@mailwoman/corpus 8.0.0 → 8.2.0
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/out/src/adapter.d.ts +15 -5
- package/out/src/adapter.d.ts.map +1 -1
- package/out/src/adapter.js +11 -9
- package/out/src/adapter.js.map +1 -1
- package/out/src/adapters/ban/adapter.d.ts +7 -0
- package/out/src/adapters/ban/adapter.d.ts.map +1 -1
- package/out/src/adapters/ban/adapter.js +9 -2
- package/out/src/adapters/ban/adapter.js.map +1 -1
- package/out/src/adapters/ban/street-decompose.d.ts.map +1 -1
- package/out/src/adapters/ban/street-decompose.js +5 -1
- package/out/src/adapters/ban/street-decompose.js.map +1 -1
- package/out/src/adapters/fcc-bdc/adapter.d.ts +14 -1
- package/out/src/adapters/fcc-bdc/adapter.d.ts.map +1 -1
- package/out/src/adapters/fcc-bdc/adapter.js +15 -2
- package/out/src/adapters/fcc-bdc/adapter.js.map +1 -1
- package/out/src/adapters/geonames/adapter.d.ts +11 -0
- package/out/src/adapters/geonames/adapter.d.ts.map +1 -1
- package/out/src/adapters/geonames/adapter.js +24 -5
- package/out/src/adapters/geonames/adapter.js.map +1 -1
- package/out/src/adapters/geonames-postal/adapter.d.ts +11 -0
- package/out/src/adapters/geonames-postal/adapter.d.ts.map +1 -1
- package/out/src/adapters/geonames-postal/adapter.js +14 -1
- package/out/src/adapters/geonames-postal/adapter.js.map +1 -1
- package/out/src/adapters/gnaf/adapter.d.ts +10 -1
- package/out/src/adapters/gnaf/adapter.d.ts.map +1 -1
- package/out/src/adapters/gnaf/adapter.js +11 -2
- package/out/src/adapters/gnaf/adapter.js.map +1 -1
- package/out/src/adapters/gnaf/assemble.d.ts +18 -6
- package/out/src/adapters/gnaf/assemble.d.ts.map +1 -1
- package/out/src/adapters/gnaf/assemble.js +16 -6
- package/out/src/adapters/gnaf/assemble.js.map +1 -1
- package/out/src/adapters/index.d.ts.map +1 -1
- package/out/src/adapters/index.js.map +1 -1
- package/out/src/adapters/openaddresses/adapter.d.ts +11 -0
- package/out/src/adapters/openaddresses/adapter.d.ts.map +1 -1
- package/out/src/adapters/openaddresses/adapter.js +18 -3
- package/out/src/adapters/openaddresses/adapter.js.map +1 -1
- package/out/src/adapters/overture/adapter.d.ts +11 -0
- package/out/src/adapters/overture/adapter.d.ts.map +1 -1
- package/out/src/adapters/overture/adapter.js +12 -1
- package/out/src/adapters/overture/adapter.js.map +1 -1
- package/out/src/adapters/state-hi-schools/adapter.d.ts +11 -0
- package/out/src/adapters/state-hi-schools/adapter.d.ts.map +1 -1
- package/out/src/adapters/state-hi-schools/adapter.js +11 -0
- package/out/src/adapters/state-hi-schools/adapter.js.map +1 -1
- package/out/src/adapters/state-ia-contractors/adapter.d.ts +11 -0
- package/out/src/adapters/state-ia-contractors/adapter.d.ts.map +1 -1
- package/out/src/adapters/state-ia-contractors/adapter.js +11 -0
- package/out/src/adapters/state-ia-contractors/adapter.js.map +1 -1
- package/out/src/adapters/state-ny-notaries/adapter.d.ts +11 -0
- package/out/src/adapters/state-ny-notaries/adapter.d.ts.map +1 -1
- package/out/src/adapters/state-ny-notaries/adapter.js +11 -0
- package/out/src/adapters/state-ny-notaries/adapter.js.map +1 -1
- package/out/src/adapters/state-tx-notaries/adapter.d.ts +11 -0
- package/out/src/adapters/state-tx-notaries/adapter.d.ts.map +1 -1
- package/out/src/adapters/state-tx-notaries/adapter.js +15 -2
- package/out/src/adapters/state-tx-notaries/adapter.js.map +1 -1
- package/out/src/adapters/synth-po-box/adapter.d.ts +11 -0
- package/out/src/adapters/synth-po-box/adapter.d.ts.map +1 -1
- package/out/src/adapters/synth-po-box/adapter.js +13 -2
- package/out/src/adapters/synth-po-box/adapter.js.map +1 -1
- package/out/src/adapters/tiger/adapter.d.ts +14 -1
- package/out/src/adapters/tiger/adapter.d.ts.map +1 -1
- package/out/src/adapters/tiger/adapter.js +19 -4
- package/out/src/adapters/tiger/adapter.js.map +1 -1
- package/out/src/adapters/tiger/street-decompose.d.ts.map +1 -1
- package/out/src/adapters/tiger/street-decompose.js.map +1 -1
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.d.ts +11 -0
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.d.ts.map +1 -1
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.js +12 -1
- package/out/src/adapters/usgov-hrsa-fqhc/adapter.js.map +1 -1
- package/out/src/adapters/usgov-imls-pls/adapter.d.ts +11 -0
- package/out/src/adapters/usgov-imls-pls/adapter.d.ts.map +1 -1
- package/out/src/adapters/usgov-imls-pls/adapter.js +11 -0
- package/out/src/adapters/usgov-imls-pls/adapter.js.map +1 -1
- package/out/src/adapters/usgov-irs-bmf/adapter.d.ts +11 -0
- package/out/src/adapters/usgov-irs-bmf/adapter.d.ts.map +1 -1
- package/out/src/adapters/usgov-irs-bmf/adapter.js +17 -2
- package/out/src/adapters/usgov-irs-bmf/adapter.js.map +1 -1
- package/out/src/adapters/usgov-nad/adapter.d.ts +11 -0
- package/out/src/adapters/usgov-nad/adapter.d.ts.map +1 -1
- package/out/src/adapters/usgov-nad/adapter.js +12 -1
- package/out/src/adapters/usgov-nad/adapter.js.map +1 -1
- package/out/src/adapters/usgov-nppes/adapter.d.ts +11 -0
- package/out/src/adapters/usgov-nppes/adapter.d.ts.map +1 -1
- package/out/src/adapters/usgov-nppes/adapter.js +11 -0
- package/out/src/adapters/usgov-nppes/adapter.js.map +1 -1
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.d.ts +11 -0
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.d.ts.map +1 -1
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.js +15 -2
- package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.js.map +1 -1
- package/out/src/adapters/wof-admin-jp/adapter.d.ts +7 -0
- package/out/src/adapters/wof-admin-jp/adapter.d.ts.map +1 -1
- package/out/src/adapters/wof-admin-jp/adapter.js +18 -3
- package/out/src/adapters/wof-admin-jp/adapter.js.map +1 -1
- package/out/src/adapters/wof-admin-json/adapter.d.ts +13 -3
- package/out/src/adapters/wof-admin-json/adapter.d.ts.map +1 -1
- package/out/src/adapters/wof-admin-json/adapter.js +15 -5
- package/out/src/adapters/wof-admin-json/adapter.js.map +1 -1
- package/out/src/adapters/wof-postalcode-json/adapter.d.ts +7 -0
- package/out/src/adapters/wof-postalcode-json/adapter.d.ts.map +1 -1
- package/out/src/adapters/wof-postalcode-json/adapter.js +9 -2
- package/out/src/adapters/wof-postalcode-json/adapter.js.map +1 -1
- package/out/src/align.d.ts +12 -4
- package/out/src/align.d.ts.map +1 -1
- package/out/src/align.js +9 -5
- package/out/src/align.js.map +1 -1
- package/out/src/build.d.ts +24 -8
- package/out/src/build.d.ts.map +1 -1
- package/out/src/build.js +3 -3
- package/out/src/build.js.map +1 -1
- package/out/src/codex/us-fips-state.d.ts +6 -2
- package/out/src/codex/us-fips-state.d.ts.map +1 -1
- package/out/src/codex/us-fips-state.js +3 -1
- package/out/src/codex/us-fips-state.js.map +1 -1
- package/out/src/golden.d.ts +21 -7
- package/out/src/golden.d.ts.map +1 -1
- package/out/src/golden.js +18 -10
- package/out/src/golden.js.map +1 -1
- package/out/src/license.d.ts +3 -1
- package/out/src/license.d.ts.map +1 -1
- package/out/src/license.js +4 -2
- package/out/src/license.js.map +1 -1
- package/out/src/parquet-wrapper/reader.d.ts +4 -2
- package/out/src/parquet-wrapper/reader.d.ts.map +1 -1
- package/out/src/parquet-wrapper/reader.js +5 -4
- package/out/src/parquet-wrapper/reader.js.map +1 -1
- package/out/src/parquet-wrapper/schema.d.ts +14 -10
- package/out/src/parquet-wrapper/schema.d.ts.map +1 -1
- package/out/src/parquet-wrapper/schema.js +6 -2
- package/out/src/parquet-wrapper/schema.js.map +1 -1
- package/out/src/parquet-wrapper/writer.d.ts +12 -4
- package/out/src/parquet-wrapper/writer.d.ts.map +1 -1
- package/out/src/parquet-wrapper/writer.js +12 -4
- package/out/src/parquet-wrapper/writer.js.map +1 -1
- package/out/src/parquet.d.ts +21 -7
- package/out/src/parquet.d.ts.map +1 -1
- package/out/src/parquet.js +15 -4
- package/out/src/parquet.js.map +1 -1
- package/out/src/runner.d.ts +39 -13
- package/out/src/runner.d.ts.map +1 -1
- package/out/src/runner.js +7 -3
- package/out/src/runner.js.map +1 -1
- package/out/src/shard-recipes/anchor-absorption.d.ts +4 -0
- package/out/src/shard-recipes/anchor-absorption.d.ts.map +1 -1
- package/out/src/shard-recipes/anchor-absorption.js +14 -5
- package/out/src/shard-recipes/anchor-absorption.js.map +1 -1
- package/out/src/shard-recipes/boundary-stress.d.ts +4 -0
- package/out/src/shard-recipes/boundary-stress.d.ts.map +1 -1
- package/out/src/shard-recipes/boundary-stress.js +13 -7
- package/out/src/shard-recipes/boundary-stress.js.map +1 -1
- package/out/src/shard-recipes/country-balanced.d.ts +4 -0
- package/out/src/shard-recipes/country-balanced.d.ts.map +1 -1
- package/out/src/shard-recipes/country-balanced.js +39 -11
- package/out/src/shard-recipes/country-balanced.js.map +1 -1
- package/out/src/shard-recipes/cz-pcfirst-preposition.d.ts +4 -0
- package/out/src/shard-recipes/cz-pcfirst-preposition.d.ts.map +1 -1
- package/out/src/shard-recipes/cz-pcfirst-preposition.js +4 -0
- package/out/src/shard-recipes/cz-pcfirst-preposition.js.map +1 -1
- package/out/src/shard-recipes/fr-admin-split.d.ts +4 -0
- package/out/src/shard-recipes/fr-admin-split.d.ts.map +1 -1
- package/out/src/shard-recipes/fr-admin-split.js +14 -3
- package/out/src/shard-recipes/fr-admin-split.js.map +1 -1
- package/out/src/shard-recipes/fr-bare-street.d.ts +4 -0
- package/out/src/shard-recipes/fr-bare-street.d.ts.map +1 -1
- package/out/src/shard-recipes/fr-bare-street.js +4 -0
- package/out/src/shard-recipes/fr-bare-street.js.map +1 -1
- package/out/src/shard-recipes/fr-fragment.d.ts +4 -0
- package/out/src/shard-recipes/fr-fragment.d.ts.map +1 -1
- package/out/src/shard-recipes/fr-fragment.js +25 -8
- package/out/src/shard-recipes/fr-fragment.js.map +1 -1
- package/out/src/shard-recipes/fr-lieudit.d.ts +4 -0
- package/out/src/shard-recipes/fr-lieudit.d.ts.map +1 -1
- package/out/src/shard-recipes/fr-lieudit.js +21 -8
- package/out/src/shard-recipes/fr-lieudit.js.map +1 -1
- package/out/src/shard-recipes/fr-order.d.ts +4 -0
- package/out/src/shard-recipes/fr-order.d.ts.map +1 -1
- package/out/src/shard-recipes/fr-order.js +27 -8
- package/out/src/shard-recipes/fr-order.js.map +1 -1
- package/out/src/shard-recipes/german.d.ts +4 -0
- package/out/src/shard-recipes/german.d.ts.map +1 -1
- package/out/src/shard-recipes/german.js +16 -6
- package/out/src/shard-recipes/german.js.map +1 -1
- package/out/src/shard-recipes/house-venue.d.ts +4 -0
- package/out/src/shard-recipes/house-venue.d.ts.map +1 -1
- package/out/src/shard-recipes/house-venue.js +4 -0
- package/out/src/shard-recipes/house-venue.js.map +1 -1
- package/out/src/shard-recipes/index.d.ts +9 -3
- package/out/src/shard-recipes/index.d.ts.map +1 -1
- package/out/src/shard-recipes/index.js +12 -4
- package/out/src/shard-recipes/index.js.map +1 -1
- package/out/src/shard-recipes/intersection.d.ts +4 -0
- package/out/src/shard-recipes/intersection.d.ts.map +1 -1
- package/out/src/shard-recipes/intersection.js +27 -13
- package/out/src/shard-recipes/intersection.js.map +1 -1
- package/out/src/shard-recipes/locale.d.ts +4 -0
- package/out/src/shard-recipes/locale.d.ts.map +1 -1
- package/out/src/shard-recipes/locale.js +9 -4
- package/out/src/shard-recipes/locale.js.map +1 -1
- package/out/src/shard-recipes/nl-postcode.d.ts +4 -0
