAAE,CAAA;IACrC,IAAI,IAAI,GAAG,CAAC,CAAA;IACZ,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,IAAI,OAAO,GAAG,CAAC,CAAA;IAEf,kCAAkC;IAClC,IAAI,MAAM,GAAG,CAAC,CAAA;IAEd,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,gBAAgB,EAAE,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,KAAK,IAAI,EAAE;QAC3F,OAAO,IAAI,EAAE,CAAC;YACb,MAAM,CAAC,GAAG,MAAM,EAAE,CAAA;YAElB,IAAI,CAAC,IAAI,KAAK,CAAC,MAAM;gBAAE,OAAM;YAC7B,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YAEtB,IAAI,CAAC;gBACJ,MAAM,UAAU,GAAG,MAAM,QAAQ,CAAC,gBAAgB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAA;gBAElE,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;oBAC/B,IAAI,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC;wBAC1B,MAAM,eAAe,GAAoB;4BACxC,GAAG,EAAE,IAAI,CAAC,GAAG;4BACb,UAAU,EAAE,EAAE,GAAG,IAAI,CAAC,UAAU,EAAE;4BAClC,OAAO,EAAE,IAAI,CAAC,OAAO;4BACrB,MAAM,EAAE,iBAAiB,QAAQ,CAAC,IAAI,EAAE;4BACxC,cAAc,EAAE,IAAI,CAAC,SAAS;4BAC9B,mBAAmB,EAAE,IAAI,CAAC,MAAM;4BAChC,kBAAkB,EAAE,IAAI,CAAC,OAAO,IAAI,EAAE;4BACtC,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,CAAC,IAAI,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE;yBAC9D,CAAA;wBAED,oDAAoD;wBACpD,KAAK,MAAM,GAAG,IAAI,eAAe,CAAC,kBAAkB,EAAE,CAAC;4BACtD,iIAAiI;4BACjI,OAAO,eAAe,CAAC,UAAU,CAAC,GAAG,CAAC,CAAA;wBACvC,CAAC;wBAED,OAAO,CAAC,IAAI,CAAC,eAAe,CAAC,CAAA;wBAE7B,IAAI,EAAE,CAAA;oBACP,CAAC;yBAAM,CAAC;wBACP,OAAO,EAAE,CAAA;oBACV,CAAC;gBACF,CAAC;YACF,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBAChB,OAAO,EAAE,CAAA;gBACT,MAAM,EAAE,CAAC,YAAY,IAAI,CAAC,SAAS,KAAM,KAAe,CAAC,OAAO,EAAE,CAAC,CAAA;YACpE,CAAC;YAED,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,CAAC;gBACxB,MAAM,EAAE,CAAC,eAAe,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,MAAM,UAAU,IAAI,aAAa,OAAO,aAAa,OAAO,EAAE,CAAC,CAAA;YACvG,CAAC;QACF,CAAC;IACF,CAAC,CAAC,CAAA;IAEF,MAAM,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;IAE1B,CAAC;QACA,YAAY,GAAG,GAAG,mBAAmB,CAAC,UAAU,CAAC,CAAA;QAEjD,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;YAC3B,MAAM,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAA;QACrC,CAAC;IACF,CAAC;IAED,MAAM,EAAE,CAAC,iBAAiB,CAAC,CAAA;IAC3B,MAAM,EAAE,CAAC,qBAAqB,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;IAC7C,MAAM,EAAE,CAAC,qBAAqB,IAAI,EAAE,CAAC,CAAA;IACrC,MAAM,EAAE,CAAC,mCAAmC,OAAO,EAAE,CAAC,CAAA;IACtD,MAAM,EAAE,CAAC,sBAAsB,OAAO,EAAE,CAAC,CAAA;IACzC,MAAM,EAAE,CAAC,UAAU,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,MAAM,GAAG,QAAQ,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAA;IACjG,MAAM,EAAE,CAAC,qBAAqB,UAAU,EAAE,CAAC,CAAA;IAE3C,OAAO,EAAE,cAAc,EAAE,KAAK,CAAC,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,CAAA;AAC5E,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"golden-promote.d.ts","sourceRoot":"","sources":["../../lib/tools/golden-promote.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AA4BH,MAAM,WAAW,YAAY;IAC5B,YAAY,EAAE,MAAM,CAAA;IACpB,WAAW,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,eAAe,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,CAAA;IAClH,IAAI,EAAE,MAAM,CAAA;IACZ,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CAClC;AAED,MAAM,WAAW,oBAAoB;IACpC;;OAEG;IACH,KAAK,EAAE,MAAM,CAAA;IACb;;OAEG;IACH,MAAM,EAAE,MAAM,CAAA;IACd;;OAEG;IACH,KAAK,CAAC,EAAE,MAAM,CAAA;IACd;;OAEG;IACH,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB;;OAEG;IACH,SAAS,CAAC,EAAE,OAAO,CAAA;IACnB;;OAEG;IACH,MAAM,CAAC,EAAE,OAAO,CAAA;CAChB;AA0DD,wBAAsB,aAAa,CAClC,OAAO,EAAE,oBAAoB,EAC7B,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,GAC7B,OAAO,CAAC,YAAY,CAAC,CAwLvB"}
|
|
@@ -0,0 +1,240 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* @copyright Sister Software
|
|
3
|
+
* @license AGPL-3.0
|
|
4
|
+
* @author Teffen Ellis, et al.
