@mailwoman/corpus 9.1.0 → 9.2.0

package/package.json

@@ -1,12 +1,12 @@
 {
   "name": "@mailwoman/corpus",
-  "version": "9.1.0",
+  "version": "9.2.0",
   "description": "Mailwoman corpus pipeline: BIO-labeled dataset builder for the neural classifier.",
   "license": "AGPL-3.0-only OR LicenseRef-Commercial",
   "repository": {
     "type": "git",
     "url": "https://github.com/sister-software/mailwoman.git",
-    "directory": "corpus"
+    "directory": "packages/corpus"
   },
   "files": [
     "data/**",
@@ -22,11 +22,55 @@
     "!*.test.tsx",
     "!**/*.test.ts",
     "!**/*.test.tsx",
-    "!test-kit/**/*"
+    "!test-kit/**/*",
+    "!test/**"
   ],
   "type": "module",
   "imports": {
-    "#test-kit": "./test-kit/index.ts"
+    "#adapters/utils": {
+      "types": "./out/src/adapters/utils/index.d.ts",
+      "default": "./out/src/adapters/utils/index.js"
+    },
+    "#runner": {
+      "types": "./out/src/runner.d.ts",
+      "default": "./out/src/runner.js"
+    },
+    "#synthesizers/utils": {
+      "types": "./out/src/synthesizers/utils.d.ts",
+      "default": "./out/src/synthesizers/utils.js"
+    },
+    "#types": {
+      "types": "./out/src/types.d.ts",
+      "default": "./out/src/types.js"
+    },
+    "#utils": {
+      "types": "./out/src/utils/index.d.ts",
+      "default": "./out/src/utils/index.js"
+    },
+    "#utils/*": {
+      "types": "./out/src/utils/*.d.ts",
+      "default": "./out/src/utils/*.js"
+    },
+    "#codex/*": {
+      "types": "./out/src/codex/*.d.ts",
+      "default": "./out/src/codex/*.js"
+    },
+    "#synthesizers/*": {
+      "types": "./out/src/synthesizers/*.d.ts",
+      "default": "./out/src/synthesizers/*.js"
+    },
+    "#tools": {
+      "types": "./out/src/tools/index.d.ts",
+      "default": "./out/src/tools/index.js"
+    },
+    "#adapters/*": {
+      "types": "./out/src/adapters/*.d.ts",
+      "default": "./out/src/adapters/*.js"
+    },
+    "#parquet-wrapper": {
+      "types": "./out/src/parquet-wrapper/index.d.ts",
+      "default": "./out/src/parquet-wrapper/index.js"
+    }
   },
   "exports": {
     "./package.json": "./package.json",
@@ -59,6 +103,10 @@
       "types": "./out/src/synthesizers/*.d.ts",
       "default": "./out/src/synthesizers/*.js"
     },
+    "./shard-recipes/*": {
+      "types": "./out/src/shard-recipes/*.d.ts",
+      "default": "./out/src/shard-recipes/*.js"
+    },
     "./adapters/utils": {
       "types": "./out/src/adapters/utils/index.d.ts",
       "default": "./out/src/adapters/utils/index.js"
@@ -78,6 +126,66 @@
     "./codex/*": {
       "types": "./out/src/codex/*.d.ts",
       "default": "./out/src/codex/*.js"
+    },
+    "./utils/parquet": {
+      "types": "./out/src/utils/parquet.d.ts",
+      "default": "./out/src/utils/parquet.js"
+    },
+    "./tools/postcode-triples": {
+      "types": "./out/src/tools/postcode-triples.d.ts",
+      "default": "./out/src/tools/postcode-triples.js"
+    },
+    "./tools/audit": {
+      "types": "./out/src/tools/audit.d.ts",
+      "default": "./out/src/tools/audit.js"
+    },
+    "./tools/gb-title-case": {
+      "types": "./out/src/tools/gb-title-case.d.ts",
+      "default": "./out/src/tools/gb-title-case.js"
+    },
+    "./tools/golden-relabel-street": {
