@mailwoman/resolver-wof-sqlite 9.0.0 → 9.2.0
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/README.md +28 -9
- package/address-point-interpolation.ts +18 -8
- package/address-point-schema.ts +18 -6
- package/address-point.ts +111 -18
- package/ancestry.ts +9 -6
- package/build-candidate.ts +287 -157
- package/build-slim.ts +3 -3
- package/candidate/alias-bags.ts +54 -0
- package/candidate/ancestors-sidecar.ts +206 -0
- package/candidate/country-display-names.ts +79 -0
- package/candidate/name-roles.ts +237 -0
- package/candidate/own-name.ts +146 -0
- package/candidate/place-attrs.ts +44 -0
- package/candidate/shard-fold.ts +137 -0
- package/candidate-ancestors-schema.ts +195 -0
- package/candidate-fts.ts +4 -2
- package/candidate-importance.ts +228 -0
- package/candidate-lookup.ts +564 -174
- package/candidate-schema.ts +60 -3
- package/candidate-scoring.ts +268 -0
- package/capital-schema.ts +90 -0
- package/capitals.ts +148 -0
- package/coincident-roles.ts +69 -10
- package/convention-schema.ts +72 -0
- package/convention.ts +2 -2
- package/coverage-manifest-schema.ts +7 -7
- package/currency-backfill.ts +249 -0
- package/exact-match.ts +104 -0
- package/fst-autocomplete.ts +105 -122
- package/fst-builder.ts +39 -47
- package/fst-deserialize-web.ts +43 -7
- package/fst-freshness.ts +2 -2
- package/fst-serialize.ts +68 -12
- package/fst-types.ts +35 -1
- package/fts-query.ts +1 -1
- package/fts.ts +16 -4
- package/geonames-postal.ts +2 -2
- package/index.ts +26 -14
- package/interpolation.ts +113 -19
- package/lookup.ts +118 -560
- package/name-score.ts +6 -4
- package/out/address-point-interpolation.d.ts.map +1 -1
- package/out/address-point-interpolation.js +13 -7
- package/out/address-point-interpolation.js.map +1 -1
- package/out/address-point-schema.d.ts +16 -6
- package/out/address-point-schema.d.ts.map +1 -1
- package/out/address-point-schema.js.map +1 -1
- package/out/address-point.d.ts.map +1 -1
- package/out/address-point.js +70 -14
- package/out/address-point.js.map +1 -1
- package/out/ancestry.d.ts +2 -2
- package/out/ancestry.d.ts.map +1 -1
- package/out/ancestry.js +5 -6
- package/out/ancestry.js.map +1 -1
- package/out/build-candidate.d.ts +108 -0
- package/out/build-candidate.d.ts.map +1 -1
- package/out/build-candidate.js +151 -120
- package/out/build-candidate.js.map +1 -1
- package/out/build-slim.d.ts +1 -1
- package/out/build-slim.js +3 -3
- package/out/build-slim.js.map +1 -1
- package/out/candidate/alias-bags.d.ts +17 -0
- package/out/candidate/alias-bags.d.ts.map +1 -0
- package/out/candidate/alias-bags.js +39 -0
- package/out/candidate/alias-bags.js.map +1 -0
- package/out/candidate/ancestors-sidecar.d.ts +33 -0
- package/out/candidate/ancestors-sidecar.d.ts.map +1 -0
- package/out/candidate/ancestors-sidecar.js +140 -0
- package/out/candidate/ancestors-sidecar.js.map +1 -0
- package/out/candidate/country-display-names.d.ts +35 -0
- package/out/candidate/country-display-names.d.ts.map +1 -0
- package/out/candidate/country-display-names.js +59 -0
- package/out/candidate/country-display-names.js.map +1 -0
- package/out/candidate/name-roles.d.ts +55 -0
- package/out/candidate/name-roles.d.ts.map +1 -0
- package/out/candidate/name-roles.js +165 -0
- package/out/candidate/name-roles.js.map +1 -0
- package/out/candidate/own-name.d.ts +50 -0
- package/out/candidate/own-name.d.ts.map +1 -0
- package/out/candidate/own-name.js +132 -0
- package/out/candidate/own-name.js.map +1 -0
- package/out/candidate/place-attrs.d.ts +43 -0
- package/out/candidate/place-attrs.d.ts.map +1 -0
- package/out/candidate/place-attrs.js +15 -0
- package/out/candidate/place-attrs.js.map +1 -0
- package/out/candidate/shard-fold.d.ts +31 -0
- package/out/candidate/shard-fold.d.ts.map +1 -0
- package/out/candidate/shard-fold.js +104 -0
- package/out/candidate/shard-fold.js.map +1 -0
- package/out/candidate-ancestors-schema.d.ts +150 -0
- package/out/candidate-ancestors-schema.d.ts.map +1 -0
- package/out/candidate-ancestors-schema.js +123 -0
- package/out/candidate-ancestors-schema.js.map +1 -0
- package/out/candidate-fts.d.ts +4 -2
- package/out/candidate-fts.d.ts.map +1 -1
- package/out/candidate-fts.js +4 -2
- package/out/candidate-fts.js.map +1 -1
- package/out/candidate-importance.d.ts +132 -0
- package/out/candidate-importance.d.ts.map +1 -0
- package/out/candidate-importance.js +174 -0
- package/out/candidate-importance.js.map +1 -0
- package/out/candidate-lookup.d.ts +22 -37
- package/out/candidate-lookup.d.ts.map +1 -1
- package/out/candidate-lookup.js +446 -132
- package/out/candidate-lookup.js.map +1 -1
- package/out/candidate-schema.d.ts +52 -4
- package/out/candidate-schema.d.ts.map +1 -1
- package/out/candidate-schema.js +8 -0
- package/out/candidate-schema.js.map +1 -1
- package/out/candidate-scoring.d.ts +34 -0
- package/out/candidate-scoring.d.ts.map +1 -0
- package/out/candidate-scoring.js +200 -0
- package/out/candidate-scoring.js.map +1 -0
- package/out/capital-schema.d.ts +51 -0
- package/out/capital-schema.d.ts.map +1 -0
- package/out/capital-schema.js +63 -0
- package/out/capital-schema.js.map +1 -0
- package/out/capitals.d.ts +69 -0
- package/out/capitals.d.ts.map +1 -0
- package/out/capitals.js +98 -0
- package/out/capitals.js.map +1 -0
- package/out/coincident-roles.d.ts +7 -0
- package/out/coincident-roles.d.ts.map +1 -1
- package/out/coincident-roles.js +42 -8
- package/out/coincident-roles.js.map +1 -1
- package/out/convention-schema.d.ts +51 -0
- package/out/convention-schema.d.ts.map +1 -0
- package/out/convention-schema.js +34 -0
- package/out/convention-schema.js.map +1 -0
- package/out/convention.d.ts +1 -1
- package/out/convention.js +2 -2
- package/out/coverage-manifest-schema.js +3 -7
- package/out/coverage-manifest-schema.js.map +1 -1
- package/out/currency-backfill.d.ts +46 -0
- package/out/currency-backfill.d.ts.map +1 -0
- package/out/currency-backfill.js +180 -0
- package/out/currency-backfill.js.map +1 -0
- package/out/exact-match.d.ts +25 -0
- package/out/exact-match.d.ts.map +1 -0
- package/out/exact-match.js +89 -0
- package/out/exact-match.js.map +1 -0
- package/out/fst-autocomplete.d.ts +24 -14
- package/out/fst-autocomplete.d.ts.map +1 -1
- package/out/fst-autocomplete.js +84 -100
- package/out/fst-autocomplete.js.map +1 -1
- package/out/fst-builder.d.ts.map +1 -1
- package/out/fst-builder.js +32 -40
- package/out/fst-builder.js.map +1 -1
- package/out/fst-deserialize-web.d.ts.map +1 -1
- package/out/fst-deserialize-web.js +36 -7
- package/out/fst-deserialize-web.js.map +1 -1
- package/out/fst-freshness.d.ts +2 -2
- package/out/fst-freshness.js +2 -2
- package/out/fst-serialize.d.ts +14 -4
- package/out/fst-serialize.d.ts.map +1 -1
- package/out/fst-serialize.js +60 -12
- package/out/fst-serialize.js.map +1 -1
- package/out/fst-types.d.ts +35 -1
- package/out/fst-types.d.ts.map +1 -1
- package/out/fts-query.js +1 -1
- package/out/fts-query.js.map +1 -1
- package/out/fts.d.ts +15 -4
- package/out/fts.d.ts.map +1 -1
- package/out/fts.js +15 -4
- package/out/fts.js.map +1 -1
- package/out/geonames-postal.d.ts +2 -2
- package/out/geonames-postal.js +2 -2
- package/out/index.d.ts +4 -2
- package/out/index.d.ts.map +1 -1
- package/out/index.js +3 -2
- package/out/index.js.map +1 -1
- package/out/interpolation.d.ts +8 -0
- package/out/interpolation.d.ts.map +1 -1
- package/out/interpolation.js +91 -19
- package/out/interpolation.js.map +1 -1
- package/out/lookup.d.ts +4 -5
- package/out/lookup.d.ts.map +1 -1
- package/out/lookup.js +102 -444
- package/out/lookup.js.map +1 -1
- package/out/name-score.d.ts +0 -10
- package/out/name-score.d.ts.map +1 -1
- package/out/name-score.js +6 -4
- package/out/name-score.js.map +1 -1
- package/out/place-importance-schema.d.ts +226 -0
- package/out/place-importance-schema.d.ts.map +1 -0
- package/out/place-importance-schema.js +288 -0
- package/out/place-importance-schema.js.map +1 -0
- package/out/poi-lookup.d.ts +1 -1
- package/out/poi-lookup.d.ts.map +1 -1
- package/out/poi-lookup.js +12 -13
- package/out/poi-lookup.js.map +1 -1
- package/out/poi-schema.d.ts +7 -3
- package/out/poi-schema.d.ts.map +1 -1
- package/out/poi-schema.js.map +1 -1
- package/out/polygon-schema.d.ts +37 -0
- package/out/polygon-schema.d.ts.map +1 -0
- package/out/polygon-schema.js +23 -0
- package/out/polygon-schema.js.map +1 -0
- package/out/postal-city-alias-lookup.d.ts +1 -1
- package/out/postal-city-alias-lookup.js +1 -1
- package/out/postal-city-candidate-schema.d.ts +2 -1
- package/out/postal-city-candidate-schema.d.ts.map +1 -1
- package/out/postal-city-candidate-schema.js.map +1 -1
- package/out/postcode-point-lookup.d.ts +1 -1
- package/out/postcode-point-lookup.js +1 -1
- package/out/primary-preference.d.ts +125 -0
- package/out/primary-preference.d.ts.map +1 -0