|
|
5
|
+
*
|
|
6
|
+
* Promote LLM-synthesized golden-set candidates into a versioned golden dir, with
|
|
7
|
+
* human-typed-likelihood filters + dedup. Companion to `golden-expand.ts`.
|
|
8
|
+
*
|
|
9
|
+
* ## What it does
|
|
10
|
+
*
|
|
11
|
+
* 1. Reads a candidates JSONL from `data/eval/golden/candidates/`
|
|
12
|
+
* 2. Reads the previous-version golden dir for forward-copy + dedup base
|
|
13
|
+
* 3. Applies filters that drop candidates unlikely to be human-typed:
|
|
14
|
+
*
|
|
15
|
+
* - Components-glued-without-commas (5+ components but <2 separators → freeform jumble)
|
|
16
|
+
* - Postcode-at-start with many other components (US/UK conventions put postcode last; FR puts it
|
|
17
|
+
* before locality only)
|
|
18
|
+
* - Suspicious-token signals (unmatched brackets, control chars, etc.)
|
|
19
|
+
* 4. Dedupes by normalized raw (case-insensitive, whitespace-collapsed)
|
|
20
|
+
* 5. Splits by country (US/FR/other) and writes `data/eval/golden/v<X.Y.Z>/{us,fr,other}.jsonl`
|
|
21
|
+
* 6. Forward-copies the prior version's adversarial.jsonl + README as-is
|
|
22
|
+
* 7. Writes MANIFEST.json with sha256 of each output file
|
|
23
|
+
*
|
|
24
|
+
* ## Usage
|
|
25
|
+
*
|
|
26
|
+
* ```sh
|
|
27
|
+
* mailwoman corpus golden promote \
|
|
28
|
+
* --input data/eval/golden/candidates/expand-20260518-162627.jsonl \
|
|
29
|
+
* --bump-to v0.1.1 \
|
|
30
|
+
* --prior v0.1.0
|
|
31
|
+
* ```
|
|
32
|
+
*/
|
|
33
|
+
import { readDirectory, pathExists } from "@mailwoman/core/fs/readers";
|
|
34
|
+
import { copyFileTo, makeDirectories, writeLocalJSONFile } from "@mailwoman/core/fs/writers";
|
|
35
|
+
import { sha256File } from "@mailwoman/core/hash";
|
|
36
|
+
import { foldCaseWhitespace } from "@mailwoman/normalize/fold";
|
|
37
|
+
import { join } from "path-ts";
|
|
38
|
+
import { createNewlineWriter, JSONSpliterator } from "spliterator";
|
|
39
|
+
// ── Types ──────────────────────────────────────────────────────────────────
|
|
40
|
+
/**
|
|
41
|
+
* Components a candidate needs before it is rich enough to be worth a golden-set slot.
|
|
42
|
+
*/
|
|
43
|
+
const MIN_GOLDEN_COMPONENTS = 5;
|
|
44
|
+
/**
|
|
45
|
+
* Components an existing entry must still carry to survive promotion.