- package/out/src/shard-recipes/nl-postcode.d.ts.map +1 -1
- package/out/src/shard-recipes/nl-postcode.js +8 -2
- package/out/src/shard-recipes/nl-postcode.js.map +1 -1
- package/out/src/shard-recipes/no-fragment.d.ts +4 -0
- package/out/src/shard-recipes/no-fragment.d.ts.map +1 -1
- package/out/src/shard-recipes/no-fragment.js +8 -2
- package/out/src/shard-recipes/no-fragment.js.map +1 -1
- package/out/src/shard-recipes/no-street-led.d.ts +4 -0
- package/out/src/shard-recipes/no-street-led.d.ts.map +1 -1
- package/out/src/shard-recipes/no-street-led.js +5 -1
- package/out/src/shard-recipes/no-street-led.js.map +1 -1
- package/out/src/shard-recipes/no-street.d.ts +4 -0
- package/out/src/shard-recipes/no-street.d.ts.map +1 -1
- package/out/src/shard-recipes/no-street.js +4 -0
- package/out/src/shard-recipes/no-street.js.map +1 -1
- package/out/src/shard-recipes/po-box-cedex.d.ts +4 -0
- package/out/src/shard-recipes/po-box-cedex.d.ts.map +1 -1
- package/out/src/shard-recipes/po-box-cedex.js +99 -43
- package/out/src/shard-recipes/po-box-cedex.js.map +1 -1
- package/out/src/shard-recipes/po-box.d.ts +4 -0
- package/out/src/shard-recipes/po-box.d.ts.map +1 -1
- package/out/src/shard-recipes/po-box.js +4 -0
- package/out/src/shard-recipes/po-box.js.map +1 -1
- package/out/src/shard-recipes/scaffold.d.ts +48 -16
- package/out/src/shard-recipes/scaffold.d.ts.map +1 -1
- package/out/src/shard-recipes/scaffold.js +10 -6
- package/out/src/shard-recipes/scaffold.js.map +1 -1
- package/out/src/shard-recipes/si-bare-village.d.ts +4 -0
- package/out/src/shard-recipes/si-bare-village.d.ts.map +1 -1
- package/out/src/shard-recipes/si-bare-village.js +4 -0
- package/out/src/shard-recipes/si-bare-village.js.map +1 -1
- package/out/src/shard-recipes/street-affix.d.ts +4 -0
- package/out/src/shard-recipes/street-affix.d.ts.map +1 -1
- package/out/src/shard-recipes/street-affix.js +29 -10
- package/out/src/shard-recipes/street-affix.js.map +1 -1
- package/out/src/shard-recipes/street-bare.d.ts +4 -0
- package/out/src/shard-recipes/street-bare.d.ts.map +1 -1
- package/out/src/shard-recipes/street-bare.js +4 -0
- package/out/src/shard-recipes/street-bare.js.map +1 -1
- package/out/src/shard-recipes/street.d.ts +4 -0
- package/out/src/shard-recipes/street.d.ts.map +1 -1
- package/out/src/shard-recipes/street.js +4 -0
- package/out/src/shard-recipes/street.js.map +1 -1
- package/out/src/shard-recipes/unit.d.ts +4 -0
- package/out/src/shard-recipes/unit.d.ts.map +1 -1
- package/out/src/shard-recipes/unit.js +45 -14
- package/out/src/shard-recipes/unit.js.map +1 -1
- package/out/src/split.d.ts +18 -6
- package/out/src/split.d.ts.map +1 -1
- package/out/src/split.js +4 -2
- package/out/src/split.js.map +1 -1
- package/out/src/synthesize-anchor-absorption.d.ts +7 -1
- package/out/src/synthesize-anchor-absorption.d.ts.map +1 -1
- package/out/src/synthesize-anchor-absorption.js +28 -14
- package/out/src/synthesize-anchor-absorption.js.map +1 -1
- package/out/src/synthesize-boundary-stress.d.ts +3 -1
- package/out/src/synthesize-boundary-stress.d.ts.map +1 -1
- package/out/src/synthesize-boundary-stress.js +35 -28
- package/out/src/synthesize-boundary-stress.js.map +1 -1
- package/out/src/synthesize-german.d.ts +15 -5
- package/out/src/synthesize-german.d.ts.map +1 -1
- package/out/src/synthesize-german.js +9 -3
- package/out/src/synthesize-german.js.map +1 -1
- package/out/src/synthesize-house-venue.d.ts.map +1 -1
- package/out/src/synthesize-house-venue.js +15 -15
- package/out/src/synthesize-house-venue.js.map +1 -1
- package/out/src/synthesize-intersection.d.ts +9 -3
- package/out/src/synthesize-intersection.d.ts.map +1 -1
- package/out/src/synthesize-intersection.js +25 -6
- package/out/src/synthesize-intersection.js.map +1 -1
- package/out/src/synthesize-no-street.d.ts +3 -1
- package/out/src/synthesize-no-street.d.ts.map +1 -1
- package/out/src/synthesize-no-street.js +5 -6
- package/out/src/synthesize-no-street.js.map +1 -1
- package/out/src/synthesize-po-box.d.ts +15 -5
- package/out/src/synthesize-po-box.d.ts.map +1 -1
- package/out/src/synthesize-po-box.js +21 -7
- package/out/src/synthesize-po-box.js.map +1 -1
- package/out/src/synthesize-street.d.ts +3 -1
- package/out/src/synthesize-street.d.ts.map +1 -1
- package/out/src/synthesize-street.js +6 -1
- package/out/src/synthesize-street.js.map +1 -1
- package/out/src/synthesize.d.ts +39 -13
- package/out/src/synthesize.d.ts.map +1 -1
- package/out/src/synthesize.js +66 -36
- package/out/src/synthesize.js.map +1 -1
- package/out/src/tokenize.d.ts +15 -5
- package/out/src/tokenize.d.ts.map +1 -1
- package/out/src/tokenize.js.map +1 -1
- package/out/src/tools/align-shard.d.ts.map +1 -1
- package/out/src/tools/align-shard.js +3 -1
- package/out/src/tools/align-shard.js.map +1 -1
- package/out/src/tools/audit.d.ts.map +1 -1
- package/out/src/tools/audit.js +20 -8
- package/out/src/tools/audit.js.map +1 -1
- package/out/src/tools/corpus-stats.d.ts.map +1 -1
- package/out/src/tools/corpus-stats.js.map +1 -1
- package/out/src/tools/fetch/ban.d.ts.map +1 -1
- package/out/src/tools/fetch/ban.js +17 -10
- package/out/src/tools/fetch/ban.js.map +1 -1
- package/out/src/tools/fetch/download.d.ts +30 -10
- package/out/src/tools/fetch/download.d.ts.map +1 -1
- package/out/src/tools/fetch/download.js +26 -6
- package/out/src/tools/fetch/download.js.map +1 -1
- package/out/src/tools/fetch/hrsa.d.ts.map +1 -1
- package/out/src/tools/fetch/hrsa.js.map +1 -1
- package/out/src/tools/fetch/imls-pls.d.ts.map +1 -1
- package/out/src/tools/fetch/imls-pls.js +18 -18
- package/out/src/tools/fetch/imls-pls.js.map +1 -1
- package/out/src/tools/fetch/index.d.ts +3 -1
- package/out/src/tools/fetch/index.d.ts.map +1 -1
- package/out/src/tools/fetch/index.js +3 -1
- package/out/src/tools/fetch/index.js.map +1 -1
- package/out/src/tools/fetch/nad.d.ts +21 -7
- package/out/src/tools/fetch/nad.d.ts.map +1 -1
- package/out/src/tools/fetch/nad.js +5 -5
- package/out/src/tools/fetch/nad.js.map +1 -1
- package/out/src/tools/fetch/nppes.d.ts.map +1 -1
- package/out/src/tools/fetch/nppes.js +9 -15
- package/out/src/tools/fetch/nppes.js.map +1 -1
- package/out/src/tools/fetch/openaddresses.d.ts +3 -1
- package/out/src/tools/fetch/openaddresses.d.ts.map +1 -1
- package/out/src/tools/fetch/openaddresses.js +32 -23
- package/out/src/tools/fetch/openaddresses.js.map +1 -1
- package/out/src/tools/fetch/ppd.d.ts +3 -1
- package/out/src/tools/fetch/ppd.d.ts.map +1 -1
- package/out/src/tools/fetch/ppd.js +14 -3
- package/out/src/tools/fetch/ppd.js.map +1 -1
- package/out/src/tools/fetch/state-hi-schools.d.ts.map +1 -1
- package/out/src/tools/fetch/state-hi-schools.js +13 -6
- package/out/src/tools/fetch/state-hi-schools.js.map +1 -1
- package/out/src/tools/fetch/state-sources.d.ts.map +1 -1
- package/out/src/tools/fetch/state-sources.js +8 -3
- package/out/src/tools/fetch/state-sources.js.map +1 -1
- package/out/src/tools/fetch/tiger-full.d.ts +9 -3
- package/out/src/tools/fetch/tiger-full.d.ts.map +1 -1
- package/out/src/tools/fetch/tiger-full.js +35 -20
- package/out/src/tools/fetch/tiger-full.js.map +1 -1
- package/out/src/tools/gb-title-case.d.ts.map +1 -1
- package/out/src/tools/gb-title-case.js.map +1 -1
- package/out/src/tools/golden-expand.d.ts +24 -8
- package/out/src/tools/golden-expand.d.ts.map +1 -1
- package/out/src/tools/golden-expand.js +20 -12
- package/out/src/tools/golden-expand.js.map +1 -1
- package/out/src/tools/golden-promote.d.ts +18 -6
- package/out/src/tools/golden-promote.d.ts.map +1 -1
- package/out/src/tools/golden-promote.js +22 -7
- package/out/src/tools/golden-promote.js.map +1 -1
- package/out/src/tools/ingest-csv.d.ts +3 -1
- package/out/src/tools/ingest-csv.d.ts.map +1 -1
- package/out/src/tools/ingest-csv.js +17 -19
- package/out/src/tools/ingest-csv.js.map +1 -1
- package/out/src/tools/jsonl-to-parquet.d.ts +18 -6
- package/out/src/tools/jsonl-to-parquet.d.ts.map +1 -1
- package/out/src/tools/jsonl-to-parquet.js +11 -5
- package/out/src/tools/jsonl-to-parquet.js.map +1 -1
- package/out/src/tools/lint-shard-vocab.d.ts +36 -12
- package/out/src/tools/lint-shard-vocab.d.ts.map +1 -1
- package/out/src/tools/lint-shard-vocab.js +25 -11
- package/out/src/tools/lint-shard-vocab.js.map +1 -1
- package/out/src/tools/lint-shard.d.ts +30 -10
- package/out/src/tools/lint-shard.d.ts.map +1 -1
- package/out/src/tools/lint-shard.js +33 -20
- package/out/src/tools/lint-shard.js.map +1 -1
- package/out/src/tools/overlay-manifest.d.ts.map +1 -1
- package/out/src/tools/overlay-manifest.js +6 -4
- package/out/src/tools/overlay-manifest.js.map +1 -1
- package/out/src/tools/shard-kryptonite.d.ts +6 -2
- package/out/src/tools/shard-kryptonite.d.ts.map +1 -1
- package/out/src/tools/shard-kryptonite.js +1 -1
- package/out/src/tools/shard-kryptonite.js.map +1 -1
- package/out/src/tools/shard-translit.d.ts +9 -3
- package/out/src/tools/shard-translit.d.ts.map +1 -1
- package/out/src/tools/shard-translit.js +2 -2
- package/out/src/tools/shard-translit.js.map +1 -1
- package/out/src/types.d.ts +51 -17
- package/out/src/types.d.ts.map +1 -1
- package/out/src/wof-json.d.ts +9 -3
- package/out/src/wof-json.d.ts.map +1 -1
- package/out/src/wof-json.js +6 -4
- package/out/src/wof-json.js.map +1 -1
- package/package.json +7 -6
- package/src/adapter.ts +24 -13
- package/src/adapters/ban/adapter.ts +12 -2
- package/src/adapters/ban/street-decompose.ts +7 -1
- package/src/adapters/fcc-bdc/adapter.ts +19 -3
- package/src/adapters/geonames/adapter.ts +28 -5
- package/src/adapters/geonames-postal/adapter.ts +17 -1
- package/src/adapters/gnaf/adapter.ts +15 -2
- package/src/adapters/gnaf/assemble.ts +40 -10
- package/src/adapters/index.ts +13 -0
- package/src/adapters/openaddresses/adapter.ts +22 -3
- package/src/adapters/overture/adapter.ts +18 -2
- package/src/adapters/state-hi-schools/adapter.ts +15 -0
- package/src/adapters/state-ia-contractors/adapter.ts +15 -0
- package/src/adapters/state-ny-notaries/adapter.ts +15 -0
- package/src/adapters/state-tx-notaries/adapter.ts +18 -2
- package/src/adapters/synth-po-box/adapter.ts +19 -2
- package/src/adapters/tiger/adapter.ts +23 -4
- package/src/adapters/tiger/street-decompose.ts +1 -0
- package/src/adapters/usgov-hrsa-fqhc/adapter.ts +18 -2
- package/src/adapters/usgov-imls-pls/adapter.ts +15 -0
- package/src/adapters/usgov-irs-bmf/adapter.ts +19 -2
- package/src/adapters/usgov-nad/adapter.ts +14 -1
- package/src/adapters/usgov-nppes/adapter.ts +13 -0
- package/src/adapters/usgov-samhsa-treatment-locator/adapter.ts +21 -3
- package/src/adapters/wof-admin-jp/adapter.ts +20 -3
- package/src/adapters/wof-admin-json/adapter.ts +22 -7
- package/src/adapters/wof-postalcode-json/adapter.ts +10 -2
- package/src/align.ts +21 -8
- package/src/build.ts +48 -11
- package/src/codex/us-fips-state.ts +6 -2
- package/src/golden.ts +28 -13
- package/src/license.ts +4 -2
- package/src/parquet-wrapper/reader.ts +6 -4
- package/src/parquet-wrapper/schema.ts +14 -10
- package/src/parquet-wrapper/writer.ts +12 -4
- package/src/parquet.ts +39 -8
- package/src/runner.ts +49 -15
- package/src/shard-recipes/anchor-absorption.ts +21 -5
- package/src/shard-recipes/boundary-stress.ts +20 -7
- package/src/shard-recipes/country-balanced.ts +78 -13