- package/out/primary-preference.js +138 -0
- package/out/primary-preference.js.map +1 -0
- package/out/proximity-rerank.d.ts +77 -0
- package/out/proximity-rerank.d.ts.map +1 -0
- package/out/proximity-rerank.js +86 -0
- package/out/proximity-rerank.js.map +1 -0
- package/out/region-keys.d.ts +47 -0
- package/out/region-keys.d.ts.map +1 -0
- package/out/region-keys.js +121 -0
- package/out/region-keys.js.map +1 -0
- package/out/reverse.d.ts.map +1 -1
- package/out/reverse.js +6 -9
- package/out/reverse.js.map +1 -1
- package/out/schema.d.ts +1 -1
- package/out/search-fetch.d.ts +57 -0
- package/out/search-fetch.d.ts.map +1 -0
- package/out/search-fetch.js +183 -0
- package/out/search-fetch.js.map +1 -0
- package/out/sharding.d.ts +3 -3
- package/out/sharding.js +1 -1
- package/out/sqlite-convention-source.d.ts +1 -1
- package/out/sqlite-convention-source.js +1 -1
- package/out/sqlite-utils.d.ts +31 -1
- package/out/sqlite-utils.d.ts.map +1 -1
- package/out/sqlite-utils.js +38 -0
- package/out/sqlite-utils.js.map +1 -1
- package/out/street-centroid-schema.d.ts +7 -2
- package/out/street-centroid-schema.d.ts.map +1 -1
- package/out/street-centroid-schema.js.map +1 -1
- package/out/street-centroid.d.ts.map +1 -1
- package/out/street-centroid.js +7 -7
- package/out/street-centroid.js.map +1 -1
- package/out/street-morphology-fst-builder.d.ts.map +1 -1
- package/out/street-morphology-fst-builder.js +5 -4
- package/out/street-morphology-fst-builder.js.map +1 -1
- package/out/street-normalize.d.ts +83 -9
- package/out/street-normalize.d.ts.map +1 -1
- package/out/street-normalize.js +177 -10
- package/out/street-normalize.js.map +1 -1
- package/out/street-segment-schema.d.ts +6 -2
- package/out/street-segment-schema.d.ts.map +1 -1
- package/out/street-segment-schema.js.map +1 -1
- package/out/types.d.ts +74 -1
- package/out/types.d.ts.map +1 -1
- package/out/unified-schema.d.ts +1 -1
- package/out/unified-schema.js +1 -1
- package/out/uprn-lookup.d.ts +85 -0
- package/out/uprn-lookup.d.ts.map +1 -0
- package/out/uprn-lookup.js +152 -0
- package/out/uprn-lookup.js.map +1 -0
- package/out/uprn-schema.d.ts +93 -0
- package/out/uprn-schema.d.ts.map +1 -0
- package/out/uprn-schema.js +78 -0
- package/out/uprn-schema.js.map +1 -0
- package/out/weights-overlay-linker.d.ts +141 -0
- package/out/weights-overlay-linker.d.ts.map +1 -0
- package/out/weights-overlay-linker.js +259 -0
- package/out/weights-overlay-linker.js.map +1 -0
- package/package.json +296 -16
- package/place-importance-schema.ts +402 -0
- package/poi-lookup.ts +12 -13
- package/poi-schema.ts +8 -3
- package/polygon-schema.ts +47 -0
- package/postal-city-alias-lookup.ts +1 -1
- package/postal-city-candidate-schema.ts +3 -1
- package/postcode-point-lookup.ts +1 -1
- package/primary-preference.ts +207 -0
- package/proximity-rerank.ts +120 -0
- package/region-keys.ts +144 -0
- package/reverse.ts +17 -16
- package/schema.ts +1 -1
- package/search-fetch.ts +256 -0
- package/sharding.ts +3 -3
- package/sqlite-convention-source.ts +1 -1
- package/sqlite-utils.ts +63 -1
- package/street-centroid-schema.ts +8 -2
- package/street-centroid.ts +13 -8
- package/street-morphology-fst-builder.ts +5 -4
- package/street-normalize.ts +254 -24
- package/street-segment-schema.ts +7 -2
- package/types.ts +74 -1
- package/unified-schema.ts +1 -1
- package/uprn-lookup.ts +210 -0
- package/uprn-schema.ts +124 -0
- package/weights-overlay-linker.ts +377 -0
- package/geo.ts +0 -121
- package/out/geo.d.ts +0 -74
- package/out/geo.d.ts.map +0 -1
- package/out/geo.js +0 -71
- package/out/geo.js.map +0 -1
package/out/sqlite-utils.js
CHANGED
|
@@ -5,6 +5,24 @@
|
|
|
5
5
|
*
|
|
6
6
|
* Small shared helpers for the SQLite-backed lookups.
|
|
7
7
|
*/
|
|
8
|
+
import { allRows, getRow } from "@mailwoman/core/utils";
|
|
9
|
+
// The row-shape assertion itself lives in `core` so the readers that cannot depend on this package reach the same
|
|
10
|
+
// seam; re-exported here because this module is where this package's readers already look for it.
|
|
11
|
+
export { allRows, getRow } from "@mailwoman/core/utils";
|
|
12
|
+
/**
|
|
13
|
+
* Prepare a single-row query while preserving its exact parameter tuple at every call site.
|
|
14
|
+
*/
|
|
15
|
+
export function prepareGet(db, sql) {
|
|
16
|
+
const statement = db.prepare(sql);
|
|
17
|
+
return (...parameters) => getRow(statement, ...parameters);
|
|
18
|
+
}
|
|
19
|
+
/**
|
|
20
|
+
* Prepare a multi-row query while preserving its exact parameter tuple at every call site.
|
|
21
|
+
*/
|
|
22
|
+
export function prepareAll(db, sql) {
|
|
23
|
+
const statement = db.prepare(sql);
|
|
24
|
+
return (...parameters) => allRows(statement, ...parameters);
|
|
25
|
+
}
|
|
8
26
|
/**
|
|
9
27
|
* True when `name` is a table in the open database. The street-level lookups use this to degrade gracefully on an
|
|
10
28
|
* empty/tableless shard — an interrupted `build-*-shard.ts`, or a stray 0-byte file (e.g. `sqlite3 <missing>.db "…"`
|
|
@@ -20,4 +38,24 @@ export function hasTable(db, name) {
|
|
|
20
38
|
return false;
|
|
21
39
|
}
|
|
22
40
|
}
|
|
41
|
+
/**
|
|
42
|
+
* True when `table` exists in the open database AND carries `column`.
|
|
43
|
+
*
|
|
44
|
+
* The column-level sibling of {@link hasTable}, and it exists for the same reason one layer down: an artifact built
|
|
45
|
+
* before a column was added is still a VALID artifact, and a reader that unconditionally names the new column in its
|
|
46
|
+
* `SELECT` turns "this gazetteer is a build behind" into `no such column` at the first keystroke. Probe once at
|
|
47
|
+
* construction and shape the query — `table_info` is a PRAGMA, so it must not sit on a per-query path.
|
|
48
|
+
*
|
|
49
|
+
* Note the interpolation: PRAGMA does not take bound parameters, so `table` is spliced. Every caller passes a
|
|
50
|
+
* module-level constant; never pass user input.
|
|
51
|
+
*/
|
|
52
|
+
export function hasColumn(db, table, column) {
|
|
53
|
+
try {
|
|
54
|
+
const rows = allRows(db.prepare(`PRAGMA table_info(${table})`));
|
|
55
|
+
return rows.some((r) => String(r.name) === column);
|
|
56
|
+
}
|
|
57
|
+
catch {
|
|
58
|
+
return false;
|
|
59
|
+
}
|
|
60
|
+
}
|
|
23
61
|
//# sourceMappingURL=sqlite-utils.js.map
|
package/out/sqlite-utils.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"sqlite-utils.js","sourceRoot":"","sources":["../sqlite-utils.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH;;;;;GAKG;AACH,MAAM,UAAU,QAAQ,CAAC,EAAgB,EAAE,IAAY;IACtD,IAAI,CAAC;QACJ,MAAM,GAAG,GAAG,EAAE,CAAC,OAAO,CAAC,uEAAuE,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAEzG,OAAO,GAAG,KAAK,SAAS,CAAA;IACzB,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAA;IACb,CAAC;AACF,CAAC"}
|
|
1
|
+
{"version":3,"file":"sqlite-utils.js","sourceRoot":"","sources":["../sqlite-utils.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,uBAAuB,CAAA;AAEvD,kHAAkH;AAClH,kGAAkG;AAClG,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,uBAAuB,CAAA;AAQvD;;GAEG;AACH,MAAM,UAAU,UAAU,CACzB,EAAgB,EAChB,GAAW;IAEX,MAAM,SAAS,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;IAEjC,OAAO,CAAC,GAAG,UAAU,EAAE,EAAE,CAAC,MAAM,CAAM,SAAS,EAAE,GAAG,UAAU,CAAC,CAAA;AAChE,CAAC;AAOD;;GAEG;AACH,MAAM,UAAU,UAAU,CACzB,EAAgB,EAChB,GAAW;IAEX,MAAM,SAAS,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;IAEjC,OAAO,CAAC,GAAG,UAAU,EAAE,EAAE,CAAC,OAAO,CAAM,SAAS,EAAE,GAAG,UAAU,CAAC,CAAA;AACjE,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,QAAQ,CAAC,EAAgB,EAAE,IAAY;IACtD,IAAI,CAAC;QACJ,MAAM,GAAG,GAAG,EAAE,CAAC,OAAO,CAAC,uEAAuE,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAEzG,OAAO,GAAG,KAAK,SAAS,CAAA;IACzB,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAA;IACb,CAAC;AACF,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,UAAU,SAAS,CAAC,EAAgB,EAAE,KAAa,EAAE,MAAc;IACxE,IAAI,CAAC;QACJ,MAAM,IAAI,GAAG,OAAO,CAAmB,EAAE,CAAC,OAAO,CAAC,qBAAqB,KAAK,GAAG,CAAC,CAAC,CAAA;QAEjF,OAAO,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,CAAA;IACnD,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAA;IACb,CAAC;AACF,CAAC"}
|
|
@@ -20,6 +20,7 @@
|
|
|
20
20
|
* every postcode/arrondissement it spans.
|
|
21
21
|
*/
|
|
22
22
|
import type { Kysely } from "kysely";
|
|
23
|
+
import type { NameKey, StreetKey } from "./street-normalize.ts";
|
|
23
24
|
/**
|
|
24
25
|
* One street roll-up. `(street_norm, postcode, locality_base)` is unique. `lat`/`lon` are the UNWEIGHTED mean of the
|
|
25
26
|
* group's member address points (each source row = one point), so a cross-group weighted mean (`SUM(lat*point_count) /
|
|
@@ -30,7 +31,7 @@ export interface StreetCentroidTable {
|
|
|
30
31
|
/**
|
|
31
32
|
* Shared `normalizeStreetForKeyLocale` of the street — the build/query-consistent probe key.