|
|
46
|
+
*/
|
|
47
|
+
const MIN_PROMOTABLE_COMPONENTS = 4;
|
|
48
|
+
/**
|
|
49
|
+
* Quote-delimited segments above which the raw line is malformed rather than merely quoted.
|
|
50
|
+
*/
|
|
51
|
+
const MAX_QUOTE_SEGMENTS = 3;
|
|
52
|
+
// ── Filters ───────────────────────────────────────────────────────────────
|
|
53
|
+
/**
|
|
54
|
+
* Heuristic: an address with 5+ components but fewer than 2 separators (commas/newlines/dashes) is most likely glued
|
|
55
|
+
* together rather than human-typed.
|
|
56
|
+
*/
|
|
57
|
+
function isComponentsGlued(entry) {
|
|
58
|
+
const componentCount = Object.keys(entry.components).length;
|
|
59
|
+
if (componentCount < MIN_GOLDEN_COMPONENTS)
|
|
60
|
+
return false;
|
|
61
|
+
const separators = (entry.raw.match(/[,\n;]/g) ?? []).length;
|
|
62
|
+
return separators < 2;
|
|
63
|
+
}
|
|
64
|
+
/**
|
|
65
|
+
* Heuristic: in US/UK conventions, postcode goes at the END of the address. If postcode appears in the first third of a
|
|
66
|
+
* multi-component raw AND there are 4+ components, the LLM probably over-aggressively reordered. FR is exempt (postcode
|
|
67
|
+
* often precedes locality there).
|
|
68
|
+
*/
|
|
69
|
+
function isPostcodeBadlyLeading(entry) {
|
|
70
|
+
if (Object.keys(entry.components).length < MIN_PROMOTABLE_COMPONENTS)
|
|
71
|
+
return false;
|
|
72
|
+
if (entry.country === "FR" || entry.country === "France")
|
|
73
|
+
return false;
|
|
74
|
+
const postcode = entry.components.postcode;
|
|
75
|
+
if (!postcode)
|
|
76
|
+
return false;
|
|
77
|
+
const postcodeIdx = entry.raw.indexOf(postcode);
|
|
78
|
+
if (postcodeIdx === -1)
|
|
79
|
+
return false;
|
|
80
|
+
return postcodeIdx < entry.raw.length / 3;
|
|
81
|
+
}
|
|
82
|
+
/**
|
|
83
|
+
* Heuristic: catch-all for visually-bad outputs — unmatched brackets, control chars, suspicious punctuation that
|
|
84
|
+
* suggests the LLM emitted markup instead of an address.
|
|
85
|
+
*/
|
|
86
|
+
function isSuspicious(entry) {
|
|
87
|
+
const raw = entry.raw;
|
|
88
|
+
// eslint-disable-next-line no-control-regex
|
|
89
|
+
if (/[\u0000-\u001F\u007F]/.test(raw))
|
|
90
|
+
return true;
|
|
91
|
+
const openBrackets = (raw.match(/[[({<]/g) ?? []).length;
|
|
92
|
+
const closeBrackets = (raw.match(/[\])}>]/g) ?? []).length;
|
|
93
|
+
if (openBrackets !== closeBrackets)
|
|
94
|
+
return true;
|
|
95
|
+
if (raw.split('"').length > MAX_QUOTE_SEGMENTS)
|
|
96
|
+
return true;
|
|
97
|
+
// too many quote marks
|
|
98
|
+
return false;
|
|
99
|
+
}
|
|
100
|
+
// ── Main ───────────────────────────────────────────────────────────────────
|
|
101
|
+
export async function promoteGolden(options, report) {
|
|
102
|
+
const prior = options.prior ?? "v0.1.0";
|
|
103
|
+
const goldenRoot = options.goldenRoot ?? "data/eval/golden";
|
|
104
|
+
const applyFilters = !options.noFilters;
|
|
105
|
+
const dryRun = options.dryRun ?? false;
|
|
106
|
+
report?.(`reading candidates: ${options.input}`);
|
|
107
|
+
// The candidates file is the REQUIRED input, so its absence must not read as "zero candidates" —
|
|
108
|
+
// that promotes the prior version unchanged and reports success. `JSONSpliterator.fromAsync`
|
|
109
|
+
// reports a missing path as a bare `TypeError` naming its own internals, so name the path here.