+      "types": "./out/src/tools/golden-relabel-street.d.ts",
+      "default": "./out/src/tools/golden-relabel-street.js"
+    },
+    "./tools/overture-subvenue": {
+      "types": "./out/src/tools/overture-subvenue.d.ts",
+      "default": "./out/src/tools/overture-subvenue.js"
+    },
+    "./tools/sub-venue-lexicon": {
+      "types": "./out/src/tools/sub-venue-lexicon.d.ts",
+      "default": "./out/src/tools/sub-venue-lexicon.js"
+    },
+    "./tools/sub-venue-promotions": {
+      "types": "./out/src/tools/sub-venue-promotions.d.ts",
+      "default": "./out/src/tools/sub-venue-promotions.js"
+    },
+    "./tools/fetch/download": {
+      "types": "./out/src/tools/fetch/download.d.ts",
+      "default": "./out/src/tools/fetch/download.js"
+    },
+    "./tools/fetch/ppd": {
+      "types": "./out/src/tools/fetch/ppd.d.ts",
+      "default": "./out/src/tools/fetch/ppd.js"
+    },
+    "./utils/tokenize": {
+      "types": "./out/src/utils/tokenize.d.ts",
+      "default": "./out/src/utils/tokenize.js"
+    },
+    "./utils/golden": {
+      "types": "./out/src/utils/golden.d.ts",
+      "default": "./out/src/utils/golden.js"
+    },
+    "./utils/license": {
+      "types": "./out/src/utils/license.d.ts",
+      "default": "./out/src/utils/license.js"
+    },
+    "./utils/split": {
+      "types": "./out/src/utils/split.d.ts",
+      "default": "./out/src/utils/split.js"
+    },
+    "./utils/wof-json": {
+      "types": "./out/src/utils/wof-json.d.ts",
+      "default": "./out/src/utils/wof-json.js"
     }
   },
   "publishConfig": {
@@ -113,6 +221,10 @@
         "types": "./out/src/synthesizers/*.d.ts",
         "default": "./out/src/synthesizers/*.js"
       },
+      "./shard-recipes/*": {
+        "types": "./out/src/shard-recipes/*.d.ts",
+        "default": "./out/src/shard-recipes/*.js"
+      },
       "./adapters/utils": {
         "types": "./out/src/adapters/utils/index.d.ts",
         "default": "./out/src/adapters/utils/index.js"
@@ -132,22 +244,130 @@
       "./codex/*": {
         "types": "./out/src/codex/*.d.ts",
         "default": "./out/src/codex/*.js"
+      },
+      "./utils/parquet": {
+        "types": "./out/src/utils/parquet.d.ts",
+        "default": "./out/src/utils/parquet.js"
+      },
+      "./tools/postcode-triples": {
+        "types": "./out/src/tools/postcode-triples.d.ts",
+        "default": "./out/src/tools/postcode-triples.js"
+      },
+      "./tools/audit": {
+        "types": "./out/src/tools/audit.d.ts",
+        "default": "./out/src/tools/audit.js"
+      },
+      "./tools/gb-title-case": {
+        "types": "./out/src/tools/gb-title-case.d.ts",
+        "default": "./out/src/tools/gb-title-case.js"
+      },
+      "./tools/golden-relabel-street": {
+        "types": "./out/src/tools/golden-relabel-street.d.ts",
+        "default": "./out/src/tools/golden-relabel-street.js"
+      },
+      "./tools/overture-subvenue": {
+        "types": "./out/src/tools/overture-subvenue.d.ts",
+        "default": "./out/src/tools/overture-subvenue.js"
+      },
+      "./tools/sub-venue-lexicon": {
+        "types": "./out/src/tools/sub-venue-lexicon.d.ts",
+        "default": "./out/src/tools/sub-venue-lexicon.js"
+      },
+      "./tools/sub-venue-promotions": {