- package/src/shard-recipes/cz-pcfirst-preposition.ts +9 -0
- package/src/shard-recipes/fr-admin-split.ts +30 -5
- package/src/shard-recipes/fr-bare-street.ts +10 -0
- package/src/shard-recipes/fr-fragment.ts +33 -8
- package/src/shard-recipes/fr-lieudit.ts +35 -9
- package/src/shard-recipes/fr-order.ts +47 -9
- package/src/shard-recipes/german.ts +32 -6
- package/src/shard-recipes/house-venue.ts +7 -0
- package/src/shard-recipes/index.ts +12 -4
- package/src/shard-recipes/intersection.ts +63 -14
- package/src/shard-recipes/locale.ts +37 -4
- package/src/shard-recipes/nl-postcode.ts +12 -2
- package/src/shard-recipes/no-fragment.ts +12 -2
- package/src/shard-recipes/no-street-led.ts +11 -1
- package/src/shard-recipes/no-street.ts +7 -0
- package/src/shard-recipes/po-box-cedex.ts +145 -45
- package/src/shard-recipes/po-box.ts +7 -0
- package/src/shard-recipes/scaffold.ts +52 -20
- package/src/shard-recipes/si-bare-village.ts +9 -0
- package/src/shard-recipes/street-affix.ts +88 -13
- package/src/shard-recipes/street-bare.ts +10 -0
- package/src/shard-recipes/street.ts +7 -0
- package/src/shard-recipes/unit.ts +79 -16
- package/src/split.ts +28 -7
- package/src/synthesize-anchor-absorption.ts +41 -14
- package/src/synthesize-boundary-stress.ts +57 -30
- package/src/synthesize-german.ts +30 -7
- package/src/synthesize-house-venue.ts +22 -16
- package/src/synthesize-intersection.ts +34 -7
- package/src/synthesize-no-street.ts +18 -10
- package/src/synthesize-po-box.ts +42 -10
- package/src/synthesize-street.ts +10 -2
- package/src/synthesize.ts +81 -38
- package/src/tokenize.ts +15 -5
- package/src/tools/align-shard.ts +6 -1
- package/src/tools/audit.ts +57 -11
- package/src/tools/corpus-stats.ts +9 -0
- package/src/tools/fetch/ban.ts +32 -10
- package/src/tools/fetch/download.ts +49 -12
- package/src/tools/fetch/hrsa.ts +3 -0
- package/src/tools/fetch/imls-pls.ts +28 -18
- package/src/tools/fetch/index.ts +3 -1
- package/src/tools/fetch/nad.ts +39 -12
- package/src/tools/fetch/nppes.ts +17 -15
- package/src/tools/fetch/openaddresses.ts +52 -24
- package/src/tools/fetch/ppd.ts +22 -3
- package/src/tools/fetch/state-hi-schools.ts +17 -6
- package/src/tools/fetch/state-sources.ts +16 -3
- package/src/tools/fetch/tiger-full.ts +66 -23
- package/src/tools/gb-title-case.ts +1 -0
- package/src/tools/golden-expand.ts +64 -20
- package/src/tools/golden-promote.ts +53 -13
- package/src/tools/ingest-csv.ts +41 -26
- package/src/tools/jsonl-to-parquet.ts +30 -10
- package/src/tools/lint-shard-vocab.ts +75 -24
- package/src/tools/lint-shard.ts +75 -31
- package/src/tools/overlay-manifest.ts +7 -4
- package/src/tools/shard-kryptonite.ts +11 -3
- package/src/tools/shard-translit.ts +20 -5
- package/src/types.ts +51 -17
- package/src/wof-json.ts +18 -7
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/usgov-nad/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,kBAAkB,CAAA;AAC1C,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAA;AAE7C,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAGrD,MAAM,CAAC,MAAM,oBAAoB,GAAG,WAAW,CAAA;AAC/C,MAAM,CAAC,MAAM,yBAAyB,GAAG,eAAe,CAAA;AAgDxD,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC;IAC7B,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,+BAA+B;IAC/B,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;CACJ,CAAC,CAAA;AAEF,SAAS,QAAQ,CAAC,GAAG,MAAwC;IAC5D,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACxB,MAAM,OAAO,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QAE3C,IAAI,OAAO;YAAE,OAAO,OAAO,CAAA;IAC5B,CAAC;IAED,OAAO,SAAS,CAAA;AACjB,CAAC;AAED,SAAS,kBAAkB,CAAC,CAAY;IACvC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEnD,IAAI,IAAI;QAAE,OAAO,IAAI,CAAA;IACrB,MAAM,GAAG,GAAG,CAAC,CAAC,UAAU,IAAI,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,EAAE,CAAA;IAEnE,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAA;IAC1B,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAClD,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAElD,OAAO,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,SAAS,CAAA;AACrE,CAAC;AASD,SAAS,kBAAkB,CAAC,CAAY;IACvC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEhD,IAAI,IAAI,EAAE,CAAC;QACV,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,SAAS,CAAA;QAC9E,MAAM,MAAM,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,SAAS,CAAA;QACtE,MAAM,IAAI,GAAG,CAAC,MAAM,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QAE7D,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;IAC9C,CAAC;
|
|
1
|
+
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/usgov-nad/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,kBAAkB,CAAA;AAC1C,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAA;AAE7C,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAGrD;;;GAGG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,WAAW,CAAA;AAC/C;;;GAGG;AACH,MAAM,CAAC,MAAM,yBAAyB,GAAG,eAAe,CAAA;AAgDxD,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC;IAC7B,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,+BAA+B;IAC/B,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;CACJ,CAAC,CAAA;AAEF,SAAS,QAAQ,CAAC,GAAG,MAAwC;IAC5D,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACxB,MAAM,OAAO,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QAE3C,IAAI,OAAO;YAAE,OAAO,OAAO,CAAA;IAC5B,CAAC;IAED,OAAO,SAAS,CAAA;AACjB,CAAC;AAED,SAAS,kBAAkB,CAAC,CAAY;IACvC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEnD,IAAI,IAAI;QAAE,OAAO,IAAI,CAAA;IACrB,MAAM,GAAG,GAAG,CAAC,CAAC,UAAU,IAAI,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,EAAE,CAAA;IAEnE,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAA;IAC1B,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAClD,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAElD,OAAO,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,SAAS,CAAA;AACrE,CAAC;AASD,SAAS,kBAAkB,CAAC,CAAY;IACvC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEhD,IAAI,IAAI,EAAE,CAAC;QACV,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;QACpD,MAAM,MAAM,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,SAAS,CAAA;QAC9E,MAAM,MAAM,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,SAAS,CAAA;QACtE,MAAM,IAAI,GAAG,CAAC,MAAM,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QAE7D,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;IAC9C,CAAC;IAED,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEnD,IAAI,IAAI;QAAE,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;IAEvC,OAAO,SAAS,CAAA;AACjB,CAAC;AAED,SAAS,eAAe,CAAC,CAAY;IACpC,OAAO,QAAQ,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC,UAAU,EAAE,CAAC,CAAC,UAAU,CAAC,CAAA;AACrE,CAAC;AAED,SAAS,eAAe,CAAC,CAAY;IACpC,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEhD,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAA;IAC1B,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAA;IAEhD,OAAO,KAAK,CAAC,CAAC,CAAC,GAAG,GAAG,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,GAAG,CAAA;AACvC,CAAC;AAED,SAAS,UAAU,CAAC,KAQnB;IACA,MAAM,UAAU,GAAG,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IACjG,MAAM,IAAI,GAAG,GAAG,KAAK,CAAC,QAAQ,KAAK,KAAK,CAAC,MAAM,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAA;IAEnE,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,UAAU,IAAI,SAAS,EAAE,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AAC/E,CAAC;AAED,MAAM,UAAU,qBAAqB;IACpC,OAAO;QACN,EAAE,EAAE,oBAAoB;QACxB,cAAc,EAAE,yBAAyB;QACzC,WAAW,EACV,6GAA6G;QAE9G,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,qDAAqD,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YACrF,CAAC;YAED,qFAAqF;YACrF,mFAAmF;YACnF,2DAA2D;YAC3D,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,CAAA;YAC7C,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,EAAE,CAAA;YAEtE,IAAI,OAAO,GAAG,CAAC,CAAA;YACf,KAAK,EAAE,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBACnC,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,MAAK;gBAC/B,sFAAsF;gBACtF,yFAAyF;gBACzF,0FAA0F;gBAC1F,kEAAkE;gBAClE,MAAM,KAAK,GAAG,eAAe,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAA;gBAEpE,IAAI,KAAK,EAAE,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBAChC,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;wBAAE,MAAM,KAAK,CAAA;oBAErC,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;wBAAE,MAAM,KAAK,CAAA;oBAElE,IAAI,CAAC,IAAI;wBAAE,SAAQ;oBAEnB,IAAI,MAAiB,CAAA;oBAErB,IAAI,CAAC;wBACJ,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAc,CAAA;oBACvC,CAAC;oBAAC,MAAM,CAAC;wBACR,SAAQ,CAAC,iCAAiC;oBAC3C,CAAC;oBAED,MAAM,KAAK,GAAG,CAAC,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAA;oBAElE,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC;wBAAE,SAAQ;oBAEvC,MAAM,QAAQ,GAAG,eAAe,CAAC,MAAM,CAAC,CAAA;oBAExC,IAAI,CAAC,QAAQ;wBAAE,SAAQ;oBAEvB,MAAM,QAAQ,GAAG,eAAe,CAAC,MAAM,CAAC,CAAA;oBAExC,IAAI,CAAC,QAAQ;wBAAE,SAAQ;oBAEvB,MAAM,UAAU,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAA;oBAC7C,MAAM,WAAW,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAA;oBAC9C,MAAM,KAAK,GAAG,QAAQ,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;oBACzC,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,QAAQ,EAAE,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,IAAI,CAAC,CAAA;oBAE9E,MAAM,UAAU,GAA+B;wBAC9C,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBAC3B,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACrD,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACnE,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBAC5D,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACnE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACzB,QAAQ;wBACR,MAAM,EAAE,KAAK;wBACb,QAAQ;qBACR,CAAA;oBAED,MAAM,GAAG,GAAG,UAAU,CAAC;wBACtB,KAAK;wBACL,WAAW;wBACX,MAAM,EAAE,UAAU,EAAE,IAAI;wBACxB,IAAI;wBACJ,QAAQ;wBACR,MAAM,EAAE,KAAK;wBACb,QAAQ;qBACR,CAAC,CAAA;oBAEF,IAAI,CAAC,GAAG;wBAAE,SAAQ;oBAElB,MAAM,OAAO,GAAG,mBAAmB,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;oBAEpD,IAAI,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM,IAAI,CAAC;wBAAE,SAAQ;oBAE9C,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI;wBAC3B,CAAC,CAAC,GAAG,oBAAoB,IAAI,MAAM,CAAC,IAAI,EAAE;wBAC1C,CAAC,CAAC,GAAG,oBAAoB,IAAI,MAAM,CAAC,QAAQ,IAAI,GAAG,KAAK,IAAI,OAAO,EAAE,EAAE,CAAA;oBAExE,MAAM;wBACL,GAAG;wBACH,UAAU,EAAE,OAAO;wBACnB,OAAO,EAAE,IAAI;wBACb,MAAM,EAAE,OAAO;wBACf,MAAM,EAAE,oBAAoB;wBAC5B,SAAS,EAAE,QAAQ;wBACnB,cAAc,EAAE,EAAE;wBAClB,OAAO,EAAE,yBAAyB;qBAClC,CAAA;oBAED,OAAO,EAAE,CAAA;gBACV,CAAC;YACF,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,qBAAqB,EAAE,CAAA"}
|
|
@@ -21,8 +21,19 @@
|
|
|
21
21
|
* License: stamped `"Public Domain"` per CMS's federal government distribution terms.
|
|
22
22
|
*/
|
|
23
23
|
import type { CorpusAdapter } from "../../types.ts";
|
|
24
|
+
/**
|
|
25
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
26
|
+
* it came from.
|
|
27
|
+
*/
|
|
24
28
|
export declare const USGOV_NPPES_ADAPTER_ID = "usgov-nppes";
|
|
29
|
+
/**
|
|
30
|
+
* License carried by this source (Public Domain), attached to each row so downstream consumers inherit the terms rather
|
|
31
|
+
* than having to look them up.
|
|
32
|
+
*/
|
|
25
33
|
export declare const USGOV_NPPES_DEFAULT_LICENSE = "Public Domain";
|
|
26
34
|
export declare function createUsgovNPPESAdapter(): CorpusAdapter;
|
|
35
|
+
/**
|
|
36
|
+
* The configured adapter instance registered with the corpus builder.