|
|
32
33
|
*/
|
|
33
|
-
street_norm:
|
|
34
|
+
street_norm: StreetKey;
|
|
34
35
|
/**
|
|
35
36
|
* The 5-digit postcode of this group, or null when the source row carried none.
|
|
36
37
|
*/
|
|
@@ -38,7 +39,7 @@ export interface StreetCentroidTable {
|
|
|
38
39
|
/**
|
|
39
40
|
* Arrondissement-stripped commune (`stripArrondissement(normalizeLocalityForKey(commune))`) — the fallback scope.
|
|
40
41
|
*/
|
|
41
|
-
locality_base:
|
|
42
|
+
locality_base: NameKey;
|
|
42
43
|
/**
|
|
43
44
|
* Weighted-mean centroid latitude of the street's member points.
|
|
44
45
|
*/
|
|
@@ -73,6 +74,10 @@ export interface StreetCentroidTable {
|
|
|
73
74
|
* name-evidence rerank folds the model's street surface with the SAME `foldStreetSurface` used to build this column
|
|
74
75
|
* (the fold-parity contract), so it must not drift from `street_norm`'s richer normalizer. Indexed (`idx_sc_name`)
|
|
75
76
|
* for a direct seek.
|
|
77
|
+
*
|
|
78
|
+
* Deliberately UNBRANDED despite the `name_key` column name: `foldStreetSurface` is a different fold from the
|
|
79
|
+
* {@link NameKey} one every other `name_key` column carries, and giving it that brand would invite exactly the
|
|
80
|
+
* cross-fold probe the brands exist to stop.
|
|
76
81
|
*/
|
|
77
82
|
name_key: string;
|
|
78
83
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-centroid-schema.d.ts","sourceRoot":"","sources":["../street-centroid-schema.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAA;AAEpC;;;;;GAKG;AACH,MAAM,WAAW,mBAAmB;IACnC;;OAEG;IACH,WAAW,EAAE,
|
|
1
|
+
{"version":3,"file":"street-centroid-schema.d.ts","sourceRoot":"","sources":["../street-centroid-schema.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAA;AAEpC,OAAO,KAAK,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,uBAAuB,CAAA;AAE/D;;;;;GAKG;AACH,MAAM,WAAW,mBAAmB;IACnC;;OAEG;IACH,WAAW,EAAE,SAAS,CAAA;IACtB;;OAEG;IACH,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAA;IACvB;;OAEG;IACH,aAAa,EAAE,OAAO,CAAA;IACtB;;OAEG;IACH,GAAG,EAAE,MAAM,CAAA;IACX;;OAEG;IACH,GAAG,EAAE,MAAM,CAAA;IACX,OAAO,EAAE,MAAM,CAAA;IACf,OAAO,EAAE,MAAM,CAAA;IACf,OAAO,EAAE,MAAM,CAAA;IACf,OAAO,EAAE,MAAM,CAAA;IACf;;OAEG;IACH,WAAW,EAAE,MAAM,CAAA;IACnB;;OAEG;IACH,UAAU,EAAE,MAAM,CAAA;IAClB;;OAEG;IACH,MAAM,EAAE,MAAM,CAAA;IACd;;OAEG;IACH,OAAO,EAAE,MAAM,CAAA;IACf;;;;;;;;;;OAUG;IACH,QAAQ,EAAE,MAAM,CAAA;CAChB;AAED;;GAEG;AACH,MAAM,WAAW,sBAAsB;IACtC,eAAe,EAAE,mBAAmB,CAAA;CACpC;AAED;;;GAGG;AACH,eAAO,MAAM,uBAAuB,+KAe1B,CAAA;AAEV;;GAEG;AACH,wBAAsB,yBAAyB,CAAC,EAAE,EAAE,MAAM,CAAC,sBAAsB,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAkBjG;AAED;;;;GAIG;AACH,wBAAsB,2BAA2B,CAAC,EAAE,EAAE,MAAM,CAAC,sBAAsB,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAUnG"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-centroid-schema.js","sourceRoot":"","sources":["../street-centroid-schema.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;
|
|
1
|
+
{"version":3,"file":"street-centroid-schema.js","sourceRoot":"","sources":["../street-centroid-schema.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AA0EH;;;GAGG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG;IACtC,aAAa;IACb,UAAU;IACV,eAAe;IACf,KAAK;IACL,KAAK;IACL,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,aAAa;IACb,YAAY;IACZ,QAAQ;IACR,SAAS;IACT,UAAU;CACD,CAAA;AAEV;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,yBAAyB,CAAC,EAAkC;IACjF,MAAM,EAAE,CAAC,MAAM;SACb,WAAW,CAAC,iBAAiB,CAAC;SAC9B,SAAS,CAAC,aAAa,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SACpD,SAAS,CAAC,UAAU,EAAE,MAAM,CAAC;SAC7B,SAAS,CAAC,eAAe,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SACtD,SAAS,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAC5C,SAAS,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAC5C,SAAS,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAChD,SAAS,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAChD,SAAS,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAChD,SAAS,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAChD,SAAS,CAAC,aAAa,EAAE,SAAS,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SACvD,SAAS,CAAC,YAAY,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SACnD,SAAS,CAAC,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAC/C,SAAS,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SAChD,SAAS,CAAC,UAAU,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;SACjD,OAAO,EAAE,CAAA;AACZ,CAAC;AAED;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,2BAA2B,CAAC,EAAkC;IACnF,MAAM,EAAE,CAAC,MAAM,CAAC,WAAW,CAAC,iBAAiB,CAAC,CAAC,EAAE,CAAC,iBAAiB,CAAC,CAAC,OAAO,CAAC,CAAC,UAAU,EAAE,aAAa,CAAC,CAAC,CAAC,OAAO,EAAE,CAAA;IAEnH,MAAM,EAAE,CAAC,MAAM;SACb,WAAW,CAAC,iBAAiB,CAAC;SAC9B,EAAE,CAAC,iBAAiB,CAAC;SACrB,OAAO,CAAC,CAAC,eAAe,EAAE,aAAa,CAAC,CAAC;SACzC,OAAO,EAAE,CAAA;IAEX,MAAM,EAAE,CAAC,MAAM,CAAC,WAAW,CAAC,aAAa,CAAC,CAAC,EAAE,CAAC,iBAAiB,CAAC,CAAC,OAAO,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,EAAE,CAAA;AACjG,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-centroid.d.ts","sourceRoot":"","sources":["../street-centroid.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAIH,OAAO,KAAK,EAAE,iBAAiB,EAAE,oBAAoB,EAAE,MAAM,qBAAqB,CAAA;AAGlF,OAAO,
|
|
1
|
+
{"version":3,"file":"street-centroid.d.ts","sourceRoot":"","sources":["../street-centroid.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAIH,OAAO,KAAK,EAAE,iBAAiB,EAAE,oBAAoB,EAAE,MAAM,qBAAqB,CAAA;AAGlF,OAAO,EAMN,KAAK,YAAY,EAEjB,MAAM,uBAAuB,CAAA;AAwC9B,qBAAa,0BAA2B,YAAW,oBAAoB;;IAMtE;;;;OAIG;gBACS,MAAM,EAAE,MAAM,EAAE,IAAI,GAAE;QAAE,YAAY,CAAC,EAAE,YAAY,CAAA;KAAO;IAmBtE,IAAI,CAAC,KAAK,EAAE;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;QAAC,QAAQ,CAAC,EAAE,MAAM,CAAA;KAAE,GAAG,iBAAiB,GAAG,IAAI;IA4B/F,KAAK,IAAI,IAAI;CAGb"}
|
package/out/street-centroid.js
CHANGED
|
@@ -19,8 +19,8 @@
|
|
|
19
19
|
* exact-after-normalization only (no fuzzy street matching in this tier).
|
|
20
20
|
*/
|
|
21
21
|
import { DatabaseSync } from "node:sqlite";
|
|
22
|
-
import { hasTable } from "./sqlite-utils.js";
|
|
23
|
-
import { normalizeLocalityForKey, normalizeStreetForKeyLocale, stripArrondissement, } from "./street-normalize.js";
|
|
22
|
+
import { hasTable, prepareGet } from "./sqlite-utils.js";
|
|
23
|
+
import { normalizeLocalityForKey, normalizeStreetForKeyLocale, streetLocaleForSurface, stripArrondissement, } from "./street-normalize.js";
|
|
24
24
|
/**
|
|
25
25
|
* Weighted-centroid aggregate over a WHERE-filtered set. `SUM(coord*n)/SUM(n)` reconstructs the grand centroid.
|
|
26
26
|
*/
|
|
@@ -57,23 +57,23 @@ export class StreetCentroidSqliteLookup {
|
|
|
57
57
|
// Degrade gracefully on an empty/tableless shard (interrupted build, stray 0-byte file): with no
|
|
58
58
|
// `street_centroid` table this lookup is a no-op miss, not a crash (mirrors the address-point reader).