|
|
110
|
+
if (!(await pathExists(options.input))) {
|
|
111
|
+
throw new Error(`Candidates file not found: ${options.input}`);
|
|
112
|
+
}
|
|
113
|
+
const candidates = await Array.fromAsync(JSONSpliterator.fromAsync(options.input));
|
|
114
|
+
report?.(` ${candidates.length} candidates loaded`);
|
|
115
|
+
// Forward-copy base: existing entries from the prior golden version go forward verbatim,
|
|
116
|
+
// and we dedupe new candidates against them so v_new = v_old ∪ accepted_candidates.
|
|
117
|
+
const priorDir = join(goldenRoot, prior);
|
|
118
|
+
const priorEntries = [];
|
|
119
|
+
const seenNormalized = new Set();
|
|
120
|
+
if (await pathExists(priorDir)) {
|
|
121
|
+
for (const f of (await readDirectory(priorDir)).filter((n) => n.endsWith(".jsonl"))) {
|
|
122
|
+
const country = f.replace(".jsonl", "").toUpperCase();
|
|
123
|
+
const entries = await Array.fromAsync(JSONSpliterator.fromAsync(join(priorDir, f)));
|
|
124
|
+
priorEntries.push({ country, entries });
|
|
125
|
+
for (const e of entries) {
|
|
126
|
+
seenNormalized.add(foldCaseWhitespace(e.raw));
|
|
127
|
+
}
|
|
128
|
+
report?.(` prior ${country}: ${entries.length} entries (forward-copy base)`);
|
|
129
|
+
}
|
|
130
|
+
}
|
|
131
|
+
else {
|
|
132
|
+
report?.(` ⚠ prior dir ${priorDir} not found — starting fresh`);
|
|
133
|
+
}
|
|
134
|
+
// Filter pass
|
|
135
|
+
const stats = {
|
|
136
|
+
candidatesIn: candidates.length,
|
|
137
|
+
filteredOut: { glued: 0, postcodeLeading: 0, suspicious: 0, duplicate: 0, forwardDup: 0 },
|
|
138
|
+
kept: 0,
|
|
139
|
+
perCountry: {},
|
|
140
|
+
};
|
|
141
|
+
const accepted = [];
|
|
142
|
+
const seenInBatch = new Set();
|
|
143
|
+
for (const cand of candidates) {
|
|
144
|
+
const norm = foldCaseWhitespace(cand.raw);
|
|
145
|
+
// Dedup pass 1: against prior versioned golden
|
|
146
|
+
if (seenNormalized.has(norm)) {
|
|
147
|
+
stats.filteredOut.forwardDup++;
|
|
148
|
+
continue;
|
|
149
|
+
}
|
|
150
|
+
// Dedup pass 2: against this batch
|
|
151
|
+
if (seenInBatch.has(norm)) {
|
|
152
|
+
stats.filteredOut.duplicate++;
|
|
153
|
+
continue;
|
|
154
|
+
}
|
|
155
|
+
if (applyFilters) {
|
|
156
|
+
if (isComponentsGlued(cand)) {
|
|
157
|
+
stats.filteredOut.glued++;
|
|
158
|
+
continue;
|
|
159
|
+
}
|
|
160
|
+
if (isPostcodeBadlyLeading(cand)) {
|
|
161
|
+
stats.filteredOut.postcodeLeading++;
|
|
162
|
+
continue;
|
|
163
|
+
}
|
|
164
|
+
if (isSuspicious(cand)) {
|
|
165
|
+
stats.filteredOut.suspicious++;
|
|
166
|
+
continue;
|
|
167
|
+
}
|
|
168
|
+
}
|
|
169
|
+
accepted.push(cand);
|
|
170
|
+
seenInBatch.add(norm);
|
|
171
|
+
stats.kept++;
|
|
172
|
+
const country = cand.country || "OTHER";
|
|
173
|
+
stats.perCountry[country] = (stats.perCountry[country] ?? 0) + 1;
|
|
174
|
+
}
|
|
175
|
+
// Bucket per country, including forward-copied entries
|
|
176
|
+
const buckets = new Map();
|
|
177
|
+
for (const { country, entries } of priorEntries) {