|
|
37
|
+
*/
|
|
27
38
|
export declare const usgovNPPESAdapter: CorpusAdapter;
|
|
28
39
|
//# sourceMappingURL=adapter.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/usgov-nppes/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AASH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AAEjF,eAAO,MAAM,sBAAsB,gBAAgB,CAAA;AACnD,eAAO,MAAM,2BAA2B,kBAAkB,CAAA;AA0C1D,wBAAgB,uBAAuB,IAAI,aAAa,
|
|
1
|
+
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/usgov-nppes/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AASH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AAEjF;;;GAGG;AACH,eAAO,MAAM,sBAAsB,gBAAgB,CAAA;AACnD;;;GAGG;AACH,eAAO,MAAM,2BAA2B,kBAAkB,CAAA;AA0C1D,wBAAgB,uBAAuB,IAAI,aAAa,CA6FvD;AAED;;GAEG;AACH,eAAO,MAAM,iBAAiB,eAA4B,CAAA"}
|
|
@@ -25,7 +25,15 @@ import { parse as csvParse } from "csv-parse";
|
|
|
25
25
|
import { stableSourceID } from "../../adapter.js";
|
|
26
26
|
import { lookupStateAbbreviation } from "../../codex/us-fips-state.js";
|
|
27
27
|
import { reconcileComponents } from "../../format.js";
|
|
28
|
+
/**
|
|
29
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
30
|
+
* it came from.
|
|
31
|
+
*/
|
|
28
32
|
export const USGOV_NPPES_ADAPTER_ID = "usgov-nppes";
|
|
33
|
+
/**
|
|
34
|
+
* License carried by this source (Public Domain), attached to each row so downstream consumers inherit the terms rather
|
|
35
|
+
* than having to look them up.
|
|
36
|
+
*/
|
|
29
37
|
export const USGOV_NPPES_DEFAULT_LICENSE = "Public Domain";
|
|
30
38
|
const HOUSE_NUMBER_PREFIX = /^(\d+(?:-\d+)?[A-Za-z]?)\s+(.+)$/;
|
|
31
39
|
function splitAddress(address) {
|
|
@@ -119,5 +127,8 @@ export function createUsgovNPPESAdapter() {
|
|
|
119
127
|
},
|
|
120
128
|
};
|
|
121
129
|
}
|
|
130
|
+
/**
|
|
131
|
+
* The configured adapter instance registered with the corpus builder.
|
|
132
|
+
*/
|
|
122
133
|
export const usgovNPPESAdapter = createUsgovNPPESAdapter();
|
|
123
134
|
//# sourceMappingURL=adapter.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/usgov-nppes/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAEH,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAA;AAE1C,OAAO,EAAE,KAAK,IAAI,QAAQ,EAAE,MAAM,WAAW,CAAA;AAE7C,OAAO,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAA;AACjD,OAAO,EAAE,uBAAuB,EAAE,MAAM,8BAA8B,CAAA;AACtE,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAGrD,MAAM,CAAC,MAAM,sBAAsB,GAAG,aAAa,CAAA;AACnD,MAAM,CAAC,MAAM,2BAA2B,GAAG,eAAe,CAAA;AAE1D,MAAM,mBAAmB,GAAG,kCAAkC,CAAA;AAe9D,SAAS,YAAY,CAAC,OAAe;IACpC,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,EAAE,CAAA;IAE9B,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAA;IACzB,MAAM,CAAC,GAAG,mBAAmB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAE3C,IAAI,CAAC;QAAE,OAAO,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE,CAAA;IAE1D,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,CAAA;AAC3B,CAAC;AAED,SAAS,UAAU,CAClB,KAAyB,EACzB,KAAyB,EACzB,MAAc,EACd,IAAY,EACZ,KAAa,EACb,QAAgB;IAEhB,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IACnE,MAAM,QAAQ,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAE7G,OAAO,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AAChE,CAAC;AAED,MAAM,UAAU,uBAAuB;IACtC,OAAO;QACN,EAAE,EAAE,sBAAsB;QAC1B,cAAc,EAAE,2BAA2B;QAC3C,WAAW,EACV,2IAA2I;QAE5I,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,uDAAuD,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YACvF,CAAC;YAED,MAAM,MAAM,GAAG,gBAAgB,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAA;
|
|
1
|
+
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/usgov-nppes/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAEH,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAA;AAE1C,OAAO,EAAE,KAAK,IAAI,QAAQ,EAAE,MAAM,WAAW,CAAA;AAE7C,OAAO,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAA;AACjD,OAAO,EAAE,uBAAuB,EAAE,MAAM,8BAA8B,CAAA;AACtE,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAGrD;;;GAGG;AACH,MAAM,CAAC,MAAM,sBAAsB,GAAG,aAAa,CAAA;AACnD;;;GAGG;AACH,MAAM,CAAC,MAAM,2BAA2B,GAAG,eAAe,CAAA;AAE1D,MAAM,mBAAmB,GAAG,kCAAkC,CAAA;AAe9D,SAAS,YAAY,CAAC,OAAe;IACpC,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,EAAE,CAAA;IAE9B,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAA;IACzB,MAAM,CAAC,GAAG,mBAAmB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAE3C,IAAI,CAAC;QAAE,OAAO,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE,CAAA;IAE1D,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,CAAA;AAC3B,CAAC;AAED,SAAS,UAAU,CAClB,KAAyB,EACzB,KAAyB,EACzB,MAAc,EACd,IAAY,EACZ,KAAa,EACb,QAAgB;IAEhB,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IACnE,MAAM,QAAQ,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAE7G,OAAO,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AAChE,CAAC;AAED,MAAM,UAAU,uBAAuB;IACtC,OAAO;QACN,EAAE,EAAE,sBAAsB;QAC1B,cAAc,EAAE,2BAA2B;QAC3C,WAAW,EACV,2IAA2I;QAE5I,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,uDAAuD,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YACvF,CAAC;YAED,MAAM,MAAM,GAAG,gBAAgB,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAA;YAErE,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CACzB,QAAQ,CAAC;gBACR,OAAO,EAAE,IAAI;gBACb,gBAAgB,EAAE,IAAI;gBACtB,YAAY,EAAE,IAAI;gBAClB,kBAAkB,EAAE,IAAI;aACxB,CAAC,CACF,CAAA;YAED,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,IAAI,CAAC;gBACJ,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,MAAiC,EAAE,CAAC;oBAC9D,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;wBAAE,MAAK;oBAE/B,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;wBAAE,MAAK;oBAE5D,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBACrC,MAAM,UAAU,GAAG,CAAC,MAAM,CAAC,kBAAkB,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAC5D,MAAM,OAAO,GAAG,CAAC,MAAM,CAAC,kDAAkD,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBACzF,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,iCAAiC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBACzE,MAAM,SAAS,GAAG,CAAC,MAAM,CAAC,qBAAqB,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAE9D,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,wDAAwD,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAChG,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,yDAAyD,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBACjG,MAAM,IAAI,GAAG,CAAC,MAAM,CAAC,uDAAuD,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAC3F,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,wDAAwD,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAChG,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,yDAAyD,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAEjG,IAAI,CAAC,IAAI,IAAI,CAAC,QAAQ;wBAAE,SAAQ;oBAEhC,MAAM,KAAK,GAAG,uBAAuB,CAAC,QAAQ,CAAC,CAAA;oBAE/C,IAAI,CAAC,KAAK;wBAAE,SAAQ;oBAEpB,MAAM,UAAU,GAAG,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;oBACjE,MAAM,KAAK,GAAG,YAAY,CAAC,UAAU,CAAC,CAAA;oBAEtC,IAAI,CAAC,KAAK;wBAAE,SAAQ;oBAEpB,MAAM,KAAK,GAAG,OAAO,IAAI,CAAC,SAAS,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,SAAS,CAAA;oBAErF,MAAM,UAAU,GAA+B;wBAC9C,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBAC3B,GAAG,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACnE,MAAM,EAAE,KAAK,CAAC,MAAM;wBACpB,QAAQ,EAAE,IAAI;wBACd,MAAM,EAAE,KAAK,CAAC,YAAY;wBAC1B,QAAQ;qBACR,CAAA;oBAED,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,EAAE,KAAK,CAAC,YAAY,EAAE,KAAK,CAAC,MAAM,EAAE,IAAI,EAAE,KAAK,CAAC,YAAY,EAAE,QAAQ,CAAC,CAAA;oBAEnG,IAAI,CAAC,GAAG;wBAAE,SAAQ;oBAElB,MAAM,OAAO,GAAG,mBAAmB,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;oBAEpD,IAAI,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM,IAAI,CAAC;wBAAE,SAAQ;oBAE9C,MAAM,QAAQ,GAAG,GAAG,CAAC,CAAC,CAAC,GAAG,sBAAsB,IAAI,GAAG,EAAE,CAAC,CAAC,CAAC,cAAc,CAAC,sBAAsB,EAAE,OAAO,CAAC,CAAA;oBAE3G,MAAM;wBACL,GAAG;wBACH,UAAU,EAAE,OAAO;wBACnB,OAAO,EAAE,IAAI;wBACb,MAAM,EAAE,OAAO;wBACf,MAAM,EAAE,sBAAsB;wBAC9B,SAAS,EAAE,QAAQ;wBACnB,cAAc,EAAE,EAAE;wBAClB,OAAO,EAAE,2BAA2B;qBACpC,CAAA;oBAED,OAAO,EAAE,CAAA;gBACV,CAAC;YACF,CAAC;oBAAS,CAAC;gBACV,MAAM,CAAC,OAAO,EAAE,CAAA;YACjB,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,uBAAuB,EAAE,CAAA"}
|
|
@@ -28,8 +28,19 @@
|
|
|
28
28
|
* distribution terms.
|
|
29
29
|
*/
|
|
30
30
|
import type { CorpusAdapter } from "../../types.ts";
|
|
31
|
+
/**
|
|
32
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
33
|
+
* it came from.
|
|
34
|
+
*/
|
|
31
35
|
export declare const USGOV_SAMHSA_ADAPTER_ID = "usgov-samhsa-treatment-locator";
|
|
36
|
+
/**
|
|
37
|
+
* License carried by this source (Public Domain), attached to each row so downstream consumers inherit the terms rather
|
|
38
|
+
* than having to look them up.
|
|
39
|
+
*/
|
|
32
40
|
export declare const USGOV_SAMHSA_DEFAULT_LICENSE = "Public Domain";
|
|
33
41
|
export declare function createUsgovSamhsaTreatmentLocatorAdapter(): CorpusAdapter;
|
|
42
|
+
/**
|
|
43
|
+
* The configured adapter instance registered with the corpus builder.
|
|
44
|
+
*/
|
|
34
45
|
export declare const usgovSamhsaTreatmentLocatorAdapter: CorpusAdapter;
|
|
35
46
|
//# sourceMappingURL=adapter.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/usgov-samhsa-treatment-locator/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AASH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AAEjF,eAAO,MAAM,uBAAuB,mCAAmC,CAAA;AACvE,eAAO,MAAM,4BAA4B,kBAAkB,CAAA;
|
|
1
|
+
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/usgov-samhsa-treatment-locator/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AASH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AAEjF;;;GAGG;AACH,eAAO,MAAM,uBAAuB,mCAAmC,CAAA;AACvE;;;GAGG;AACH,eAAO,MAAM,4BAA4B,kBAAkB,CAAA;AA0F3D,wBAAgB,wCAAwC,IAAI,aAAa,CAqFxE;AAED;;GAEG;AACH,eAAO,MAAM,kCAAkC,eAA6C,CAAA"}
|
|
@@ -32,7 +32,15 @@ import { parse as csvParse } from "csv-parse";
|
|
|
32
32
|
import { stableSourceID } from "../../adapter.js";
|
|
33
33
|
import { lookupStateAbbreviation } from "../../codex/us-fips-state.js";
|
|
34
34
|
import { reconcileComponents } from "../../format.js";
|
|
35
|
+
/**
|
|
36
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
37
|
+
* it came from.
|
|
38
|
+
*/
|
|
35
39
|
export const USGOV_SAMHSA_ADAPTER_ID = "usgov-samhsa-treatment-locator";
|
|
40
|
+
/**
|
|
41
|
+
* License carried by this source (Public Domain), attached to each row so downstream consumers inherit the terms rather
|
|
42
|
+
* than having to look them up.
|
|
43
|
+
*/
|
|
36
44
|
export const USGOV_SAMHSA_DEFAULT_LICENSE = "Public Domain";
|
|
37
45
|
const HOUSE_NUMBER_PREFIX = /^(\d+(?:-\d+)?[A-Za-z]?)\s+(.+)$/;
|
|
38
46
|
function splitAddress(address) {
|
|
@@ -80,7 +88,9 @@ function composeVenue(name1, name2) {
|
|
|
80
88
|
return n2;
|
|
81
89
|
return `${n1} - ${n2}`;
|
|
82
90
|
}
|
|
83
|
-
/**
|
|
91
|
+
/**
|
|
92
|
+
* Same envelope-style format as HRSA: venue prefix, street body, city/state/zip suffix.