|
|
59
59
|
if (hasTable(this.#db, "street_centroid")) {
|
|
60
|
-
this.#byPostcode = this.#db
|
|
61
|
-
this.#byLocality = this.#db
|
|
60
|
+
this.#byPostcode = prepareGet(this.#db, `SELECT ${AGG_SELECT} FROM street_centroid WHERE postcode = ? AND street_norm = ?`);
|
|
61
|
+
this.#byLocality = prepareGet(this.#db, `SELECT ${AGG_SELECT} FROM street_centroid WHERE locality_base = ? AND street_norm = ?`);
|
|
62
62
|
}
|
|
63
63
|
}
|
|
64
64
|
find(query) {
|
|
65
65
|
if (!this.#byPostcode || !this.#byLocality)
|
|
66
66
|
return null;
|
|
67
|
-
const streetNorm = normalizeStreetForKeyLocale(query.street, this.#locale);
|
|
67
|
+
const streetNorm = normalizeStreetForKeyLocale(query.street, streetLocaleForSurface(query.street, this.#locale));
|
|
68
68
|
if (!streetNorm)
|
|
69
69
|
return null;
|
|
70
70
|
let row;
|
|
71
71
|
if (query.postcode?.trim()) {
|
|
72
|
-
row = this.#byPostcode
|
|
72
|
+
row = this.#byPostcode(query.postcode.trim(), streetNorm);
|
|
73
73
|
}
|
|
74
74
|
if ((!row || row.lat == null) && query.locality?.trim()) {
|
|
75
75
|
const base = stripArrondissement(normalizeLocalityForKey(query.locality));
|
|
76
|
-
row = this.#byLocality
|
|
76
|
+
row = this.#byLocality(base, streetNorm);
|
|
77
77
|
}
|
|
78
78
|
if (!row || row.lat == null || row.lon == null)
|
|
79
79
|
return null;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-centroid.js","sourceRoot":"","sources":["../street-centroid.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAI1C,OAAO,EAAE,QAAQ,EAAE,MAAM,mBAAmB,CAAA;
|
|
1
|
+
{"version":3,"file":"street-centroid.js","sourceRoot":"","sources":["../street-centroid.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAI1C,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAoB,MAAM,mBAAmB,CAAA;AAC1E,OAAO,EACN,uBAAuB,EACvB,2BAA2B,EAG3B,sBAAsB,EAEtB,mBAAmB,GACnB,MAAM,uBAAuB,CAAA;AAgB9B;;GAEG;AACH,MAAM,UAAU,GACf,oDAAoD;IACpD,oDAAoD;IACpD,sGAAsG;IACtG,gDAAgD,CAAA;AAEjD;;GAEG;AACH,SAAS,aAAa,CAAC,MAAc,EAAE,MAAc,EAAE,MAAc,EAAE,MAAc;IACpF,MAAM,CAAC,GAAG,SAAS,CAAA;IACnB,MAAM,KAAK,GAAG,CAAC,CAAS,EAAU,EAAE,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,GAAG,GAAG,CAAA;IACxD,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,GAAG,MAAM,CAAC,CAAA;IACnC,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,GAAG,MAAM,CAAC,CAAA;IACnC,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,MAAM,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAA;IAC3C,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,CAAA;IACnF,MAAM,IAAI,GAAG,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAEzD,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,CAAC,CAAA;AAC5B,CAAC;AAED,MAAM,OAAO,0BAA0B;IAC7B,GAAG,CAAc;IACjB,OAAO,CAAc;IACrB,WAAW,CAAwE;IACnF,WAAW,CAAyE;IAE7F;;;;OAIG;IACH,YAAY,MAAc,EAAE,OAAwC,EAAE;QACrE,IAAI,CAAC,GAAG,GAAG,IAAI,YAAY,CAAC,MAAM,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAA;QACvD,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,YAAY,IAAI,IAAI,CAAA;QAExC,iGAAiG;QACjG,uGAAuG;QACvG,IAAI,QAAQ,CAAC,IAAI,CAAC,GAAG,EAAE,iBAAiB,CAAC,EAAE,CAAC;YAC3C,IAAI,CAAC,WAAW,GAAG,UAAU,CAC5B,IAAI,CAAC,GAAG,EACR,UAAU,UAAU,8DAA8D,CAClF,CAAA;YAED,IAAI,CAAC,WAAW,GAAG,UAAU,CAC5B,IAAI,CAAC,GAAG,EACR,UAAU,UAAU,mEAAmE,CACvF,CAAA;QACF,CAAC;IACF,CAAC;IAED,IAAI,CAAC,KAA+D;QACnE,IAAI,CAAC,IAAI,CAAC,WAAW,IAAI,CAAC,IAAI,CAAC,WAAW;YAAE,OAAO,IAAI,CAAA;QACvD,MAAM,UAAU,GAAG,2BAA2B,CAAC,KAAK,CAAC,MAAM,EAAE,sBAAsB,CAAC,KAAK,CAAC,MAAM,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC,CAAA;QAEhH,IAAI,CAAC,UAAU;YAAE,OAAO,IAAI,CAAA;QAE5B,IAAI,GAAuB,CAAA;QAE3B,IAAI,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,EAAE,CAAC;YAC5B,GAAG,GAAG,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,EAAE,UAAU,CAAC,CAAA;QAC1D,CAAC;QAED,IAAI,CAAC,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,EAAE,CAAC;YACzD,MAAM,IAAI,GAAG,mBAAmB,CAAC,uBAAuB,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAA;YACzE,GAAG,GAAG,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,UAAU,CAAC,CAAA;QACzC,CAAC;QAED,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,IAAI,IAAI,IAAI,GAAG,CAAC,GAAG,IAAI,IAAI;YAAE,OAAO,IAAI,CAAA;QAE3D,OAAO;YACN,GAAG,EAAE,GAAG,CAAC,GAAG;YACZ,GAAG,EAAE,GAAG,CAAC,GAAG;YACZ,YAAY,EAAE,aAAa,CAAC,GAAG,CAAC,OAAQ,EAAE,GAAG,CAAC,OAAQ,EAAE,GAAG,CAAC,OAAQ,EAAE,GAAG,CAAC,OAAQ,CAAC;YACnF,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,EAAE;YACxB,OAAO,EAAE,GAAG,CAAC,OAAO,IAAI,EAAE;SAC1B,CAAA;IACF,CAAC;IAED,KAAK;QACJ,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,CAAA;IACjB,CAAC;CACD"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-morphology-fst-builder.d.ts","sourceRoot":"","sources":["../street-morphology-fst-builder.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AAQH,OAAO,EAAE,UAAU,EAAmB,MAAM,kBAAkB,CAAA;AAC9D,OAAO,KAAK,EAAE,aAAa,EAAc,MAAM,gBAAgB,CAAA;AAW/D,MAAM,WAAW,4BAA4B;IAC5C;;OAEG;IACH,eAAe,EAAE,MAAM,CAAA;IACvB;;OAEG;IACH,OAAO,CAAC,EAAE,MAAM,EAAE,CAAA;IAClB;;;;;;;;OAQG;IACH,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB;;OAEG;IACH,UAAU,CAAC,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,KAAK,IAAI,CAAA;CACrD;AAED,MAAM,WAAW,8BAA8B;IAC9C,OAAO,EAAE,UAAU,CAAA;IACnB,UAAU,EAAE,aAAa,CAAA;IACzB,cAAc,EAAE,MAAM,CAAA;IACtB,YAAY,EAAE,MAAM,CAAA;IACpB,WAAW,EAAE,MAAM,CAAA;IACnB,OAAO,EAAE,MAAM,EAAE,CAAA;CACjB;AAuBD,wBAAgB,wBAAwB,CAAC,IAAI,EAAE,4BAA4B,GAAG,8BAA8B,
|
|
1
|
+
{"version":3,"file":"street-morphology-fst-builder.d.ts","sourceRoot":"","sources":["../street-morphology-fst-builder.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AAQH,OAAO,EAAE,UAAU,EAAmB,MAAM,kBAAkB,CAAA;AAC9D,OAAO,KAAK,EAAE,aAAa,EAAc,MAAM,gBAAgB,CAAA;AAW/D,MAAM,WAAW,4BAA4B;IAC5C;;OAEG;IACH,eAAe,EAAE,MAAM,CAAA;IACvB;;OAEG;IACH,OAAO,CAAC,EAAE,MAAM,EAAE,CAAA;IAClB;;;;;;;;OAQG;IACH,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB;;OAEG;IACH,UAAU,CAAC,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,KAAK,IAAI,CAAA;CACrD;AAED,MAAM,WAAW,8BAA8B;IAC9C,OAAO,EAAE,UAAU,CAAA;IACnB,UAAU,EAAE,aAAa,CAAA;IACzB,cAAc,EAAE,MAAM,CAAA;IACtB,YAAY,EAAE,MAAM,CAAA;IACpB,WAAW,EAAE,MAAM,CAAA;IACnB,OAAO,EAAE,MAAM,EAAE,CAAA;CACjB;AAuBD,wBAAgB,wBAAwB,CAAC,IAAI,EAAE,4BAA4B,GAAG,8BAA8B,CAsJ3G"}
|
|
@@ -131,10 +131,11 @@ export function buildStreetMorphologyFST(opts) {
|
|
|
131
131
|
placetype: "street_affix",
|
|
132
132
|
name: canonical,
|
|
133
133
|
parentChain: [],
|
|
134
|
-
// Fixed
|
|
135
|
-
// anything but street-typing). The morphology prior caps bias separately; this value
|
|
136
|
-
// just feeds the cap formula `
|
|
137
|
-
|
|
134
|
+
// Fixed referential score: street affixes are structurally unambiguous (Avenue is almost
|
|
135
|
+
// never anything but street-typing). The morphology prior caps bias separately; this value
|
|
136
|
+
// just feeds the cap formula `referential * cap`. No encyclopedic field — a street affix is
|
|
137
|
+
// not a place and has no article; absence here is the correct statement.