|
|
178
|
+
// Existing files keyed by filename uppercase (us.jsonl → US, adversarial.jsonl → ADVERSARIAL)
|
|
179
|
+
buckets.set(country, [...entries]);
|
|
180
|
+
}
|
|
181
|
+
for (const cand of accepted) {
|
|
182
|
+
const key = (cand.country || "OTHER").toUpperCase();
|
|
183
|
+
if (!buckets.has(key)) {
|
|
184
|
+
buckets.set(key, []);
|
|
185
|
+
}
|
|
186
|
+
buckets.get(key).push(cand);
|
|
187
|
+
}
|
|
188
|
+
// Output
|
|
189
|
+
const outDir = join(goldenRoot, options.bumpTo);
|
|
190
|
+
report?.(`=== plan ===`);
|
|
191
|
+
report?.(`output dir: ${outDir}${dryRun ? " (dry-run)" : ""}`);
|
|
192
|
+
for (const [key, entries] of buckets) {
|
|
193
|
+
report?.(` ${key.toLowerCase()}.jsonl: ${entries.length} entries`);
|
|
194
|
+
}
|
|
195
|
+
report?.(`=== filter stats ===`);
|
|
196
|
+
report?.(`candidates in: ${stats.candidatesIn}`);
|
|
197
|
+
report?.(` filtered (glued): ${stats.filteredOut.glued}`);
|
|
198
|
+
report?.(` filtered (postcode-lead): ${stats.filteredOut.postcodeLeading}`);
|
|
199
|
+
report?.(` filtered (suspicious): ${stats.filteredOut.suspicious}`);
|
|
200
|
+
report?.(` filtered (dup-in-batch): ${stats.filteredOut.duplicate}`);
|
|
201
|
+
report?.(` filtered (dup-vs-prior): ${stats.filteredOut.forwardDup}`);
|
|
202
|
+
report?.(` kept: ${stats.kept}`);
|
|
203
|
+
report?.(`per-country (kept):`);
|
|
204
|
+
for (const [c, n] of Object.entries(stats.perCountry).toSorted((a, b) => b[1] - a[1])) {
|
|
205
|
+
report?.(` ${c}: ${n}`);
|
|
206
|
+
}
|
|
207
|
+
if (dryRun) {
|
|
208
|
+
report?.(`(dry-run: no files written)`);
|
|
209
|
+
return stats;
|
|
210
|
+
}
|
|
211
|
+
await makeDirectories(outDir);
|
|
212
|
+
const manifest = {
|
|
213
|
+
promoted_at: new Date().toISOString(),
|
|
214
|
+
from: options.input,
|
|
215
|
+
from_sha256: await sha256File(options.input),
|
|
216
|
+
files: {},
|
|
217
|
+
};
|
|
218
|
+
for (const [key, entries] of buckets) {
|
|
219
|
+
const filename = `${key.toLowerCase()}.jsonl`;
|
|
220
|
+
const path = join(outDir, filename);
|
|
221
|
+
{
|
|
222
|
+
await using out = createNewlineWriter(path);
|
|
223
|
+
for (const entry of entries) {
|
|
224
|
+
await out.write(JSON.stringify(entry));
|
|
225
|
+
}
|
|
226
|
+
}
|
|
227
|
+
manifest.files[filename] = { entries: entries.length, sha256: await sha256File(path) };
|
|
228
|
+
}
|
|
229
|
+
// Forward-copy non-.jsonl files (README.md, etc.) from prior
|
|
230
|
+
if (await pathExists(priorDir)) {
|
|
231
|
+
for (const f of (await readDirectory(priorDir)).filter((n) => !n.endsWith(".jsonl"))) {
|
|
232
|
+
await copyFileTo(join(priorDir, f), join(outDir, f));
|
|
233
|
+
report?.(` forward-copied: ${f}`);
|
|
234
|
+
}
|
|
235
|
+
}
|
|
236
|
+
await writeLocalJSONFile(manifest, join(outDir, "MANIFEST.json"));
|
|
237
|
+
report?.(`✓ promoted to ${outDir}`);
|
|
238
|
+
return stats;
|
|
239
|
+
}
|
|
240
|
+
//# sourceMappingURL=golden-promote.js.map
|