|
|
93
|
+
*/
|
|
84
94
|
function composeRaw(venue, house, street, city, state, postcode) {
|
|
85
95
|
const streetPart = [house, street].filter(Boolean).join(" ").trim();
|
|
86
96
|
const cityPart = [city.trim(), [state, postcode].filter(Boolean).join(" ").trim()].filter(Boolean).join(", ");
|
|
@@ -133,7 +143,7 @@ export function createUsgovSamhsaTreatmentLocatorAdapter() {
|
|
|
133
143
|
if (!raw)
|
|
134
144
|
continue;
|
|
135
145
|
const aligned = reconcileComponents(components, raw);
|
|
136
|
-
if (Object.keys(aligned).length
|
|
146
|
+
if (!Object.keys(aligned).length)
|
|
137
147
|
continue;
|
|
138
148
|
const frID = (record.frid ?? "").trim();
|
|
139
149
|
const sourceID = frID
|
|
@@ -158,5 +168,8 @@ export function createUsgovSamhsaTreatmentLocatorAdapter() {
|
|
|
158
168
|
},
|
|
159
169
|
};
|
|
160
170
|
}
|
|
171
|
+
/**
|
|
172
|
+
* The configured adapter instance registered with the corpus builder.
|
|
173
|
+
*/
|
|
161
174
|
export const usgovSamhsaTreatmentLocatorAdapter = createUsgovSamhsaTreatmentLocatorAdapter();
|
|
162
175
|
//# sourceMappingURL=adapter.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/usgov-samhsa-treatment-locator/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAA;AAE1C,OAAO,EAAE,KAAK,IAAI,QAAQ,EAAE,MAAM,WAAW,CAAA;AAE7C,OAAO,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAA;AACjD,OAAO,EAAE,uBAAuB,EAAE,MAAM,8BAA8B,CAAA;AACtE,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAGrD,MAAM,CAAC,MAAM,uBAAuB,GAAG,gCAAgC,CAAA;AACvE,MAAM,CAAC,MAAM,4BAA4B,GAAG,eAAe,CAAA;
|
|
1
|
+
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/usgov-samhsa-treatment-locator/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAA;AAE1C,OAAO,EAAE,KAAK,IAAI,QAAQ,EAAE,MAAM,WAAW,CAAA;AAE7C,OAAO,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAA;AACjD,OAAO,EAAE,uBAAuB,EAAE,MAAM,8BAA8B,CAAA;AACtE,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAGrD;;;GAGG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG,gCAAgC,CAAA;AACvE;;;GAGG;AACH,MAAM,CAAC,MAAM,4BAA4B,GAAG,eAAe,CAAA;AAoB3D,MAAM,mBAAmB,GAAG,kCAAkC,CAAA;AAE9D,SAAS,YAAY,CAAC,OAAe;IACpC,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,EAAE,CAAA;IAE9B,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAA;IACzB,MAAM,CAAC,GAAG,mBAAmB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAE3C,IAAI,CAAC;QAAE,OAAO,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE,CAAA;IAE1D,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,CAAA;AAC3B,CAAC;AAED;;;;GAIG;AACH,SAAS,iBAAiB,CAAC,OAAe,EAAE,OAA2B;IACtE,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,EAAE,CAAA;IACzB,MAAM,EAAE,GAAG,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;IAEjC,IAAI,CAAC,EAAE,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAA;IAEzB,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAA;IAElB,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAA;IAElB,OAAO,GAAG,EAAE,KAAK,EAAE,EAAE,CAAA;AACtB,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,YAAY,CAAC,KAAa,EAAE,KAAyB;IAC7D,MAAM,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,CAAA;IACvB,MAAM,EAAE,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;IAE/B,IAAI,CAAC,EAAE,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAA;IAEzB,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAA;IAElB,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAA;IAElB,OAAO,GAAG,EAAE,MAAM,EAAE,EAAE,CAAA;AACvB,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAClB,KAAa,EACb,KAAyB,EACzB,MAAc,EACd,IAAY,EACZ,KAAa,EACb,QAAgB;IAEhB,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IACnE,MAAM,QAAQ,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAE7G,OAAO,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACvE,CAAC;AAED,MAAM,UAAU,wCAAwC;IACvD,OAAO;QACN,EAAE,EAAE,uBAAuB;QAC3B,cAAc,EAAE,4BAA4B;QAC5C,WAAW,EACV,gJAAgJ;QAEjJ,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,wDAAwD,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YACxF,CAAC;YAED,MAAM,MAAM,GAAG,gBAAgB,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAA;YAErE,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CACzB,QAAQ,CAAC;gBACR,OAAO,EAAE,IAAI;gBACb,gBAAgB,EAAE,IAAI;gBACtB,YAAY,EAAE,IAAI;gBAClB,kBAAkB,EAAE,IAAI;aACxB,CAAC,CACF,CAAA;YAED,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,IAAI,CAAC;gBACJ,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,MAAsC,EAAE,CAAC;oBACnE,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;wBAAE,MAAK;oBAE/B,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;wBAAE,MAAK;oBAE5D,MAAM,KAAK,GAAG,YAAY,CAAC,MAAM,CAAC,KAAK,IAAI,EAAE,EAAE,MAAM,CAAC,KAAK,CAAC,CAAA;oBAC5D,MAAM,MAAM,GAAG,iBAAiB,CAAC,MAAM,CAAC,OAAO,IAAI,EAAE,EAAE,MAAM,CAAC,OAAO,CAAC,CAAA;oBACtE,MAAM,KAAK,GAAG,YAAY,CAAC,MAAM,CAAC,CAAA;oBAClC,MAAM,IAAI,GAAG,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBACvC,MAAM,SAAS,GAAG,CAAC,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAC7C,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAE1C,IAAI,CAAC,KAAK,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,IAAI,CAAC,QAAQ;wBAAE,SAAQ;oBACpD,MAAM,KAAK,GAAG,uBAAuB,CAAC,SAAS,CAAC,CAAA;oBAEhD,IAAI,CAAC,KAAK;wBAAE,SAAQ;oBAEpB,mEAAmE;oBACnE,MAAM,UAAU,GAA+B;wBAC9C,KAAK;wBACL,GAAG,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACnE,MAAM,EAAE,KAAK,CAAC,MAAM;wBACpB,QAAQ,EAAE,IAAI;wBACd,MAAM,EAAE,KAAK,CAAC,YAAY;wBAC1B,QAAQ;qBACR,CAAA;oBAED,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,EAAE,KAAK,CAAC,YAAY,EAAE,KAAK,CAAC,MAAM,EAAE,IAAI,EAAE,KAAK,CAAC,YAAY,EAAE,QAAQ,CAAC,CAAA;oBAEnG,IAAI,CAAC,GAAG;wBAAE,SAAQ;oBAElB,MAAM,OAAO,GAAG,mBAAmB,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;oBAEpD,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM;wBAAE,SAAQ;oBAE1C,MAAM,IAAI,GAAG,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;oBAEvC,MAAM,QAAQ,GAAG,IAAI;wBACpB,CAAC,CAAC,GAAG,uBAAuB,IAAI,IAAI,EAAE;wBACtC,CAAC,CAAC,cAAc,CAAC,uBAAuB,EAAE,OAAO,CAAC,CAAA;oBAEnD,MAAM;wBACL,GAAG;wBACH,UAAU,EAAE,OAAO;wBACnB,OAAO,EAAE,IAAI;wBACb,MAAM,EAAE,OAAO;wBACf,MAAM,EAAE,uBAAuB;wBAC/B,SAAS,EAAE,QAAQ;wBACnB,cAAc,EAAE,EAAE;wBAClB,OAAO,EAAE,4BAA4B;qBACrC,CAAA;oBAED,OAAO,EAAE,CAAA;gBACV,CAAC;YACF,CAAC;oBAAS,CAAC;gBACV,MAAM,CAAC,OAAO,EAAE,CAAA;YACjB,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,kCAAkC,GAAG,wCAAwC,EAAE,CAAA"}
|
|
@@ -29,6 +29,10 @@
|
|
|
29
29
|
* from MLIT/JapanPost).
|
|
30
30
|
*/
|
|
31
31
|
import type { CanonicalRow, CorpusAdapter } from "../../types.ts";
|
|
32
|
+
/**
|
|
33
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
34
|
+
* it came from.
|
|
35
|
+
*/
|
|
32
36
|
export declare const WOF_ADMIN_JP_ADAPTER_ID = "wof-admin-jp";
|
|
33
37
|
interface PlaceRow {
|
|
34
38
|
id: number;
|
|
@@ -53,6 +57,9 @@ export declare function synthesizeJpAddress(chain: PlaceRow[], jpnNames: Map<num
|
|
|
53
57
|
* yields canonical rows.
|
|
54
58
|
*/
|
|
55
59
|
export declare function createWOFAdminJpAdapter(): CorpusAdapter;
|
|
60
|
+
/**
|
|
61
|
+
* The configured adapter instance registered with the corpus builder.
|
|
62
|
+
*/
|
|
56
63
|
export declare const wofAdminJpAdapter: CorpusAdapter;
|
|
57
64
|
export {};
|
|
58
65
|
//# sourceMappingURL=adapter.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-jp/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAIH,OAAO,KAAK,EAAkB,YAAY,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAA;
|
|
1
|
+
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-jp/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAIH,OAAO,KAAK,EAAkB,YAAY,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAA;AAOjF;;;GAGG;AACH,eAAO,MAAM,uBAAuB,iBAAiB,CAAA;AAErD,UAAU,QAAQ;IACjB,EAAE,EAAE,MAAM,CAAA;IACV,IAAI,EAAE,MAAM,CAAA;IACZ,SAAS,EAAE,MAAM,CAAA;IACjB,SAAS,EAAE,MAAM,CAAA;IACjB,OAAO,EAAE,MAAM,CAAA;CACf;AAkCD;;;;;;GAMG;AACH,wBAAgB,mBAAmB,CAClC,KAAK,EAAE,QAAQ,EAAE,EACjB,QAAQ,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,GAC3B;IACF,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,YAAY,CAAC,YAAY,CAAC,CAAA;CACtC,GAAG,IAAI,CAyBP;AAED;;;GAGG;AACH,wBAAgB,uBAAuB,IAAI,aAAa,CAyDvD;AAED;;GAEG;AACH,eAAO,MAAM,iBAAiB,eAA4B,CAAA"}
|
|
@@ -29,13 +29,23 @@
|
|
|
29
29
|
* from MLIT/JapanPost).
|
|
30
30
|
*/
|
|
31
31
|
import { DatabaseSync } from "node:sqlite";
|
|
32
|
+
/**
|
|
33
|
+
* Ancestry hops walked upward from a WOF record before giving up — deeper than any real JP admin chain.
|
|
34
|
+
*/
|
|
35
|
+
const MAX_ANCESTRY_DEPTH = 6;
|
|
36
|
+
/**
|
|
37
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
38
|
+
* it came from.
|
|
39
|
+
*/
|
|
32
40
|
export const WOF_ADMIN_JP_ADAPTER_ID = "wof-admin-jp";
|
|
33
|
-
/**
|
|
41
|
+
/**
|
|
42
|
+
* Walk parent chain up to 6 levels.
|
|
43
|
+
*/
|
|
34
44
|
function chainOf(db, startID, _jpnNames) {
|
|
35
45
|
const stmt = db.prepare(`SELECT id, name, placetype, parent_id, country FROM spr WHERE id = ?`);
|
|
36
46
|
const out = [];
|
|
37
47
|
let id = startID;
|
|
38
|
-
for (let i = 0; i <
|
|
48
|
+
for (let i = 0; i < MAX_ANCESTRY_DEPTH && id > 0; i++) {
|
|
39
49
|
const row = stmt.get(id);
|
|
40
50
|
if (!row)
|
|
41
51
|
break;
|
|
@@ -44,7 +54,9 @@ function chainOf(db, startID, _jpnNames) {
|
|
|
44
54
|
}
|
|
45
55
|
return out;
|
|
46
56
|
}
|
|
47
|
-
/**
|
|
57
|
+
/**
|
|
58
|
+
* Pick the best display name for a place: prefer Japanese variant, fall back to English.
|
|
59
|
+
*/
|
|
48
60
|
function pickName(row, jpnNames) {
|
|
49
61
|
return jpnNames.get(row.id) ?? row.name;
|
|
50
62
|
}
|
|
@@ -127,5 +139,8 @@ export function createWOFAdminJpAdapter() {
|
|
|
127
139
|
},
|
|
128
140
|
};
|
|
129
141
|
}
|
|
142
|
+
/**
|
|
143
|
+
* The configured adapter instance registered with the corpus builder.