|
|
138
|
+
referential: 1,
|
|
138
139
|
lat: 0,
|
|
139
140
|
lon: 0,
|
|
140
141
|
};
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-morphology-fst-builder.js","sourceRoot":"","sources":["../street-morphology-fst-builder.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AAEH,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AAC7D,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAA;AAG7C,OAAO,EAAE,UAAU,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAA;AAG9D;;;;GAIG;AACH,MAAM,uBAAuB,GAAG,aAAa,CAAA;AAE7C,MAAM,qBAAqB,GAAG,kBAAkB,CAAA;AAoChD;;;;;GAKG;AACH,SAAS,SAAS,CAAC,IAAY;IAC9B,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAA;IAE3B,IAAI,CAAC,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAA;IAE3D,MAAM,KAAK,GAAG,OAAO;SACnB,KAAK,CAAC,GAAG,CAAC;SACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IAE7B,IAAI,CAAC,KAAK,CAAC,MAAM;QAAE,OAAO,IAAI,CAAA;IAE9B,OAAO,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAA;AACjD,CAAC;AAED,MAAM,UAAU,wBAAwB,CAAC,IAAkC;IAC1E,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,IAAI,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;IAC9C,MAAM,gBAAgB,GAAG,IAAI,CAAC,gBAAgB,IAAI,CAAC,CAAA;IAEnD,iGAAiG;IACjG,IAAI,OAAiB,CAAA;IAErB,oGAAoG;IACpG,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC;QACzC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAA;IACvB,CAAC;SAAM,CAAC;QACP,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE;YAC5D,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,KAAK,CAAC,CAAA;YAEpD,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,WAAW,EAAE;gBAAE,OAAO,KAAK,CAAA;YAErD,IAAI,CAAC;gBACJ,QAAQ,CAAC,IAAI,CAAC,UAAU,EAAE,qBAAqB,CAAC,CAAC,CAAA;gBAEjD,OAAO,IAAI,CAAA;YACZ,CAAC;YAAC,MAAM,CAAC;gBACR,OAAO,KAAK,CAAA;YACb,CAAC;QACF,CAAC,CAAC,CAAA;IACH,CAAC;IAED,QAAQ,CAAC,UAAU,EAAE,SAAS,OAAO,CAAC,MAAM,iBAAiB,qBAAqB,EAAE,CAAC,CAAA;IAErF,4FAA4F;IAC5F,wEAAwE;IACxE,MAAM,mBAAmB,GAAG,IAAI,GAAG,EAAuB,CAAA;IAE1D,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,MAAM,EAAE,qBAAqB,CAAC,CAAA;QAC1E,MAAM,OAAO,GAAG,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;QAE9C,KAAK,MAAM,IAAI,IAAI,eAAe,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;YAClD,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,CAAA;YAE9B,IAAI,CAAC,MAAM;gBAAE,SAAQ;YACrB,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,IAAI,IAAI,GAAG,EAAU,CAAA;YAE/E,KAAK,MAAM,OAAO,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;gBACvC,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;YACtB,CAAC;YAED,mBAAmB,CAAC,GAAG,CAAC,MAAM,CAAC,SAAS,EAAE,QAAQ,CAAC,CAAA;QACpD,CAAC;IACF,CAAC;IAED,QAAQ,CAAC,SAAS,EAAE,aAAa,mBAAmB,CAAC,IAAI,oBAAoB,CAAC,CAAA;IAE9E,mEAAmE;IACnE,MAAM,gBAAgB,GAAG,CAAC,GAAG,mBAAmB,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAA;IACnE,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAkB,CAAA;IAElD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,gBAAgB,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClD,gBAAgB,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC,CAAE,EAAE,uBAAuB,GAAG,CAAC,CAAC,CAAA;IACxE,CAAC;IAED,2FAA2F;IAC3F,8FAA8F;IAC9F,MAAM,KAAK,GAAc,CAAC,EAAE,KAAK,EAAE,IAAI,GAAG,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,CAAA;IAE3D,SAAS,UAAU,CAAC,MAAgB,EAAE,KAAiB;QACtD,IAAI,CAAC,MAAM,CAAC,MAAM;YAAE,OAAM;QAC1B,IAAI,OAAO,GAAG,CAAC,CAAA;QAEf,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;YACxB,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAE,CAAA;YAC5B,IAAI,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAA;YAE5B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;gBACxB,IAAI,GAAG,KAAK,CAAC,MAAM,CAAA;gBACnB,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,GAAG,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,CAAA;gBAC5C,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;YACxB,CAAC;YAED,OAAO,GAAG,IAAI,CAAA;QACf,CAAC;QAED,MAAM,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAE,CAAC,MAAM,CAAA;QAEvC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,SAAS,KAAK,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;YACvF,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACrB,CAAC;IACF,CAAC;IAED,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IAEpB,KAAK,MAAM,SAAS,IAAI,gBAAgB,EAAE,CAAC;QAC1C,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,CAAC,SAAS,CAAE,CAAA;QACpD,MAAM,KAAK,GAAG,gBAAgB,CAAC,GAAG,CAAC,SAAS,CAAE,CAAA;QAE9C,MAAM,KAAK,GAAe;YACzB,KAAK;YACL,SAAS,EAAE,cAAc;YACzB,IAAI,EAAE,SAAS;YACf,WAAW,EAAE,EAAE;YACf,
|
|
1
|
+
{"version":3,"file":"street-morphology-fst-builder.js","sourceRoot":"","sources":["../street-morphology-fst-builder.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AAEH,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AAC7D,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAA;AAG7C,OAAO,EAAE,UAAU,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAA;AAG9D;;;;GAIG;AACH,MAAM,uBAAuB,GAAG,aAAa,CAAA;AAE7C,MAAM,qBAAqB,GAAG,kBAAkB,CAAA;AAoChD;;;;;GAKG;AACH,SAAS,SAAS,CAAC,IAAY;IAC9B,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAA;IAE3B,IAAI,CAAC,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAA;IAE3D,MAAM,KAAK,GAAG,OAAO;SACnB,KAAK,CAAC,GAAG,CAAC;SACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IAE7B,IAAI,CAAC,KAAK,CAAC,MAAM;QAAE,OAAO,IAAI,CAAA;IAE9B,OAAO,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAA;AACjD,CAAC;AAED,MAAM,UAAU,wBAAwB,CAAC,IAAkC;IAC1E,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,IAAI,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;IAC9C,MAAM,gBAAgB,GAAG,IAAI,CAAC,gBAAgB,IAAI,CAAC,CAAA;IAEnD,iGAAiG;IACjG,IAAI,OAAiB,CAAA;IAErB,oGAAoG;IACpG,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC;QACzC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAA;IACvB,CAAC;SAAM,CAAC;QACP,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE;YAC5D,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,KAAK,CAAC,CAAA;YAEpD,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,WAAW,EAAE;gBAAE,OAAO,KAAK,CAAA;YAErD,IAAI,CAAC;gBACJ,QAAQ,CAAC,IAAI,CAAC,UAAU,EAAE,qBAAqB,CAAC,CAAC,CAAA;gBAEjD,OAAO,IAAI,CAAA;YACZ,CAAC;YAAC,MAAM,CAAC;gBACR,OAAO,KAAK,CAAA;YACb,CAAC;QACF,CAAC,CAAC,CAAA;IACH,CAAC;IAED,QAAQ,CAAC,UAAU,EAAE,SAAS,OAAO,CAAC,MAAM,iBAAiB,qBAAqB,EAAE,CAAC,CAAA;IAErF,4FAA4F;IAC5F,wEAAwE;IACxE,MAAM,mBAAmB,GAAG,IAAI,GAAG,EAAuB,CAAA;IAE1D,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,MAAM,EAAE,qBAAqB,CAAC,CAAA;QAC1E,MAAM,OAAO,GAAG,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;QAE9C,KAAK,MAAM,IAAI,IAAI,eAAe,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;YAClD,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,CAAA;YAE9B,IAAI,CAAC,MAAM;gBAAE,SAAQ;YACrB,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,IAAI,IAAI,GAAG,EAAU,CAAA;YAE/E,KAAK,MAAM,OAAO,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;gBACvC,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;YACtB,CAAC;YAED,mBAAmB,CAAC,GAAG,CAAC,MAAM,CAAC,SAAS,EAAE,QAAQ,CAAC,CAAA;QACpD,CAAC;IACF,CAAC;IAED,QAAQ,CAAC,SAAS,EAAE,aAAa,mBAAmB,CAAC,IAAI,oBAAoB,CAAC,CAAA;IAE9E,mEAAmE;IACnE,MAAM,gBAAgB,GAAG,CAAC,GAAG,mBAAmB,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAA;IACnE,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAkB,CAAA;IAElD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,gBAAgB,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClD,gBAAgB,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC,CAAE,EAAE,uBAAuB,GAAG,CAAC,CAAC,CAAA;IACxE,CAAC;IAED,2FAA2F;IAC3F,8FAA8F;IAC9F,MAAM,KAAK,GAAc,CAAC,EAAE,KAAK,EAAE,IAAI,GAAG,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,CAAA;IAE3D,SAAS,UAAU,CAAC,MAAgB,EAAE,KAAiB;QACtD,IAAI,CAAC,MAAM,CAAC,MAAM;YAAE,OAAM;QAC1B,IAAI,OAAO,GAAG,CAAC,CAAA;QAEf,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;YACxB,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAE,CAAA;YAC5B,IAAI,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAA;YAE5B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;gBACxB,IAAI,GAAG,KAAK,CAAC,MAAM,CAAA;gBACnB,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,GAAG,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,CAAA;gBAC5C,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;YACxB,CAAC;YAED,OAAO,GAAG,IAAI,CAAA;QACf,CAAC;QAED,MAAM,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAE,CAAC,MAAM,CAAA;QAEvC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,SAAS,KAAK,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;YACvF,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACrB,CAAC;IACF,CAAC;IAED,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IAEpB,KAAK,MAAM,SAAS,IAAI,gBAAgB,EAAE,CAAC;QAC1C,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,CAAC,SAAS,CAAE,CAAA;QACpD,MAAM,KAAK,GAAG,gBAAgB,CAAC,GAAG,CAAC,SAAS,CAAE,CAAA;QAE9C,MAAM,KAAK,GAAe;YACzB,KAAK;YACL,SAAS,EAAE,cAAc;YACzB,IAAI,EAAE,SAAS;YACf,WAAW,EAAE,EAAE;YACf,yFAAyF;YACzF,2FAA2F;YAC3F,4FAA4F;YAC5F,yEAAyE;YACzE,WAAW,EAAE,CAAC;YACd,GAAG,EAAE,CAAC;YACN,GAAG,EAAE,CAAC;SACN,CAAA;QAED,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAChC,MAAM,MAAM,GAAG,eAAe,CAAC,OAAO,CAAC,CAAA;YAEvC,IAAI,CAAC,MAAM,CAAC,MAAM;gBAAE,SAAQ;YAC5B,qFAAqF;YACrF,2FAA2F;YAC3F,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YAE9B,IAAI,MAAM,CAAC,MAAM,GAAG,gBAAgB;gBAAE,SAAQ;YAC9C,UAAU,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;YAEzB,WAAW,EAAE,CAAA;YAEb,YAAY,EAAE,CAAA;QACf,CAAC;IACF,CAAC;IAED,QAAQ,CAAC,MAAM,EAAE,eAAe,KAAK,CAAC,MAAM,YAAY,WAAW,qBAAqB,CAAC,CAAA;IAEzF,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;IACjE,MAAM,OAAO,GAAG,UAAU,CAAC,SAAS,CAAC,KAAK,CAAC,CAAA;IAE3C,MAAM,UAAU,GAAkB;QACjC,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACjC,SAAS,EAAE,OAAO,EAAE,kFAAkF;QACtG,UAAU,EAAE,KAAK,CAAC,MAAM;QACxB,UAAU,EAAE,gBAAgB,CAAC,MAAM;QACnC,SAAS;QACT,cAAc,EAAE,WAAW;QAC3B,iBAAiB,EAAE,CAAC,EAAE,uDAAuD;QAC7E,QAAQ,EAAE,IAAI,CAAC,eAAe;KAC9B,CAAA;IAED,OAAO;QACN,OAAO;QACP,UAAU;QACV,cAAc,EAAE,gBAAgB,CAAC,MAAM;QACvC,YAAY;QACZ,WAAW;QACX,OAAO;KACP,CAAA;AACF,CAAC"}
|
|
@@ -21,11 +21,39 @@
|
|
|
21
21
|
* folds to its digit form (`tenth street` → `10th street`, #723) so the grid-city ordinal
|
|
22
22
|
* cross-streets the source data spells with digits become reachable.
|
|
23
23
|
*/
|
|
24
|
+
import type { Tagged } from "type-fest";
|
|
25
|
+
/**
|
|
26
|
+
* A place name folded by {@link normalizeLocalityForKey} — the value stored in, and required to probe, every
|
|
27
|
+
* `name_key`-style column: the candidate gazetteer, the postal-city side-index, the street-centroid shard, an ancestor
|
|
28
|
+
* chain's `parent_name_key`, and the address-point locality scope.
|
|
29
|
+
*
|
|
30
|
+
* The brand is here because the fold is applied at BUILD time and is therefore mandatory at QUERY time, while a
|
|
31
|
+
* near-miss approximation of it (`toLowerCase()`, `trim()`) is still a `string`: it binds to the parameter, returns
|
|
32
|
+
* fewer rows, and the shortfall reads as a coverage gap in the data rather than a defect in the probe. Requiring the
|
|
33
|
+
* brand at the seam turns that silent under-match into a compile error. Mint one only by calling the fold.