+        "types": "./out/src/tools/sub-venue-promotions.d.ts",
+        "default": "./out/src/tools/sub-venue-promotions.js"
+      },
+      "./tools/fetch/download": {
+        "types": "./out/src/tools/fetch/download.d.ts",
+        "default": "./out/src/tools/fetch/download.js"
+      },
+      "./tools/fetch/ppd": {
+        "types": "./out/src/tools/fetch/ppd.d.ts",
+        "default": "./out/src/tools/fetch/ppd.js"
+      },
+      "./utils/tokenize": {
+        "types": "./out/src/utils/tokenize.d.ts",
+        "default": "./out/src/utils/tokenize.js"
+      },
+      "./utils/golden": {
+        "types": "./out/src/utils/golden.d.ts",
+        "default": "./out/src/utils/golden.js"
+      },
+      "./utils/license": {
+        "types": "./out/src/utils/license.d.ts",
+        "default": "./out/src/utils/license.js"
+      },
+      "./utils/split": {
+        "types": "./out/src/utils/split.d.ts",
+        "default": "./out/src/utils/split.js"
+      },
+      "./utils/wof-json": {
+        "types": "./out/src/utils/wof-json.d.ts",
+        "default": "./out/src/utils/wof-json.js"
+      }
+    },
+    "imports": {
+      "#adapters/utils": {
+        "types": "./out/src/adapters/utils/index.d.ts",
+        "default": "./out/src/adapters/utils/index.js"
+      },
+      "#runner": {
+        "types": "./out/src/runner.d.ts",
+        "default": "./out/src/runner.js"
+      },
+      "#synthesizers/utils": {
+        "types": "./out/src/synthesizers/utils.d.ts",
+        "default": "./out/src/synthesizers/utils.js"
+      },
+      "#types": {
+        "types": "./out/src/types.d.ts",
+        "default": "./out/src/types.js"
+      },
+      "#utils": {
+        "types": "./out/src/utils/index.d.ts",
+        "default": "./out/src/utils/index.js"
+      },
+      "#utils/*": {
+        "types": "./out/src/utils/*.d.ts",
+        "default": "./out/src/utils/*.js"
+      },
+      "#codex/*": {
+        "types": "./out/src/codex/*.d.ts",
+        "default": "./out/src/codex/*.js"
+      },
+      "#synthesizers/*": {
+        "types": "./out/src/synthesizers/*.d.ts",
+        "default": "./out/src/synthesizers/*.js"
+      },
+      "#tools": {
+        "types": "./out/src/tools/index.d.ts",
+        "default": "./out/src/tools/index.js"
+      },
+      "#adapters/*": {
+        "types": "./out/src/adapters/*.d.ts",
+        "default": "./out/src/adapters/*.js"
+      },
+      "#parquet-wrapper": {
+        "types": "./out/src/parquet-wrapper/index.d.ts",
+        "default": "./out/src/parquet-wrapper/index.js"
       }
     }
   },
   "dependencies": {
     "@dsnp/parquetjs": "1.8.9",
-    "@mailwoman/ban": "9.1.0",
-    "@mailwoman/codex": "9.1.0",
-    "@mailwoman/core": "9.1.0",
-    "@mailwoman/formatter": "9.1.0",
+    "@mailwoman/ban": "9.2.0",
+    "@mailwoman/codex": "9.2.0",
+    "@mailwoman/core": "9.2.0",
+    "@mailwoman/formatter": "9.2.0",
     "fast-glob": "^3.3.3",
     "fastest-levenshtein": "^1.0.16",
     "lru-cache": "^11.5.2",
     "path-ts": "^2.1.0",
-    "spliterator": "^6.1.0"
+    "spliterator": "^6.2.0"
   },
   "devDependencies": {
-    "@types/node": ">=26.1.2"
+    "@types/node": ">=26.1.2",
+    "adm-zip": "^0.6.0",
+    "kysely": "^0.29.4"
   }
 }