|
|
144
|
+
*/
|
|
130
145
|
export const wofAdminJpAdapter = createWOFAdminJpAdapter();
|
|
131
146
|
//# sourceMappingURL=adapter.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-jp/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAI1C,MAAM,CAAC,MAAM,uBAAuB,GAAG,cAAc,CAAA;AAgBrD
|
|
1
|
+
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-jp/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAI1C;;GAEG;AACH,MAAM,kBAAkB,GAAG,CAAC,CAAA;AAE5B;;;GAGG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG,cAAc,CAAA;AAgBrD;;GAEG;AACH,SAAS,OAAO,CAAC,EAAgB,EAAE,OAAe,EAAE,SAA8B;IACjF,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CAAC,sEAAsE,CAAC,CAAA;IAC/F,MAAM,GAAG,GAAe,EAAE,CAAA;IAC1B,IAAI,EAAE,GAAG,OAAO,CAAA;IAEhB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,kBAAkB,IAAI,EAAE,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACvD,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,CAAyB,CAAA;QAEhD,IAAI,CAAC,GAAG;YAAE,MAAK;QACf,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACb,EAAE,GAAG,GAAG,CAAC,SAAS,CAAA;IACnB,CAAC;IAED,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;GAEG;AACH,SAAS,QAAQ,CAAC,GAAa,EAAE,QAA6B;IAC7D,OAAO,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,GAAG,CAAC,IAAI,CAAA;AACxC,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,mBAAmB,CAClC,KAAiB,EACjB,QAA6B;IAK7B,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,KAAK,QAAQ,CAAC,CAAA;IAC1D,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,KAAK,UAAU,IAAI,CAAC,CAAC,SAAS,KAAK,QAAQ,CAAC,CAAA;IAE1F,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ;QAAE,OAAO,IAAI,CAAA;IAErC,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,KAAK,eAAe,CAAC,CAAA;IAExE,MAAM,UAAU,GAAG,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAA;IAC7C,MAAM,YAAY,GAAG,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAA;IACjD,MAAM,iBAAiB,GAAG,aAAa,CAAC,CAAC,CAAC,QAAQ,CAAC,aAAa,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAElF,MAAM,UAAU,GAA+B;QAC9C,MAAM,EAAE,UAAU;QAClB,QAAQ,EAAE,YAAY;QACtB,OAAO,EAAE,IAAI;KACb,CAAA;IAED,IAAI,iBAAiB,EAAE,CAAC;QACvB,UAAU,CAAC,kBAAkB,GAAG,iBAAiB,CAAA;IAClD,CAAC;IAED,MAAM,GAAG,GAAG,CAAC,UAAU,EAAE,YAAY,EAAE,iBAAiB,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IAElF,OAAO,EAAE,GAAG,EAAE,UAAU,EAAE,CAAA;AAC3B,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,uBAAuB;IACtC,OAAO;QACN,EAAE,EAAE,uBAAuB;QAC3B,cAAc,EAAE,WAAW;QAC3B,WAAW,EAAE,kFAAkF;QAE/F,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,wDAAwD,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YACxF,CAAC;YAED,MAAM,EAAE,GAAG,IAAI,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAA;YAE/D,IAAI,CAAC;gBACJ,MAAM,YAAY,GAAG,EAAE,CAAC,OAAO,CAAC,mDAAmD,CAAC,CAAA;gBACpF,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAA;gBAE1C,KAAK,MAAM,GAAG,IAAI,YAAY,CAAC,GAAG,EAAoC,EAAE,CAAC;oBACxE,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,CAAC;wBAC3B,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,IAAI,CAAC,CAAA;oBAC/B,CAAC;gBACF,CAAC;gBAED,MAAM,KAAK,GAAG,EAAE,CAAC,OAAO,CAAC,qEAAqE,CAAC,CAAC,GAAG,EAEhG,CAAA;gBAEH,IAAI,OAAO,GAAG,CAAC,CAAA;gBAEf,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBAC1B,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;wBAAE,MAAK;oBAE/B,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;wBAAE,MAAK;oBAE5D,MAAM,KAAK,GAAG,OAAO,CAAC,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAA;oBAC5C,MAAM,KAAK,GAAG,mBAAmB,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAA;oBAElD,IAAI,CAAC,KAAK;wBAAE,SAAQ;oBAEpB,MAAM;wBACL,GAAG,EAAE,KAAK,CAAC,GAAG;wBACd,UAAU,EAAE,KAAK,CAAC,UAAU;wBAC5B,OAAO,EAAE,IAAI;wBACb,MAAM,EAAE,OAAO;wBACf,MAAM,EAAE,uBAAuB;wBAC/B,SAAS,EAAE,GAAG,uBAAuB,IAAI,IAAI,CAAC,EAAE,EAAE;wBAClD,cAAc,EAAE,EAAE;wBAClB,OAAO,EAAE,WAAW;qBACpB,CAAA;oBAED,OAAO,EAAE,CAAA;gBACV,CAAC;YACF,CAAC;oBAAS,CAAC;gBACV,EAAE,CAAC,KAAK,EAAE,CAAA;YACX,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,uBAAuB,EAAE,CAAA"}
|
|
@@ -42,9 +42,13 @@ import type { ComponentTag } from "@mailwoman/core/types";
|
|
|
42
42
|
import type { CorpusAdapter } from "../../types.ts";
|
|
43
43
|
import { type WOFRecord } from "../../wof-json.ts";
|
|
44
44
|
interface VariantSpec {
|
|
45
|
-
/**
|
|
45
|
+
/**
|
|
46
|
+
* Hierarchy-variant id appended to `source_id`.
|
|
47
|
+
*/
|
|
46
48
|
suffix: string;
|
|
47
|
-
/**
|
|
49
|
+
/**
|
|
50
|
+
* Component tag → display string the adapter will hand to the runner.
|
|
51
|
+
*/
|
|
48
52
|
components: Partial<Record<ComponentTag, string>>;
|
|
49
53
|
}
|
|
50
54
|
/**
|
|
@@ -70,6 +74,10 @@ export declare function nameSlotsFor(rec: WOFRecord): Array<{
|
|
|
70
74
|
key: string;
|
|
71
75
|
value: string;
|
|
72
76
|
}>;
|
|
77
|
+
/**
|
|
78
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
79
|
+
* it came from.
|
|
80
|
+
*/
|
|
73
81
|
export declare const WOF_ADMIN_ADAPTER_ID = "wof-admin";
|
|
74
82
|
/**
|
|
75
83
|
* Construct the wof-admin JSON-bundle adapter. The adapter is stateless across runs; calling this twice with the same
|
|
@@ -77,7 +85,9 @@ export declare const WOF_ADMIN_ADAPTER_ID = "wof-admin";
|
|
|
77
85
|
* insensitive to filesystem walk ordering).
|
|
78
86
|
*/
|
|
79
87
|
export declare function createWOFAdminAdapter(): CorpusAdapter;
|
|
80
|
-
/**
|
|
88
|
+
/**
|
|
89
|
+
* Single shared instance, suitable for `defaultAdapterRegistry`.
|
|
90
|
+
*/
|
|
81
91
|
export declare const wofAdminAdapter: CorpusAdapter;
|
|
82
92
|
export {};
|
|
83
93
|
//# sourceMappingURL=adapter.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AAGH,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAGzD,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AACjF,OAAO,EAAsD,KAAK,SAAS,EAAE,MAAM,mBAAmB,CAAA;
|
|
1
|
+
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AAGH,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAGzD,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AACjF,OAAO,EAAsD,KAAK,SAAS,EAAE,MAAM,mBAAmB,CAAA;AAoDtG,UAAU,WAAW;IACpB;;OAEG;IACH,MAAM,EAAE,MAAM,CAAA;IAEd;;OAEG;IACH,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAA;CACjD;AAED;;;;;;;;;GASG;AACH,wBAAgB,WAAW,CAAC,GAAG,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,EAAE,QAAQ,EAAE,MAAM,GAAG,WAAW,EAAE,CAkElG;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAAC,GAAG,EAAE,SAAS,GAAG,KAAK,CAAC;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CAclF;AAED;;;GAGG;AACH,eAAO,MAAM,oBAAoB,cAAc,CAAA;AAE/C;;;;GAIG;AACH,wBAAgB,qBAAqB,IAAI,aAAa,CAkErD;AAED;;GAEG;AACH,eAAO,MAAM,eAAe,eAA0B,CAAA"}
|
|
@@ -54,12 +54,16 @@ const COUNTRY_DISPLAY_NAME = {
|
|
|
54
54
|
US: "United States of America",
|
|
55
55
|
FR: "France",
|
|
56
56
|
};
|
|
57
|
-
/**
|
|
57
|
+
/**
|
|
58
|
+
* BCP-47 locale defaulting for the corpus row's `locale` field. Defaulted by country.
|
|
59
|
+
*/
|
|
58
60
|
const LOCALE_BY_COUNTRY = {
|
|
59
61
|
US: "en-US",
|
|
60
62
|
FR: "fr-FR",
|
|
61
63
|
};
|
|
62
|
-
/**
|
|
64
|
+
/**
|
|
65
|
+
* Map a WOF placetype to a Mailwoman `ComponentTag`, or `undefined` to skip.
|
|
66
|
+
*/
|
|
63
67
|
function placetypeToTag(placetype) {
|
|
64
68
|
switch (placetype) {
|
|
65
69
|
case "country":
|
|
@@ -168,6 +172,10 @@ export function nameSlotsFor(rec) {
|
|
|
168
172
|
}
|
|
169
173
|
return slots;
|
|
170
174
|
}
|
|
175
|
+
/**
|
|
176
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
177
|
+
* it came from.
|
|
178
|
+
*/
|
|
171
179
|
export const WOF_ADMIN_ADAPTER_ID = "wof-admin";
|
|
172
180
|
/**
|
|
173
181
|
* Construct the wof-admin JSON-bundle adapter. The adapter is stateless across runs; calling this twice with the same
|
|
@@ -197,7 +205,7 @@ export function createWOFAdminAdapter() {
|
|
|
197
205
|
}
|
|
198
206
|
const ancestry = buildAncestryIndex(byID);
|
|
199
207
|
// Pass 2: emit rows in sorted-id order for deterministic JSONL.
|
|
200
|
-
const ids = [...byID.keys()].
|
|
208
|
+
const ids = [...byID.keys()].toSorted((a, b) => a - b);
|
|
201
209
|
let emitted = 0;
|
|
202
210
|
for (const id of ids) {
|
|
203
211
|
if (opts.signal?.aborted)
|
|
@@ -214,7 +222,7 @@ export function createWOFAdminAdapter() {
|
|
|
214
222
|
if (!raw)
|
|
215
223
|
continue;
|
|
216
224
|
const aligned = reconcileComponents(variant.components, raw);
|
|
217
|
-
if (Object.keys(aligned).length
|
|
225
|
+
if (!Object.keys(aligned).length)
|
|
218
226
|
continue;
|
|
219
227
|
yield {
|
|
220
228
|
raw,
|
|
@@ -233,6 +241,8 @@ export function createWOFAdminAdapter() {
|
|
|
233
241
|
},
|
|
234
242
|
};
|
|
235
243
|
}
|
|
236
|
-
/**
|
|
244
|
+
/**
|
|
245
|
+
* Single shared instance, suitable for `defaultAdapterRegistry`.