|
|
34
|
+
*/
|
|
35
|
+
export type NameKey = Tagged<string, "NameKey">;
|
|
36
|
+
/**
|
|
37
|
+
* A street name folded by {@link normalizeStreetForKey} / {@link normalizeStreetForKeyLocale} — the value stored in a
|
|
38
|
+
* `street_norm` address-point column and required to probe it.
|
|
39
|
+
*
|
|
40
|
+
* Distinct from {@link RouteKey}: the route fold is a FURTHER canonicalization applied to some columns and not others,
|
|
41
|
+
* so the two are not interchangeable even though both are folded streets.
|
|
42
|
+
*/
|
|
43
|
+
export type StreetKey = Tagged<string, "StreetKey">;
|
|
44
|
+
/**
|
|
45
|
+
* A street key that has additionally passed {@link canonicalizeRouteKey} — the numbered-route canonical form.
|
|
46
|
+
*
|
|
47
|
+
* A separate brand from {@link StreetKey} because it identifies a DIFFERENT set of columns, and the two folds disagree
|
|
48
|
+
* exactly on the rows that motivate the route fold: binding a plain street key to a route-folded column silently misses
|
|
49
|
+
* every row whose source spelled the route differently, which is the miss class the fold exists to close.
|
|
50
|
+
*/
|
|
51
|
+
export type RouteKey = Tagged<string, "RouteKey">;
|
|
24
52
|
/**
|
|
25
53
|
* Normalize a street name for address-point keying. Same function at build time and lookup time — see module docstring
|
|
26
54
|
* for the contract.
|
|
27
55
|
*/
|
|
28
|
-
export declare function normalizeStreetForKey(street: string):
|
|
56
|
+
export declare function normalizeStreetForKey(street: string): StreetKey;
|
|
29
57
|
/**
|
|
30
58
|
* Street-name locale for the address-point key. The US path is the full USPS pipeline ({@link normalizeStreetForKey});
|
|
31
59
|
* the international paths fold + apply a SMALL, consistent per-locale type-token canonicalization. Same discipline as
|
|
@@ -34,32 +62,60 @@ export declare function normalizeStreetForKey(street: string): string;
|
|
|
34
62
|
* article, so no salient-token / multi-key index is built yet (deferred until probing shows the normalizer can't absorb
|
|
35
63
|
* the false-negatives).
|
|
36
64
|
*/
|
|
37
|
-
export type StreetLocale = "us" | "fr" | "de" | "nl";
|
|
65
|
+
export type StreetLocale = "us" | "en" | "fr" | "de" | "nl" | "pl" | "vn" | "id";
|
|
38
66
|
/**
|
|
39
67
|
* Normalize a street name for the address-point key in a non-US locale. Same function build-side and probe-side (the
|
|
40
68
|
* one-function discipline). US delegates to {@link normalizeStreetForKey}.
|
|
41
69
|
*
|
|
70
|
+
* - **en** — the shared English address-key pipeline (directionals + suffix aliases), currently identical to US.
|
|
42
71
|
* - **fr** — fold + expand leading type abbreviations and Saint/Sainte (token map).
|
|
43
72
|
* - **de** — fold + ß→ss + canonicalize the GLUED `-str(.)` suffix to `-strasse` ("Lindenstr." → "lindenstrasse",
|
|
44
73
|
* "Lindenstraße" → "lindenstrasse"); an already-full "-strasse" is left intact.
|
|
45
74
|
* - **nl** — fold + canonicalize the glued `-str` suffix to `-straat` ("Kerkstr." → "kerkstraat").
|
|
75
|
+
* - **pl** — fold + ł→l (Ł does not NFKD-decompose, so the generic fold keeps it and an undiacritized query would miss;
|
|
76
|
+
* every other Polish diacritic is a combining form the fold already strips) + STRIP the leading type token — see
|
|
77
|
+
* {@link PL_LEADING_TYPE} for the measured reason expansion was wrong for this source.
|
|
78
|
+
* - **vn** — fold + đ→d AND ð→d (both non-decomposing, and OSM mixes the two codepoints inside single values — see the
|
|
79
|
+
* branch comment). Deliberately NO type-abbreviation map yet: the common abbreviation is the single letter "Đ." for
|
|
80
|
+
* Đường, and expanding a bare folded "d" token would rewrite initials — measure the miss rate on the built shard
|
|
81
|
+
* before adding anything.
|
|
82
|
+
* - **id** — fold + expand leading type abbreviations (jl/jln→jalan, gg→gang); Indonesian street surfaces are otherwise
|
|
83
|
+
* ASCII-clean.
|
|
46
84
|
*/
|
|
47
|
-
export declare function normalizeStreetForKeyLocale(street: string, locale: StreetLocale):
|
|
85
|
+
export declare function normalizeStreetForKeyLocale(street: string, locale: StreetLocale): StreetKey;
|
|
48
86
|
/**
|
|
49
87
|
* Normalize a locality name for address-point keying (fold only — no street semantics).
|
|
50
88
|
*/
|
|
51
|
-
export declare function normalizeLocalityForKey(locality: string):
|
|
89
|
+
export declare function normalizeLocalityForKey(locality: string): NameKey;
|
|
90
|
+
/**
|
|
91
|
+
* Surface-driven street-locale ROUTER for bilingual shards — the Québec finishing move on the CA rooftop shard.
|
|
92
|
+
*
|
|
93
|
+
* A country registers ONE street locale, but Canada's street surfaces are two languages: under the `en` rules a French
|
|
94
|
+
* surface passes through mostly unchanged (both sides fold identically, so those rows stay reachable), and what breaks
|
|
95
|
+
* is abbreviation variance — the `en` rules cannot fold "boul"/"Ste-" to the full French word, so an abbreviated query
|
|
96
|
+
* misses a full-word row and vice versa. Routing on the SURFACE (not the province) also carries bilingual NB and the
|
|
97
|
+
* French street names outside Québec for free.
|
|
98
|
+
*
|
|
99
|
+
* ONE function, called by the shard BUILDER and the query PROBE alike — the #861 discipline: routing is part of the
|
|
100
|
+
* fold contract, and two transcriptions of this predicate would diverge exactly where it matters. Only an `en` base
|
|
101
|
+
* re-routes: a `fr`/`de`/`nl` shard already speaks its own rules, and the US pipeline stays untouched.
|
|
102
|
+
*
|
|
103
|
+
* Measured basis (CA shard, 2026-08-19): 183,963 distinct surfaces, 43,762 French-lead; 29,682 fold differently under
|
|
104
|
+
* fr-vs-en (888,265 rows), and the non-French half of those are English surfaces this router keeps on `en` unchanged.
|
|
105
|
+
*/
|
|
106
|
+
export declare function streetLocaleForSurface(street: string, base: StreetLocale): StreetLocale;
|
|
52
107
|
/**
|
|
53
108
|
* Strip a trailing French arrondissement designator from a FOLDED commune key ("paris 8e arrondissement" → "paris",
|
|
54
109
|
* "lyon 1er arrondissement" → "lyon", "marseille 10e arrondissement" → "marseille"). Paris, Lyon and Marseille are the
|
|
55
110
|
* only French communes subdivided into _arrondissements municipaux_; a national register (BAN) names each row per
|
|
56
111
|
* arrondissement, but a query names the base commune ("Place Bellecour, Lyon", never "…, Lyon 2e"). Applied on BOTH
|
|
57
112
|
* sides of the #1042 street-centroid key — build-side (deriving the `locality_base` column) and query-side (folding the
|
|
58
|
-
* probe commune) — so the two agree by construction (the one-function discipline).
|
|
59
|
-
*
|
|
60
|
-
*
|
|
113
|
+
* probe commune) — so the two agree by construction (the one-function discipline). The {@link NameKey} parameter is the
|
|
114
|
+
* enforcement of "must already be folded": stripping an arrondissement off an unfolded surface yields a key neither
|
|
115
|
+
* side stores. A no-op for every other commune, and the strip leaves a fold fixed point, so the result is still a
|
|
116
|
+
* {@link NameKey}. Returns the input unchanged if the strip would empty it.
|
|
61
117
|
*/
|
|
62
|
-
export declare function stripArrondissement(localityNorm:
|
|
118
|
+
export declare function stripArrondissement(localityNorm: NameKey): NameKey;
|
|
63
119
|
/**
|
|
64
120
|
* Strip a locality QUALIFIER for a query-side fallback — when an OA locality's exact normalized name misses the
|
|
65
121
|
* gazetteer's canonical name, retry with the qualifier removed. OA address data carries disambiguating qualifiers the
|
|
@@ -90,5 +146,23 @@ export declare function stripLocalityQualifier(locality: string): string;
|
|
|
90
146
|
* is ambiguous (designator unknown) and it stays unfolded — a bare-route query therefore misses rather than guessing a
|
|
91
147
|
* designator.
|
|
92
148
|
*/
|
|
93
|
-
export declare function canonicalizeRouteKey(streetNorm:
|
|
149
|
+
export declare function canonicalizeRouteKey(streetNorm: StreetKey): RouteKey;
|
|
150
|
+
/**
|
|
151
|
+
* Ordered lookup-key variants for a US/EN street span — the primary normalized key first, then the null-only recovery
|
|
152
|
+
* forms a reader may probe when the primary misses. Two register mismatches motivate them, both measured on live
|
|
153
|
+
* queries (2026-08-14):
|
|
154
|
+
*
|
|
155
|
+
* - **Doubled type** — a user types the type twice ("Saint Pauls PL St"), and the normalizer canonicalizes only the LAST
|
|
156
|
+
* type token, leaving `saint pauls pl street`, which matches nothing anywhere. The signature is visible in the key
|
|
157
|
+
* itself: a canonical type word in last position DIRECTLY after an uncanonicalized type abbreviation. The variant
|
|
158
|
+
* drops the trailing word and canonicalizes what remains (`saint pauls place`). A street genuinely named with two
|
|
159
|
+
* types keys identically on both sides and is caught by the primary probe first.
|
|
160
|
+
* - **Saint↔St register split** — the artifacts preserve each SOURCE's spelling (NYC situs keys `st pauls place`, Nassau
|
|
161
|
+
* keys `saint pauls place`), and a query arrives in whichever register the user typed. A leading `saint` or `st`
|
|
162
|
+
* token is swapped for its sibling; leading position only — a leading `st` is always the hagionym in US street names,
|
|
163
|
+
* and interior tokens ("Mount Saint Helens Dr") are out of scope until measured.
|
|
164
|
+
*
|
|
165
|
+
* Deduplicated and ordered most-literal-first, so probing the list in order preserves the primary key's precedence.