package/src/adapters/ban/adapter.ts

@@ -31,11 +31,13 @@
  *   country !== FR).
  */
 
-import { stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
+import { isPresent } from "@mailwoman/core/objects"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { CSVSpliterator } from "spliterator"
 
+import { stableSourceID } from "#adapters/utils"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 import { decomposeFrStreet } from "./street-decompose.ts"
 
 /**
@@ -81,13 +83,13 @@ function composeHouseNumber(numero: string, rep: string): string {
  */
 function composeRaw(house: string, street: string, postcode: string, locality: string): string {
 	const parts: string[] = []
-	const streetPart = [house, street].filter(Boolean).join(" ").trim()
+	const streetPart = [house, street].filter(isPresent).join(" ").trim()
 
 	if (streetPart) {
 		parts.push(streetPart)
 	}
 
-	const cityPart = [postcode, locality].filter(Boolean).join(" ").trim()
+	const cityPart = [postcode, locality].filter(isPresent).join(" ").trim()
 
 	if (cityPart) {
 		parts.push(cityPart)

package/src/adapters/fcc-bdc/adapter.ts

@@ -36,11 +36,12 @@
 
 import { DatabaseSync } from "node:sqlite"
 
-import { splitStreetLine } from "@mailwoman/corpus/adapters/utils"
-import { lookupStateAbbreviation } from "@mailwoman/corpus/codex/us-fips-state"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
 import { formatAddress, reconcileComponents } from "@mailwoman/formatter"
 
+import { splitStreetLine } from "#adapters/utils"
+import { lookupStateAbbreviation } from "#codex/us-fips-state"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.

package/src/adapters/geonames/adapter.ts

@@ -30,12 +30,13 @@
 
 import { existsSync } from "node:fs"
 
-import { stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { dirname, join } from "path-ts"
 import { TSVSpliterator } from "spliterator"
 
+import { stableSourceID } from "#adapters/utils"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.

package/src/adapters/geonames-postal/adapter.ts

@@ -23,11 +23,12 @@
  *        `"CC-BY-4.0"` per row (attribute "GeoNames").
  */
 
-import { stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { TSVSpliterator } from "spliterator"
 
+import { stableSourceID } from "#adapters/utils"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.
@@ -49,7 +50,7 @@ export function createGeonamesPostalAdapter(): CorpusAdapter {
 		id: GEONAMES_POSTAL_ADAPTER_ID,
 		defaultLicense: GEONAMES_POSTAL_DEFAULT_LICENSE,
 		description:
-			"GeoNames postal codes (CC-BY-4.0) — multi-locale postcode→locality→region, names inline; international postcode-first order.",
+			"GeoNames postcodes (CC-BY-4.0) — multi-locale postcode→locality→region, names inline; international postcode-first order.",
 
 		async *rows(opts: AdapterOptions): AsyncIterable<CanonicalRow> {
 			// `header: false` — the GeoNames postal dump is headerless, and the spliterator would

package/src/adapters/gnaf/adapter.ts

@@ -29,11 +29,12 @@
  */
 
 import { tryParsingJSON } from "@mailwoman/core/objects"
-import { stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { TextSpliterator } from "spliterator"
 
+import { stableSourceID } from "#adapters/utils"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.

package/src/adapters/index.ts

@@ -19,8 +19,8 @@
  *   implementations; the registered ids are unchanged).
  */
 
-import { defaultAdapterRegistry } from "@mailwoman/corpus/adapters/utils"
-import type { CorpusAdapter } from "@mailwoman/corpus/types"
+import { defaultAdapterRegistry } from "#adapters/utils"
+import type { CorpusAdapter } from "#types"
 
 import { banAdapter } from "./ban/adapter.ts"
 import { fccBdcAdapter } from "./fcc-bdc/adapter.ts"

package/src/adapters/openaddresses/adapter.ts

@@ -37,12 +37,13 @@
  */
 
 import { tryParsingJSON } from "@mailwoman/core/objects"
-import { stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
-import { SHARE_ALIKE_PATTERN } from "@mailwoman/corpus/utils"
 import { formatAddress, reconcileComponents } from "@mailwoman/formatter"
 import { TextSpliterator } from "spliterator"
 
+import { stableSourceID } from "#adapters/utils"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+import { SHARE_ALIKE_PATTERN } from "#utils"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.

package/src/adapters/overture/adapter.ts

@@ -31,11 +31,12 @@
  */
 
 import { tryParsingJSON } from "@mailwoman/core/objects"
-import { stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
 import { formatAddress, reconcileComponents } from "@mailwoman/formatter"
 import { TextSpliterator } from "spliterator"
 