|
|
246
|
+
*/
|
|
237
247
|
export const wofAdminAdapter = createWOFAdminAdapter();
|
|
238
248
|
//# sourceMappingURL=adapter.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AAKH,OAAO,EAAE,aAAa,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAEpE,OAAO,EAAE,kBAAkB,EAAE,gBAAgB,EAAE,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AAEtG;;;;;;;;;GASG;AACH,MAAM,oBAAoB,GAA2B;IACpD,EAAE,EAAE,0BAA0B;IAC9B,EAAE,EAAE,QAAQ;CACZ,CAAA;AAED
|
|
1
|
+
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/wof-admin-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AAKH,OAAO,EAAE,aAAa,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAEpE,OAAO,EAAE,kBAAkB,EAAE,gBAAgB,EAAE,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AAEtG;;;;;;;;;GASG;AACH,MAAM,oBAAoB,GAA2B;IACpD,EAAE,EAAE,0BAA0B;IAC9B,EAAE,EAAE,QAAQ;CACZ,CAAA;AAED;;GAEG;AACH,MAAM,iBAAiB,GAA2B;IACjD,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,OAAO;CACX,CAAA;AAED;;GAEG;AACH,SAAS,cAAc,CAAC,SAAwC;IAC/D,QAAQ,SAAS,EAAE,CAAC;QACnB,KAAK,SAAS,CAAC;QACf,KAAK,QAAQ;YACZ,OAAO,SAAS,CAAA;QACjB,KAAK,aAAa,CAAC;QACnB,KAAK,QAAQ;YACZ,OAAO,QAAQ,CAAA;QAChB,KAAK,aAAa,CAAC;QACnB,KAAK,QAAQ,CAAC;QACd,KAAK,YAAY;YAChB,OAAO,WAAW,CAAA;QACnB,KAAK,UAAU;YACd,OAAO,UAAU,CAAA;QAClB,KAAK,SAAS,CAAC;QACf,KAAK,WAAW,CAAC;QACjB,KAAK,eAAe,CAAC;QACrB,KAAK,WAAW;YACf,OAAO,oBAAoB,CAAA;QAC5B;YACC,OAAO,SAAS,CAAA;IAClB,CAAC;AACF,CAAC;AAcD;;;;;;;;;GASG;AACH,MAAM,UAAU,WAAW,CAAC,GAAc,EAAE,QAAqB,EAAE,QAAgB;IAClF,MAAM,OAAO,GAAG,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;IAE7C,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,CAAA;IAEvB,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,QAAQ,CAAC,CAAA;IAC7E,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,SAAS,CAAC,CAAA;IAC/E,MAAM,cAAc,GAAG,oBAAoB,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,OAAO,EAAE,IAAI,IAAI,GAAG,CAAC,OAAO,CAAA;IAExF,MAAM,QAAQ,GAAkB,EAAE,CAAA;IAElC,QAAQ,OAAO,EAAE,CAAC;QACjB,KAAK,UAAU,CAAC;QAChB,KAAK,oBAAoB,CAAC,CAAC,CAAC;YAC3B,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAA;YAEtE,IAAI,MAAM,EAAE,CAAC;gBACZ,QAAQ,CAAC,IAAI,CAAC;oBACb,MAAM,EAAE,aAAa;oBACrB,UAAU,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,IAAI,EAAE;iBACxD,CAAC,CAAA;YACH,CAAC;YAED,IAAI,MAAM,IAAI,OAAO,EAAE,CAAC;gBACvB,QAAQ,CAAC,IAAI,CAAC;oBACb,MAAM,EAAE,qBAAqB;oBAC7B,UAAU,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,IAAI,EAAE,OAAO,EAAE,cAAc,EAAE;iBACjF,CAAC,CAAA;YACH,CAAC;iBAAM,IAAI,CAAC,MAAM,IAAI,OAAO,EAAE,CAAC;gBAC/B,QAAQ,CAAC,IAAI,CAAC;oBACb,MAAM,EAAE,cAAc;oBACtB,UAAU,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,OAAO,EAAE,cAAc,EAAE;iBAC5D,CAAC,CAAA;YACH,CAAC;YAED,OAAO,QAAQ,CAAA;QAChB,CAAC;QAED,KAAK,QAAQ,CAAC,CAAC,CAAC;YACf,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAA;YAEnE,IAAI,OAAO,EAAE,CAAC;gBACb,QAAQ,CAAC,IAAI,CAAC;oBACb,MAAM,EAAE,cAAc;oBACtB,UAAU,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,cAAc,EAAE;iBACzD,CAAC,CAAA;YACH,CAAC;YAED,OAAO,QAAQ,CAAA;QAChB,CAAC;QAED,KAAK,SAAS,CAAC,CAAC,CAAC;YAChB,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAA;YAEpE,OAAO,QAAQ,CAAA;QAChB,CAAC;QAED,KAAK,WAAW,CAAC,CAAC,CAAC;YAClB,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,SAAS,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAA;YAEtE,OAAO,QAAQ,CAAA;QAChB,CAAC;QAED;YACC,OAAO,EAAE,CAAA;IACX,CAAC;AACF,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAAC,GAAc;IAC1C,MAAM,OAAO,GAAG,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;IAC7C,MAAM,iBAAiB,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,CAAC,oBAAoB,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAA;IAE5G,MAAM,IAAI,GAAG,IAAI,GAAG,CAAS,CAAC,iBAAiB,CAAC,CAAC,CAAA;IACjD,MAAM,KAAK,GAA0C,CAAC,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,iBAAiB,EAAE,CAAC,CAAA;IAEnG,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,GAAG,CAAC,YAAY,EAAE,CAAC;QAChD,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;YAAE,SAAQ;QAC7B,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAA;QACf,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,gBAAgB,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAA;IACrD,CAAC;IAED,OAAO,KAAK,CAAA;AACb,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,WAAW,CAAA;AAE/C;;;;GAIG;AACH,MAAM,UAAU,qBAAqB;IACpC,OAAO;QACN,EAAE,EAAE,oBAAoB;QACxB,cAAc,EAAE,SAAS;QACzB,WAAW,EACV,mHAAmH;QAEpH,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,0EAA0E;YAC1E,sFAAsF;YACtF,wFAAwF;YACxF,qFAAqF;YACrF,wFAAwF;YACxF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAqB,CAAA;YAEzC,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,CAAC;gBAC/E,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,OAAM;gBAEhC,IAAI,IAAI,CAAC,OAAO,IAAI,GAAG,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO;oBAAE,SAAQ;gBAE1D,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC;oBAAE,SAAQ;gBAC5C,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,CAAA;YACtB,CAAC;YAED,MAAM,QAAQ,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAA;YAEzC,gEAAgE;YAChE,MAAM,GAAG,GAAG,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;YACtD,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;gBACtB,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,OAAM;gBAChC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,CAAE,CAAA;gBACzB,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,EAAE,CAAA;gBACpC,MAAM,KAAK,GAAG,YAAY,CAAC,GAAG,CAAC,CAAA;gBAE/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBAC1B,MAAM,QAAQ,GAAG,WAAW,CAAC,GAAG,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAA;oBAEpD,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;wBAChC,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;4BAAE,OAAM;wBAE7D,MAAM,GAAG,GAAG,aAAa,CAAC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;wBAE/E,IAAI,CAAC,GAAG;4BAAE,SAAQ;wBAClB,MAAM,OAAO,GAAG,mBAAmB,CAAC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;wBAE5D,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM;4BAAE,SAAQ;wBAE1C,MAAM;4BACL,GAAG;4BACH,UAAU,EAAE,OAAO;4BACnB,OAAO,EAAE,GAAG,CAAC,OAAO;4BACpB,MAAM,EAAE,iBAAiB,CAAC,GAAG,CAAC,OAAO,CAAC;4BACtC,MAAM,EAAE,oBAAoB;4BAC5B,SAAS,EAAE,GAAG,oBAAoB,IAAI,GAAG,CAAC,EAAE,IAAI,IAAI,CAAC,GAAG,IAAI,OAAO,CAAC,MAAM,EAAE;4BAC5E,cAAc,EAAE,EAAE;4BAClB,OAAO,EAAE,SAAS;yBAClB,CAAA;wBAED,OAAO,EAAE,CAAA;oBACV,CAAC;gBACF,CAAC;YACF,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,qBAAqB,EAAE,CAAA"}
|
|
@@ -56,8 +56,15 @@ export declare function nameSlotsFor(rec: WOFRecord): Array<{
|
|
|
56
56
|
key: string;
|
|
57
57
|
value: string;
|
|
58
58
|
}>;
|
|
59
|
+
/**
|
|
60
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
61
|
+
* it came from.
|
|
62
|
+
*/
|
|
59
63
|
export declare const WOF_POSTALCODE_ADAPTER_ID = "wof-postalcode";
|
|
60
64
|
export declare function createWOFPostalcodeAdapter(): CorpusAdapter;
|
|
65
|
+
/**
|
|
66
|
+
* The configured adapter instance registered with the corpus builder.
|
|
67
|
+
*/
|
|
61
68
|
export declare const wofPostalcodeAdapter: CorpusAdapter;
|
|
62
69
|
export {};
|
|
63
70
|
//# sourceMappingURL=adapter.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/wof-postalcode-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqCG;AAGH,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAGzD,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AACjF,OAAO,EAAsD,KAAK,SAAS,EAAE,MAAM,mBAAmB,CAAA;AA6BtG,UAAU,WAAW;IACpB,MAAM,EAAE,MAAM,CAAA;IACd,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAA;CACjD;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,EAAE,QAAQ,EAAE,MAAM,GAAG,WAAW,EAAE,CAqC1G;AAED;;;GAGG;AACH,wBAAgB,YAAY,CAAC,GAAG,EAAE,SAAS,GAAG,KAAK,CAAC;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CAWlF;AAED,eAAO,MAAM,yBAAyB,mBAAmB,CAAA;AAEzD,wBAAgB,0BAA0B,IAAI,aAAa,
|
|
1
|
+
{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../src/adapters/wof-postalcode-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqCG;AAGH,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAGzD,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,gBAAgB,CAAA;AACjF,OAAO,EAAsD,KAAK,SAAS,EAAE,MAAM,mBAAmB,CAAA;AA6BtG,UAAU,WAAW;IACpB,MAAM,EAAE,MAAM,CAAA;IACd,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAA;CACjD;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,EAAE,QAAQ,EAAE,MAAM,GAAG,WAAW,EAAE,CAqC1G;AAED;;;GAGG;AACH,wBAAgB,YAAY,CAAC,GAAG,EAAE,SAAS,GAAG,KAAK,CAAC;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CAWlF;AAED;;;GAGG;AACH,eAAO,MAAM,yBAAyB,mBAAmB,CAAA;AAEzD,wBAAgB,0BAA0B,IAAI,aAAa,CAmE1D;AAED;;GAEG;AACH,eAAO,MAAM,oBAAoB,eAA+B,CAAA"}
|
|
@@ -114,6 +114,10 @@ export function nameSlotsFor(rec) {
|
|
|
114
114
|
}
|
|
115
115
|
return slots;
|
|
116
116
|
}
|
|
117
|
+
/**
|
|
118
|
+
* Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
|
|
119
|
+
* it came from.
|
|
120
|
+
*/
|
|
117
121
|
export const WOF_POSTALCODE_ADAPTER_ID = "wof-postalcode";
|
|
118
122
|
export function createWOFPostalcodeAdapter() {
|
|
119
123
|
return {
|
|
@@ -136,7 +140,7 @@ export function createWOFPostalcodeAdapter() {
|
|
|
136
140
|
}
|
|
137
141
|
const ancestry = buildAncestryIndex(byID);
|
|
138
142
|
// Pass 2: emit postcode rows only, sorted by id for determinism.
|
|
139
|
-
const ids = [...byID.keys()].
|
|
143
|
+
const ids = [...byID.keys()].toSorted((a, b) => a - b);
|
|
140
144
|
let emitted = 0;
|
|
141
145
|
for (const id of ids) {
|
|
142
146
|
if (opts.signal?.aborted)
|
|
@@ -155,7 +159,7 @@ export function createWOFPostalcodeAdapter() {
|
|
|
155
159
|
if (!raw)
|
|
156
160
|
continue;
|
|
157
161
|
const aligned = reconcileComponents(variant.components, raw);
|
|
158
|
-
if (Object.keys(aligned).length
|
|
162
|
+
if (!Object.keys(aligned).length)
|
|
159
163
|
continue;
|
|
160
164
|
yield {
|
|
161
165
|
raw,
|
|
@@ -174,5 +178,8 @@ export function createWOFPostalcodeAdapter() {
|
|
|
174
178
|
},
|
|
175
179
|
};
|
|
176
180
|
}
|
|
181
|
+
/**
|
|
182
|
+
* The configured adapter instance registered with the corpus builder.
|
|
183
|
+
*/
|
|
177
184
|
export const wofPostalcodeAdapter = createWOFPostalcodeAdapter();
|
|
178
185
|
//# sourceMappingURL=adapter.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/wof-postalcode-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqCG;AAKH,OAAO,EAAE,aAAa,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAEpE,OAAO,EAAE,kBAAkB,EAAE,gBAAgB,EAAE,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AAEtG,MAAM,oBAAoB,GAA2B;IACpD,EAAE,EAAE,0BAA0B;IAC9B,EAAE,EAAE,QAAQ;CACZ,CAAA;AAED,MAAM,iBAAiB,GAA2B;IACjD,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,OAAO;CACX,CAAA;AAED,SAAS,cAAc,CAAC,SAAwC;IAC/D,QAAQ,SAAS,EAAE,CAAC;QACnB,KAAK,SAAS,CAAC;QACf,KAAK,QAAQ;YACZ,OAAO,SAAS,CAAA;QACjB,KAAK,aAAa,CAAC;QACnB,KAAK,QAAQ;YACZ,OAAO,QAAQ,CAAA;QAChB,KAAK,UAAU;YACd,OAAO,UAAU,CAAA;QAClB,KAAK,YAAY;YAChB,OAAO,UAAU,CAAA;QAClB;YACC,OAAO,SAAS,CAAA;IAClB,CAAC;AACF,CAAC;AAOD;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,GAAc,EAAE,QAAqB,EAAE,QAAgB;IAC1F,IAAI,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,UAAU;QAAE,OAAO,EAAE,CAAA;IAE3D,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,UAAU,CAAC,CAAA;IACjF,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,QAAQ,CAAC,CAAA;IAC7E,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,SAAS,CAAC,CAAA;IAC/E,MAAM,cAAc,GAAG,oBAAoB,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,OAAO,EAAE,IAAI,IAAI,GAAG,CAAC,OAAO,CAAA;IAExF,MAAM,QAAQ,GAAkB,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAA;IAExF,IAAI,QAAQ,EAAE,CAAC;QACd,QAAQ,CAAC,IAAI,CAAC;YACb,MAAM,EAAE,eAAe;YACvB,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,IAAI,EAAE;SAC3D,CAAC,CAAA;IACH,CAAC;IAED,IAAI,QAAQ,IAAI,MAAM,EAAE,CAAC;QACxB,QAAQ,CAAC,IAAI,CAAC;YACb,MAAM,EAAE,sBAAsB;YAC9B,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,MAAM,CAAC,IAAI,EAAE;SAChF,CAAC,CAAA;IACH,CAAC;IAED,IAAI,QAAQ,IAAI,MAAM,IAAI,OAAO,EAAE,CAAC;QACnC,QAAQ,CAAC,IAAI,CAAC;YACb,MAAM,EAAE,8BAA8B;YACtC,UAAU,EAAE;gBACX,QAAQ,EAAE,QAAQ;gBAClB,QAAQ,EAAE,QAAQ,CAAC,IAAI;gBACvB,MAAM,EAAE,MAAM,CAAC,IAAI;gBACnB,OAAO,EAAE,cAAc;aACvB;SACD,CAAC,CAAA;IACH,CAAC;IAED,OAAO,QAAQ,CAAA;AAChB,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAAC,GAAc;IAC1C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAA;IACxC,MAAM,KAAK,GAA0C,CAAC,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,GAAG,CAAC,IAAI,EAAE,CAAC,CAAA;IAE1F,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,GAAG,CAAC,YAAY,EAAE,CAAC;QAChD,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;YAAE,SAAQ;QAC7B,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAA;QACf,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,gBAAgB,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAA;IACrD,CAAC;IAED,OAAO,KAAK,CAAA;AACb,CAAC;AAED,MAAM,CAAC,MAAM,yBAAyB,GAAG,gBAAgB,CAAA;AAEzD,MAAM,UAAU,0BAA0B;IACzC,OAAO;QACN,EAAE,EAAE,yBAAyB;QAC7B,cAAc,EAAE,SAAS;QACzB,WAAW,EACV,wHAAwH;QAEzH,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,uFAAuF;YACvF,sFAAsF;YACtF,sFAAsF;YACtF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAqB,CAAA;YAEzC,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,CAAC;gBAC/E,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,OAAM;gBAEhC,IAAI,IAAI,CAAC,OAAO,IAAI,GAAG,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO;oBAAE,SAAQ;gBAE1D,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC;oBAAE,SAAQ;gBAC5C,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,CAAA;YACtB,CAAC;YAED,MAAM,QAAQ,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAA;YAEzC,iEAAiE;YACjE,MAAM,GAAG,GAAG,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,