|
|
166
|
+
*/
|
|
167
|
+
export declare function streetKeyVariants(street: string, locale?: StreetLocale): StreetKey[];
|
|
94
168
|
//# sourceMappingURL=street-normalize.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"street-normalize.d.ts","sourceRoot":"","sources":["../street-normalize.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;
|
|
1
|
+
{"version":3,"file":"street-normalize.d.ts","sourceRoot":"","sources":["../street-normalize.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAGH,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,WAAW,CAAA;AAEvC;;;;;;;;;GASG;AACH,MAAM,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAA;AAE/C;;;;;;GAMG;AACH,MAAM,MAAM,SAAS,GAAG,MAAM,CAAC,MAAM,EAAE,WAAW,CAAC,CAAA;AAEnD;;;;;;GAMG;AACH,MAAM,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,EAAE,UAAU,CAAC,CAAA;AA4DjD;;;GAGG;AACH,wBAAgB,qBAAqB,CAAC,MAAM,EAAE,MAAM,GAAG,SAAS,CAkE/D;AAED;;;;;;;GAOG;AACH,MAAM,MAAM,YAAY,GAAG,IAAI,GAAG,IAAI,GAAG,IAAI,GAAG,IAAI,GAAG,IAAI,GAAG,IAAI,GAAG,IAAI,GAAG,IAAI,CAAA;AAgDhF;;;;;;;;;;;;;;;;;;GAkBG;AACH,wBAAgB,2BAA2B,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,YAAY,GAAG,SAAS,CAoE3F;AAED;;GAEG;AACH,wBAAgB,uBAAuB,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAEjE;AAWD;;;;;;;;;;;;;;;GAeG;AACH,wBAAgB,sBAAsB,CAAC,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,YAAY,GAAG,YAAY,CAEvF;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,mBAAmB,CAAC,YAAY,EAAE,OAAO,GAAG,OAAO,CAIlE;AAED;;;;;;;;;;;;GAYG;AACH,wBAAgB,sBAAsB,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CA+C/D;AAED;;;;;;;;;;;;;;;GAeG;AACH,wBAAgB,oBAAoB,CAAC,UAAU,EAAE,SAAS,GAAG,QAAQ,CASpE;AAWD;;;;;;;;;;;;;;;;GAgBG;AACH,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,GAAE,YAAmB,GAAG,SAAS,EAAE,CA8C1F"}
|
package/out/street-normalize.js
CHANGED
|
@@ -157,23 +157,56 @@ const FR_STREET_ABBREV = new Map([
|
|
|
157
157
|
["ste", "sainte"],
|
|
158
158
|
["sts", "saints"],
|
|
159
159
|
]);
|
|
160
|
+
/**
|
|
161
|
+
* Polish leading street-type tokens, STRIPPED rather than expanded. Measured on the 5.56M-row PL OSM shard
|
|
162
|
+
* (2026-08-19): OSM Poland tags `addr:street` BARE — `ulica%` covers 22 rows and `ul.%` eight, so an EXPANSION rule
|
|
163
|
+
* makes a typed query ("ul. Świętokrzyska" → "ulica swietokrzyska") miss the 3,846 bare "swietokrzyska" rows. Stripping
|
|
164
|
+
* the leading type on BOTH sides keys typed and bare surfaces identically. The full words are stripped too: the
|
|
165
|
+
* aleja/plac/osiedle populations (48,597 / 26,815 / 39,552 rows) spell the word out, and "Plac Zamkowy" must key the
|
|
166
|
+
* same as a query's "plac zamkowy" or bare "Zamkowy". Never stripped when it is the only token.
|
|
167
|
+
*/
|
|
168
|
+
const PL_LEADING_TYPE = new Set(["ul", "ulica", "al", "aleja", "aleje", "pl", "plac", "os", "osiedle"]);
|
|
169
|
+
/**
|
|
170
|
+
* Indonesian street-type abbreviations → canonical full form, leading position ("Jl. Thamrin", "Gg. Waru").
|
|
171
|
+
*/
|
|
172
|
+
const ID_STREET_ABBREV = new Map([
|
|
173
|
+
["jl", "jalan"],
|
|
174
|
+
["jln", "jalan"],
|
|
175
|
+
["gg", "gang"],
|
|
176
|
+
]);
|
|
160
177
|
/**
|
|
161
178
|
* Normalize a street name for the address-point key in a non-US locale. Same function build-side and probe-side (the
|
|
162
179
|
* one-function discipline). US delegates to {@link normalizeStreetForKey}.
|
|
163
180
|
*
|
|
181
|
+
* - **en** — the shared English address-key pipeline (directionals + suffix aliases), currently identical to US.
|
|
164
182
|
* - **fr** — fold + expand leading type abbreviations and Saint/Sainte (token map).
|
|
165
183
|
* - **de** — fold + ß→ss + canonicalize the GLUED `-str(.)` suffix to `-strasse` ("Lindenstr." → "lindenstrasse",
|
|
166
184
|
* "Lindenstraße" → "lindenstrasse"); an already-full "-strasse" is left intact.
|
|
167
185
|
* - **nl** — fold + canonicalize the glued `-str` suffix to `-straat` ("Kerkstr." → "kerkstraat").
|
|
186
|
+
* - **pl** — fold + ł→l (Ł does not NFKD-decompose, so the generic fold keeps it and an undiacritized query would miss;
|
|
187
|
+
* every other Polish diacritic is a combining form the fold already strips) + STRIP the leading type token — see
|
|
188
|
+
* {@link PL_LEADING_TYPE} for the measured reason expansion was wrong for this source.
|
|
189
|
+
* - **vn** — fold + đ→d AND ð→d (both non-decomposing, and OSM mixes the two codepoints inside single values — see the
|
|
190
|
+
* branch comment). Deliberately NO type-abbreviation map yet: the common abbreviation is the single letter "Đ." for
|
|
191
|
+
* Đường, and expanding a bare folded "d" token would rewrite initials — measure the miss rate on the built shard
|
|
192
|
+
* before adding anything.
|
|
193
|
+
* - **id** — fold + expand leading type abbreviations (jl/jln→jalan, gg→gang); Indonesian street surfaces are otherwise
|
|
194
|
+
* ASCII-clean.
|
|
168
195
|
*/
|
|
169
196
|
export function normalizeStreetForKeyLocale(street, locale) {
|
|
170
|
-
if (locale === "us")
|
|
197
|
+
if (locale === "us" || locale === "en")
|
|
171
198
|
return normalizeStreetForKey(street);
|
|
172
199
|
// Hyphen → space so a compound name keys the same whether the source or the query writes the
|
|
173
200
|
// hyphen ("Champs-Élysées", "St-Honoré") or a space — both sides fold identically, so this is pure
|
|
174
201
|
// robustness. It also splits a hyphenated abbreviation ("St-Honoré" → "st honore") into tokens the
|
|
175
|
-
// per-locale type/Saint map can see.
|
|
176
|
-
|
|
202
|
+
// per-locale type/Saint map can see. Letter maps for non-decomposing letters (ß, ł, đ) live in the
|
|
203
|
+
// per-locale branches, NEVER here: widening the shared pipeline would silently change keys under
|
|
204
|
+
// every already-built shard of the other locales.
|
|
205
|
+
const tokens = fold(street)
|
|
206
|
+
.replaceAll("ß", "ss")
|
|
207
|
+
.replaceAll("-", " ")
|
|
208
|
+
.split(/\s+/)
|
|
209
|
+
.filter((value) => value.length > 0);
|
|
177
210
|
if (!tokens.length)
|
|
178
211
|
return "";
|
|
179
212
|
switch (locale) {
|
|
@@ -198,6 +231,29 @@ export function normalizeStreetForKeyLocale(street, locale) {
|
|
|
198
231
|
}
|
|
199
232
|
}
|
|
200
233
|
break;
|
|
234
|
+
case "pl":
|
|
235
|
+
for (let i = 0; i < tokens.length; i++) {
|
|
236
|
+
tokens[i] = tokens[i].replaceAll("ł", "l");
|
|
237
|
+
}
|
|
238
|
+
if (tokens.length > 1 && PL_LEADING_TYPE.has(tokens[0])) {
|
|
239
|
+
tokens.shift();
|
|
240
|
+
}
|
|
241
|
+
break;
|
|
242
|
+
case "id":
|
|
243
|
+
for (let i = 0; i < tokens.length; i++) {
|
|
244
|
+
tokens[i] = ID_STREET_ABBREV.get(tokens[i]) ?? tokens[i];
|
|
245
|
+
}
|
|
246
|
+
break;
|
|
247
|
+
case "vn":
|
|
248
|
+
// đ (U+0111, d-with-stroke) AND ð (U+00F0, eth): OSM Vietnamese data mixes the two visually
|
|
249
|
+
// identical codepoints inside single values — the 2026-08-19 shard measured `Đường Trần Hưng
|
|
250
|
+
// Đạo` carrying d-with-stroke at the front and ETH in `Đạo`, splitting the country's most
|
|
251
|
+
// common street into two keys with the MAJORITY variant (630 vs 421 rows) unreachable by a
|
|
252
|
+
// correctly-typed query. No abbreviation map — see the locale table.
|
|
253
|
+
for (let i = 0; i < tokens.length; i++) {
|
|
254
|
+
tokens[i] = tokens[i].replaceAll("đ", "d").replaceAll("ð", "d");
|
|
255
|
+
}
|
|
256
|
+
break;
|
|
201
257
|
}
|
|
202
258
|
return tokens.join(" ");
|
|
203
259
|
}
|
|
@@ -207,15 +263,42 @@ export function normalizeStreetForKeyLocale(street, locale) {
|
|
|
207
263
|
export function normalizeLocalityForKey(locality) {
|
|
208
264
|
return fold(locality);
|
|
209
265
|
}
|
|
266
|
+
/**
|
|
267
|
+
* Leading French street-type token. French types LEAD the name ("avenue du Parc", "boul Saint-Laurent") while English
|
|
268
|
+
* types TRAIL ("Fifth Avenue", "Grosvenor Place"), so a leading type-token is the discriminating signal — "1 Avenue NE"
|
|
269
|
+
* starts with a digit and never matches. The separator lookahead is explicit rather than `\b` because a word boundary
|
|
270
|
+
* after an accented final letter ("carré") is not one to an ASCII-word `\b`.
|
|
271
|
+
*/
|
|
272
|
+
const FRENCH_LEAD_TYPE = /^(?:rue|ruelle|av|ave|avenue|boul|bd|boulevard|ch|che|chemin|all[ée]e|imp|impasse|mont[ée]e|c[ôo]te|pl|place|prom|promenade|rang|rte|route|autoroute|carr[eé]|croissant|terrasse|sentier)(?=[\s.-]|$)/i;
|
|
273
|
+
/**
|
|
274
|
+
* Surface-driven street-locale ROUTER for bilingual shards — the Québec finishing move on the CA rooftop shard.