+import { stableSourceID } from "#adapters/utils"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.

package/src/adapters/state-hi-schools/adapter.ts

@@ -29,12 +29,14 @@
  *   License: stamped `"Public Domain"` per Hawaii state government open-data terms.
  */
 
-import { splitStreetLine, stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import { lookupStateAbbreviation } from "@mailwoman/corpus/codex/us-fips-state"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
+import { isPresent } from "@mailwoman/core/objects"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { CSVSpliterator } from "spliterator"
 
+import { splitStreetLine, stableSourceID } from "#adapters/utils"
+import { lookupStateAbbreviation } from "#codex/us-fips-state"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.
@@ -119,14 +121,14 @@ export function createStateHiSchoolsAdapter(): CorpusAdapter {
 					postcode: zip,
 				}
 
-				const streetPart = [split.house_number, split.street].filter(Boolean).join(" ").trim()
+				const streetPart = [split.house_number, split.street].filter(isPresent).join(" ").trim()
 
 				const raw = [
 					name,
 					streetPart,
-					[city, [HI_STATE_ABBR, zip].filter(Boolean).join(" ")].filter(Boolean).join(", "),
+					[city, [HI_STATE_ABBR, zip].filter(isPresent).join(" ")].filter(isPresent).join(", "),
 				]
-					.filter(Boolean)
+					.filter(isPresent)
 					.join(", ")
 
 				const aligned = reconcileComponents(components, raw)

package/src/adapters/state-ia-contractors/adapter.ts

@@ -16,12 +16,14 @@
  *   License: stamped `"Public Domain"` per Iowa state government open-data terms.
  */
 
-import { splitStreetLine, stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import { lookupStateAbbreviation } from "@mailwoman/corpus/codex/us-fips-state"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
+import { isPresent } from "@mailwoman/core/objects"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { CSVSpliterator } from "spliterator"
 
+import { splitStreetLine, stableSourceID } from "#adapters/utils"
+import { lookupStateAbbreviation } from "#codex/us-fips-state"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.
@@ -83,7 +85,7 @@ export function createStateIaContractorsAdapter(): CorpusAdapter {
 
 				if (!state) continue
 
-				const fullAddress = [address1, address2].filter(Boolean).join(" ")
+				const fullAddress = [address1, address2].filter(isPresent).join(" ")
 				const split = splitStreetLine(fullAddress)
 
 				if (!split) continue
@@ -99,10 +101,14 @@ export function createStateIaContractorsAdapter(): CorpusAdapter {
 					postcode: zip,
 				}
 
-				const streetPart = [split.house_number, split.street].filter(Boolean).join(" ").trim()
+				const streetPart = [split.house_number, split.street].filter(isPresent).join(" ").trim()
 
-				const raw = [venue, streetPart, [city, [stateAbbr, zip].filter(Boolean).join(" ")].filter(Boolean).join(", ")]
-					.filter(Boolean)
+				const raw = [
+					venue,
+					streetPart,
+					[city, [stateAbbr, zip].filter(isPresent).join(" ")].filter(isPresent).join(", "),
+				]
+					.filter(isPresent)
 					.join(", ")
 
 				const aligned = reconcileComponents(components, raw)

package/src/adapters/state-ny-notaries/adapter.ts

@@ -14,12 +14,14 @@
  *   License: stamped `"Public Domain"` per New York state government open-data terms.
  */
 
-import { splitStreetLine, stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import { lookupStateAbbreviation } from "@mailwoman/corpus/codex/us-fips-state"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
+import { isPresent } from "@mailwoman/core/objects"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { CSVSpliterator } from "spliterator"
 
+import { splitStreetLine, stableSourceID } from "#adapters/utils"
+import { lookupStateAbbreviation } from "#codex/us-fips-state"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.
@@ -31,30 +33,6 @@ export const STATE_NY_NOTARIES_ADAPTER_ID = "state-ny-notaries"
  */
 export const STATE_NY_NOTARIES_DEFAULT_LICENSE = "Public Domain"
 