|
|
1
|
+
{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../src/adapters/wof-postalcode-json/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqCG;AAKH,OAAO,EAAE,aAAa,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAA;AAEpE,OAAO,EAAE,kBAAkB,EAAE,gBAAgB,EAAE,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AAEtG,MAAM,oBAAoB,GAA2B;IACpD,EAAE,EAAE,0BAA0B;IAC9B,EAAE,EAAE,QAAQ;CACZ,CAAA;AAED,MAAM,iBAAiB,GAA2B;IACjD,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,OAAO;CACX,CAAA;AAED,SAAS,cAAc,CAAC,SAAwC;IAC/D,QAAQ,SAAS,EAAE,CAAC;QACnB,KAAK,SAAS,CAAC;QACf,KAAK,QAAQ;YACZ,OAAO,SAAS,CAAA;QACjB,KAAK,aAAa,CAAC;QACnB,KAAK,QAAQ;YACZ,OAAO,QAAQ,CAAA;QAChB,KAAK,UAAU;YACd,OAAO,UAAU,CAAA;QAClB,KAAK,YAAY;YAChB,OAAO,UAAU,CAAA;QAClB;YACC,OAAO,SAAS,CAAA;IAClB,CAAC;AACF,CAAC;AAOD;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,GAAc,EAAE,QAAqB,EAAE,QAAgB;IAC1F,IAAI,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,UAAU;QAAE,OAAO,EAAE,CAAA;IAE3D,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,UAAU,CAAC,CAAA;IACjF,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,QAAQ,CAAC,CAAA;IAC7E,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,SAAS,CAAC,CAAA;IAC/E,MAAM,cAAc,GAAG,oBAAoB,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,OAAO,EAAE,IAAI,IAAI,GAAG,CAAC,OAAO,CAAA;IAExF,MAAM,QAAQ,GAAkB,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAA;IAExF,IAAI,QAAQ,EAAE,CAAC;QACd,QAAQ,CAAC,IAAI,CAAC;YACb,MAAM,EAAE,eAAe;YACvB,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,IAAI,EAAE;SAC3D,CAAC,CAAA;IACH,CAAC;IAED,IAAI,QAAQ,IAAI,MAAM,EAAE,CAAC;QACxB,QAAQ,CAAC,IAAI,CAAC;YACb,MAAM,EAAE,sBAAsB;YAC9B,UAAU,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,MAAM,CAAC,IAAI,EAAE;SAChF,CAAC,CAAA;IACH,CAAC;IAED,IAAI,QAAQ,IAAI,MAAM,IAAI,OAAO,EAAE,CAAC;QACnC,QAAQ,CAAC,IAAI,CAAC;YACb,MAAM,EAAE,8BAA8B;YACtC,UAAU,EAAE;gBACX,QAAQ,EAAE,QAAQ;gBAClB,QAAQ,EAAE,QAAQ,CAAC,IAAI;gBACvB,MAAM,EAAE,MAAM,CAAC,IAAI;gBACnB,OAAO,EAAE,cAAc;aACvB;SACD,CAAC,CAAA;IACH,CAAC;IAED,OAAO,QAAQ,CAAA;AAChB,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAAC,GAAc;IAC1C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAA;IACxC,MAAM,KAAK,GAA0C,CAAC,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,GAAG,CAAC,IAAI,EAAE,CAAC,CAAA;IAE1F,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,GAAG,CAAC,YAAY,EAAE,CAAC;QAChD,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;YAAE,SAAQ;QAC7B,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAA;QACf,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,gBAAgB,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAA;IACrD,CAAC;IAED,OAAO,KAAK,CAAA;AACb,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,yBAAyB,GAAG,gBAAgB,CAAA;AAEzD,MAAM,UAAU,0BAA0B;IACzC,OAAO;QACN,EAAE,EAAE,yBAAyB;QAC7B,cAAc,EAAE,SAAS;QACzB,WAAW,EACV,wHAAwH;QAEzH,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,uFAAuF;YACvF,sFAAsF;YACtF,sFAAsF;YACtF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAqB,CAAA;YAEzC,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,CAAC;gBAC/E,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,OAAM;gBAEhC,IAAI,IAAI,CAAC,OAAO,IAAI,GAAG,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO;oBAAE,SAAQ;gBAE1D,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC;oBAAE,SAAQ;gBAC5C,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,CAAA;YACtB,CAAC;YAED,MAAM,QAAQ,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAA;YAEzC,iEAAiE;YACjE,MAAM,GAAG,GAAG,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;YACtD,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;gBACtB,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,OAAM;gBAChC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,CAAE,CAAA;gBAEzB,IAAI,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,UAAU;oBAAE,SAAQ;gBAE1D,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,EAAE,CAAA;gBACpC,MAAM,KAAK,GAAG,YAAY,CAAC,GAAG,CAAC,CAAA;gBAE/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBAC1B,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAA;oBAE5D,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;wBAChC,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;4BAAE,OAAM;wBAE7D,MAAM,GAAG,GAAG,aAAa,CAAC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;wBAE/E,IAAI,CAAC,GAAG;4BAAE,SAAQ;wBAClB,MAAM,OAAO,GAAG,mBAAmB,CAAC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;wBAE5D,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM;4BAAE,SAAQ;wBAE1C,MAAM;4BACL,GAAG;4BACH,UAAU,EAAE,OAAO;4BACnB,OAAO,EAAE,GAAG,CAAC,OAAO;4BACpB,MAAM,EAAE,iBAAiB,CAAC,GAAG,CAAC,OAAO,CAAC;4BACtC,MAAM,EAAE,yBAAyB;4BACjC,SAAS,EAAE,GAAG,yBAAyB,IAAI,GAAG,CAAC,EAAE,IAAI,IAAI,CAAC,GAAG,IAAI,OAAO,CAAC,MAAM,EAAE;4BACjF,cAAc,EAAE,EAAE;4BAClB,OAAO,EAAE,SAAS;yBAClB,CAAA;wBAED,OAAO,EAAE,CAAA;oBACV,CAAC;gBACF,CAAC;YACF,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,0BAA0B,EAAE,CAAA"}
|
package/out/src/align.d.ts
CHANGED
|
@@ -35,9 +35,13 @@
|
|
|
35
35
|
import type { ComponentTag } from "@mailwoman/core/types";
|
|
36
36
|
import { type Tokenizer } from "./tokenize.ts";
|
|
37
37
|
import type { CanonicalRow, LabeledRow, QuarantinedRow } from "./types.ts";
|
|
38
|
-
/**
|
|
38
|
+
/**
|
|
39
|
+
* Options for `alignRow`.
|
|
40
|
+
*/
|
|
39
41
|
export interface AlignOptions {
|
|
40
|
-
/**
|
|
42
|
+
/**
|
|
43
|
+
* Tokenizer to use. Defaults to `whitespaceTokenizer()`.
|
|
44
|
+
*/
|
|
41
45
|
tokenizer?: Tokenizer;
|
|
42
46
|
/**
|
|
43
47
|
* Max Levenshtein edit distance to accept when a verbatim substring match fails. Set `0` to require verbatim matches
|
|
@@ -53,7 +57,9 @@ export interface AlignOptions {
|
|
|
53
57
|
*/
|
|
54
58
|
caseInsensitive?: boolean;
|
|
55
59
|
}
|
|
56
|
-
/**
|
|
60
|
+
/**
|
|
61
|
+
* Either a successful labeled row or a quarantined one.
|
|
62
|
+
*/
|
|
57
63
|
export type AlignmentResult = {
|
|
58
64
|
kind: "labeled";
|
|
59
65
|
row: LabeledRow;
|
|
@@ -70,7 +76,9 @@ export interface ComponentSpan {
|
|
|
70
76
|
start: number;
|
|
71
77
|
end: number;
|
|
72
78
|
}
|
|
73
|
-
/**
|
|
79
|
+
/**
|
|
80
|
+
* Align a single row.
|
|
81
|
+
*/
|
|
74
82
|
export declare function alignRow(row: CanonicalRow, opts?: AlignOptions): AlignmentResult;
|
|
75
83
|
/**
|
|
76
84
|
* Enforce the #519 span-triple invariants — in-bounds, sorted ascending by start, non-overlapping — loudly.
|
package/out/src/align.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"align.d.ts","sourceRoot":"","sources":["../../src/align.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAiCG;AAEH,OAAO,KAAK,EAAY,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAGnE,OAAO,EAAuC,KAAK,SAAS,EAAE,MAAM,eAAe,CAAA;AACnF,OAAO,KAAK,EAAE,YAAY,EAAE,UAAU,EAAE,cAAc,EAAE,MAAM,YAAY,CAAA;AAE1E
|
|
1
|
+
{"version":3,"file":"align.d.ts","sourceRoot":"","sources":["../../src/align.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAiCG;AAEH,OAAO,KAAK,EAAY,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAGnE,OAAO,EAAuC,KAAK,SAAS,EAAE,MAAM,eAAe,CAAA;AACnF,OAAO,KAAK,EAAE,YAAY,EAAE,UAAU,EAAE,cAAc,EAAE,MAAM,YAAY,CAAA;AAE1E;;GAEG;AACH,MAAM,WAAW,YAAY;IAC5B;;OAEG;IACH,SAAS,CAAC,EAAE,SAAS,CAAA;IAErB;;;;;;OAMG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IAExB;;;OAGG;IACH,eAAe,CAAC,EAAE,OAAO,CAAA;CACzB;AAED;;GAEG;AACH,MAAM,MAAM,eAAe,GAAG;IAAE,IAAI,EAAE,SAAS,CAAC;IAAC,GAAG,EAAE,UAAU,CAAA;CAAE,GAAG;IAAE,IAAI,EAAE,aAAa,CAAC;IAAC,GAAG,EAAE,cAAc,CAAA;CAAE,CAAA;AAEjH;;;GAGG;AACH,MAAM,WAAW,aAAa;IAC7B,GAAG,EAAE,YAAY,CAAA;IACjB,KAAK,EAAE,MAAM,CAAA;IACb,GAAG,EAAE,MAAM,CAAA;CACX;AAED;;GAEG;AACH,wBAAgB,QAAQ,CAAC,GAAG,EAAE,YAAY,EAAE,IAAI,GAAE,YAAiB,GAAG,eAAe,CAsFpF;AAED;;;;;;;;GAQG;AACH,wBAAgB,oBAAoB,CACnC,KAAK,EAAE,SAAS,aAAa,EAAE,EAC/B,GAAG,EAAE,IAAI,CAAC,YAAY,EAAE,KAAK,GAAG,QAAQ,GAAG,WAAW,CAAC,GACrD,IAAI,CA4BN"}
|
package/out/src/align.js
CHANGED
|
@@ -34,7 +34,9 @@
|
|
|
34
34
|
*/
|
|
35
35
|
import { distance as levenshteinDistance } from "fastest-levenshtein";
|
|
36
36
|
import { whitespaceTokenizer } from "./tokenize.js";
|
|
37
|
-
/**
|
|
37
|
+
/**
|
|
38
|
+
* Align a single row.
|
|
39
|
+
*/
|
|
38
40
|
export function alignRow(row, opts = {}) {
|
|
39
41
|
const tokenizer = opts.tokenizer ?? whitespaceTokenizer();
|
|
40
42
|
const maxEditDistance = opts.maxEditDistance ?? 2;
|
|
@@ -61,7 +63,7 @@ export function alignRow(row, opts = {}) {
|
|
|
61
63
|
// component owns ("Alaska Regional Dr, Alaska" — region "Alaska" stealing the street's first
|
|
62
64
|
// word quarantined the street; pilot2's residual class). Emit order is unaffected — spans are
|
|
63
65
|
// re-sorted by start below.
|
|
64
|
-
const entries = Object.entries(components).
|
|
66
|
+
const entries = Object.entries(components).toSorted((a, b) => (b[1]?.length ?? 0) - (a[1]?.length ?? 0));
|
|
65
67
|
for (const [tag, value] of entries) {
|
|
66
68
|
if (!value)
|
|
67
69
|
continue;
|
|
@@ -144,7 +146,7 @@ export function assertSpanInvariants(spans, row) {
|
|
|
144
146
|
*/
|
|
145
147
|
function locateSpan(args) {
|
|
146
148
|
const { haystack, needle, claimed, maxEditDistance } = args;
|
|
147
|
-
if (needle.length
|
|
149
|
+
if (!needle.length)
|
|
148
150
|
return undefined;
|
|
149
151
|
// Pass 1: verbatim substring. Word-boundary-aligned matches are PREFERRED over intra-word ones
|
|
150
152
|
// — leftmost-substring alone let a short value claim the inside of an earlier word (region "AK"
|
|
@@ -156,7 +158,7 @@ function locateSpan(args) {
|
|
|
156
158
|
let from = 0;
|
|
157
159
|
while (true) {
|
|
158
160
|
const idx = haystack.indexOf(needle, from);
|
|
159
|
-
if (idx
|
|
161
|
+
if (idx === -1)
|
|
160
162
|
break;
|
|
161
163
|
const end = idx + needle.length;
|
|
162
164
|
if (!overlapsClaimed(idx, end, claimed)) {
|
|
@@ -186,7 +188,9 @@ function locateSpan(args) {
|
|
|
186
188
|
return undefined;
|
|
187
189
|
}
|
|
188
190
|
const WORD_CHAR = /[\p{L}\p{N}]/u;
|
|
189
|
-
/**
|
|
191
|
+
/**
|
|
192
|
+
* Both needle edges sit on word boundaries of the haystack (string edges count as boundaries).
|
|
193
|
+
*/
|
|
190
194
|
function isBoundaryAligned(haystack, start, end) {
|
|
191
195
|
const before = start === 0 || !WORD_CHAR.test(haystack[start - 1]);
|
|
192
196
|
const after = end === haystack.length || !WORD_CHAR.test(haystack[end]);
|