|
|
275
|
+
*
|
|
276
|
+
* A country registers ONE street locale, but Canada's street surfaces are two languages: under the `en` rules a French
|
|
277
|
+
* surface passes through mostly unchanged (both sides fold identically, so those rows stay reachable), and what breaks
|
|
278
|
+
* is abbreviation variance — the `en` rules cannot fold "boul"/"Ste-" to the full French word, so an abbreviated query
|
|
279
|
+
* misses a full-word row and vice versa. Routing on the SURFACE (not the province) also carries bilingual NB and the
|
|
280
|
+
* French street names outside Québec for free.
|
|
281
|
+
*
|
|
282
|
+
* ONE function, called by the shard BUILDER and the query PROBE alike — the #861 discipline: routing is part of the
|
|
283
|
+
* fold contract, and two transcriptions of this predicate would diverge exactly where it matters. Only an `en` base
|
|
284
|
+
* re-routes: a `fr`/`de`/`nl` shard already speaks its own rules, and the US pipeline stays untouched.
|
|
285
|
+
*
|
|
286
|
+
* Measured basis (CA shard, 2026-08-19): 183,963 distinct surfaces, 43,762 French-lead; 29,682 fold differently under
|
|
287
|
+
* fr-vs-en (888,265 rows), and the non-French half of those are English surfaces this router keeps on `en` unchanged.
|
|
288
|
+
*/
|
|
289
|
+
export function streetLocaleForSurface(street, base) {
|
|
290
|
+
return base === "en" && FRENCH_LEAD_TYPE.test(street.trimStart()) ? "fr" : base;
|
|
291
|
+
}
|
|
210
292
|
/**
|
|
211
293
|
* Strip a trailing French arrondissement designator from a FOLDED commune key ("paris 8e arrondissement" → "paris",
|
|
212
294
|
* "lyon 1er arrondissement" → "lyon", "marseille 10e arrondissement" → "marseille"). Paris, Lyon and Marseille are the
|
|
213
295
|
* only French communes subdivided into _arrondissements municipaux_; a national register (BAN) names each row per
|
|
214
296
|
* arrondissement, but a query names the base commune ("Place Bellecour, Lyon", never "…, Lyon 2e"). Applied on BOTH
|
|
215
297
|
* sides of the #1042 street-centroid key — build-side (deriving the `locality_base` column) and query-side (folding the
|
|
216
|
-
* probe commune) — so the two agree by construction (the one-function discipline).
|
|
217
|
-
*
|
|
218
|
-
*
|
|
298
|
+
* probe commune) — so the two agree by construction (the one-function discipline). The {@link NameKey} parameter is the
|
|
299
|
+
* enforcement of "must already be folded": stripping an arrondissement off an unfolded surface yields a key neither
|
|
300
|
+
* side stores. A no-op for every other commune, and the strip leaves a fold fixed point, so the result is still a
|
|
301
|
+
* {@link NameKey}. Returns the input unchanged if the strip would empty it.
|
|
219
302
|
*/
|
|
220
303
|
export function stripArrondissement(localityNorm) {
|
|
221
304
|
const stripped = localityNorm.replace(/\s+\d+(?:er|e)\s+arrondissement$/, "").trim();
|
|
@@ -239,10 +322,31 @@ export function stripLocalityQualifier(locality) {
|
|
|
239
322
|
if (s.includes("/")) {
|
|
240
323
|
s = s.split("/")[0].trim();
|
|
241
324
|
}
|
|
242
|
-
|
|
243
|
-
|
|
244
|
-
|
|
245
|
-
|
|
325
|
+
const words = [...s.matchAll(/\S+/gu)].map((match) => ({ value: match[0], index: match.index }));
|
|
326
|
+
let qualifierStart;
|
|
327
|
+
for (let index = 1; index < words.length; index++) {
|
|
328
|
+
const word = words[index];
|
|
329
|
+
const first = word.value[0] ?? "";
|
|
330
|
+
const abbreviated = /[a-zà-ÿ]/iu.test(first) && word.value[1] === "." && word.value.length > 2;
|
|
331
|
+
const oneWordQualifier = ["im", "ob", "bei", "unter", "vor"].includes(word.value.toLowerCase());
|
|
332
|
+
const twoWordQualifier = ["an", "in"].includes(word.value.toLowerCase()) && words[index + 1]?.value.toLowerCase() === "der";
|
|
333
|
+
const hasQualifierValue = oneWordQualifier ? index + 1 < words.length : index + 2 < words.length;
|
|
334
|
+
if (abbreviated || ((oneWordQualifier || twoWordQualifier) && hasQualifierValue)) {
|
|
335
|
+
qualifierStart = word.index;
|
|
336
|
+
break;
|
|
337
|
+
}
|
|
338
|
+
}
|
|
339
|
+
const suffix = words.at(-1);
|
|
340
|
+
if (qualifierStart === undefined &&
|
|
341
|
+
suffix &&
|
|
342
|
+
words.length > 1 &&
|
|
343
|
+
(["S", "N", "E", "W", "V", "Ø", "Sø", "Fyn", "Thy", "Sjælland", "Jylland"].includes(suffix.value) ||
|
|
344
|
+
/^[A-ZÅÄÖ]{2}$/u.test(suffix.value))) {
|
|
345
|
+
qualifierStart = suffix.index;
|
|
346
|
+
}
|
|
347
|
+
if (qualifierStart !== undefined) {
|
|
348
|
+
s = s.slice(0, qualifierStart).trimEnd();
|
|
349
|
+
}
|
|
246
350
|
s = s.trim();
|
|
247
351
|
return s === locality.trim() ? "" : s;
|
|
248
352
|
}
|
|
@@ -264,8 +368,71 @@ export function stripLocalityQualifier(locality) {
|
|
|
264
368
|
*/
|
|
265
369
|
export function canonicalizeRouteKey(streetNorm) {
|
|
266
370
|
const match = /^(us|state|[a-z]{2}) (?:route|rte|rt|highway|hwy) (\d.*)$/.exec(streetNorm);
|
|
371
|
+
// A street naming no route is already its own route key. {@link StreetKey} and {@link RouteKey} are
|
|
372
|
+
// SIBLING brands rather than nested ones — precisely so neither column accepts the other's key — so
|
|
373
|
+
// re-minting here is an explicit hop through the unbranded string.
|
|
267
374
|
if (!match)
|
|
268
375
|
return streetNorm;
|
|
269
376
|
return `${match[1] === "us" ? "us" : "state"} route ${match[2]}`;
|
|
270
377
|
}
|
|
378
|
+
/**
|
|
379
|
+
* The canonical street-type words (lowercase) — the VALUE side of the codex suffix table. Membership means "this
|
|
380
|
+
* normalized token is a fully-spelled street type" ("place", "street", "road" …), which is how the doubled-type
|
|
381
|
+
* collapse below recognizes its shape without any parse-tree knowledge.
|
|
382
|
+
*/
|
|
383
|
+
const CANONICAL_TYPE_WORDS = new Set([...US_STREET_SUFFIX_LOOKUP.values()].map((v) => v.toLowerCase()));
|
|
384
|
+
/**
|
|
385
|
+
* Ordered lookup-key variants for a US/EN street span — the primary normalized key first, then the null-only recovery
|
|
386
|
+
* forms a reader may probe when the primary misses. Two register mismatches motivate them, both measured on live
|
|
387
|
+
* queries (2026-08-14):
|
|
388
|
+
*
|
|
389
|
+
* - **Doubled type** — a user types the type twice ("Saint Pauls PL St"), and the normalizer canonicalizes only the LAST
|
|
390
|
+
* type token, leaving `saint pauls pl street`, which matches nothing anywhere. The signature is visible in the key
|
|
391
|
+
* itself: a canonical type word in last position DIRECTLY after an uncanonicalized type abbreviation. The variant
|
|
392
|
+
* drops the trailing word and canonicalizes what remains (`saint pauls place`). A street genuinely named with two
|
|
393
|
+
* types keys identically on both sides and is caught by the primary probe first.
|
|
394
|
+
* - **Saint↔St register split** — the artifacts preserve each SOURCE's spelling (NYC situs keys `st pauls place`, Nassau
|
|
395
|
+
* keys `saint pauls place`), and a query arrives in whichever register the user typed. A leading `saint` or `st`
|
|
396
|
+
* token is swapped for its sibling; leading position only — a leading `st` is always the hagionym in US street names,
|
|
397
|
+
* and interior tokens ("Mount Saint Helens Dr") are out of scope until measured.
|
|
398
|
+
*
|
|
399
|
+
* Deduplicated and ordered most-literal-first, so probing the list in order preserves the primary key's precedence.
|
|
400
|
+
*/
|
|
401
|
+
export function streetKeyVariants(street, locale = "us") {
|
|
402
|
+
const primary = locale === "us" ? normalizeStreetForKey(street) : normalizeStreetForKeyLocale(street, locale);
|
|
403
|
+
const variants = primary ? [primary] : [];
|
|
404
|
+
if (!primary || (locale !== "us" && locale !== "en"))
|
|
405
|
+
return variants;
|
|
406
|
+
const tokens = primary.split(" ");
|
|
407
|
+
const last = tokens.at(-1);
|
|
408
|
+
const secondLast = tokens.at(-2);
|
|
409
|
+
if (tokens.length > 2 &&
|
|
410
|
+
last &&
|
|
411
|
+
secondLast &&
|
|
412
|
+
CANONICAL_TYPE_WORDS.has(last) &&
|
|
413
|
+
!CANONICAL_TYPE_WORDS.has(secondLast) &&
|
|
414
|
+
US_STREET_SUFFIX_LOOKUP.has(secondLast)) {
|
|
415
|
+
const collapsed = normalizeStreetForKey(tokens.slice(0, -1).join(" "));
|
|
416
|
+
if (collapsed && !variants.includes(collapsed)) {
|
|
417
|
+
variants.push(collapsed);
|
|
418
|
+
}
|
|
419
|
+
}
|
|
420
|
+
// Snapshot before the swap pass — it appends while reading, and iterating the live array would
|
|
421
|
+
// re-visit its own additions.
|
|
422
|
+
const preSwap = variants.slice();
|
|
423
|
+
for (const variant of preSwap) {
|
|
424
|
+
const [head, ...rest] = variant.split(" ");
|
|
425
|
+
const swapped = head === "saint" ? "st" : head === "st" ? "saint" : null;
|
|
426
|
+
if (swapped && rest.length) {
|
|
427
|
+
// Substituting a leading hagionym token in an already-normalized key leaves a fold fixed point —
|
|
428
|
+
// the suffix pass never rewrites position 0 — so the swap yields a {@link StreetKey} without
|
|
429
|
+
// re-folding.
|
|
430
|
+
const candidate = [swapped, ...rest].join(" ");
|
|
431
|
+
if (!variants.includes(candidate)) {
|
|
432
|
+
variants.push(candidate);
|
|
433
|
+
}
|
|
434
|
+
}
|
|
435
|
+
}
|
|
436
|
+
return variants;
|
|
437
|
+
}
|
|
271
438
|
//# sourceMappingURL=street-normalize.js.map
|