-interface NyNotaryRow {
-	"Commission Holder Name": string
-	"Commission Number (UID)": string
-	"Business Name (if available)": string
-	"Business Address 1 (if available)": string
-	"Business Address 2 (if available)": string
-	" Business City (if available)": string
-	"Business State (if available)": string
-	"Business Zip (if available)": string
-	"Commissioned County": string
-}
-
-const RAW_NY_COLUMNS = [
-	"Commission Holder Name",
-	"Commission Number (UID)",
-	"Business Name (if available)",
-	"Business Address 1 (if available)",
-	"Business Address 2 (if available)",
-	" Business City (if available)",
-	"Business State (if available)",
-	"Business Zip (if available)",
-	"Commissioned County",
-] as const
-
 export function createStateNyNotariesAdapter(): CorpusAdapter {
 	return {
 		id: STATE_NY_NOTARIES_ADAPTER_ID,
@@ -103,7 +81,7 @@ export function createStateNyNotariesAdapter(): CorpusAdapter {
 
 				if (!state) continue
 
-				const fullAddress = [address1, address2].filter(Boolean).join(" ")
+				const fullAddress = [address1, address2].filter(isPresent).join(" ")
 				const split = splitStreetLine(fullAddress)
 
 				if (!split) continue
@@ -120,10 +98,14 @@ export function createStateNyNotariesAdapter(): CorpusAdapter {
 					...(county ? { subregion: county } : {}),
 				}
 
-				const streetPart = [split.house_number, split.street].filter(Boolean).join(" ").trim()
+				const streetPart = [split.house_number, split.street].filter(isPresent).join(" ").trim()
 
-				const raw = [venue, streetPart, [city, [stateAbbr, zip].filter(Boolean).join(" ")].filter(Boolean).join(", ")]
-					.filter(Boolean)
+				const raw = [
+					venue,
+					streetPart,
+					[city, [stateAbbr, zip].filter(isPresent).join(" ")].filter(isPresent).join(", "),
+				]
+					.filter(isPresent)
 					.join(", ")
 
 				const aligned = reconcileComponents(components, raw)

package/src/adapters/state-tx-notaries/adapter.ts

@@ -15,12 +15,14 @@
  *   License: stamped `"Public Domain"` per Texas state government open-data terms.
  */
 
-import { splitStreetLine, stableSourceID } from "@mailwoman/corpus/adapters/utils"
-import { lookupStateAbbreviation } from "@mailwoman/corpus/codex/us-fips-state"
-import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "@mailwoman/corpus/types"
+import { isPresent } from "@mailwoman/core/objects"
 import { reconcileComponents } from "@mailwoman/formatter"
 import { CSVSpliterator } from "spliterator"
 
+import { splitStreetLine, stableSourceID } from "#adapters/utils"
+import { lookupStateAbbreviation } from "#codex/us-fips-state"
+import type { AdapterOptions, CanonicalRow, CorpusAdapter } from "#types"
+
 /**
  * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
  * it came from.
@@ -103,7 +105,7 @@ export function createStateTxNotariesAdapter(): CorpusAdapter {
 
 				if (!split) continue
 
-				const venue = [firstName, lastName].filter(Boolean).join(" ") || undefined
+				const venue = [firstName, lastName].filter(isPresent).join(" ") || undefined
 
 				const components: CanonicalRow["components"] = {
 					...(venue ? { venue } : {}),
@@ -114,10 +116,14 @@ export function createStateTxNotariesAdapter(): CorpusAdapter {
 					...(zip ? { postcode: zip } : {}),
 				}
 
-				const streetPart = [split.house_number, split.street].filter(Boolean).join(" ").trim()
+				const streetPart = [split.house_number, split.street].filter(isPresent).join(" ").trim()
 
-				const raw = [venue, streetPart, [city, [stateAbbr, zip].filter(Boolean).join(" ")].filter(Boolean).join(", ")]
-					.filter(Boolean)
+				const raw = [
+					venue,
+					streetPart,
+					[city, [stateAbbr, zip].filter(isPresent).join(" ")].filter(isPresent).join(", "),
+				]
+					.filter(isPresent)
 					.join(", ")
 
 				const aligned = reconcileComponents(components, raw)