@mailwoman/corpus 8.6.0 → 9.1.0

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.
Files changed (497) hide show
  1. package/README.md +2 -2
  2. package/data/PROVENANCE.md +214 -0
  3. package/data/sub-venue-lexicon.json +18269 -0
  4. package/out/src/adapters/ban/adapter.d.ts +5 -4
  5. package/out/src/adapters/ban/adapter.d.ts.map +1 -1
  6. package/out/src/adapters/ban/adapter.js +62 -69
  7. package/out/src/adapters/ban/adapter.js.map +1 -1
  8. package/out/src/adapters/ban/street-decompose.d.ts.map +1 -1
  9. package/out/src/adapters/ban/street-decompose.js +18 -25
  10. package/out/src/adapters/ban/street-decompose.js.map +1 -1
  11. package/out/src/adapters/fcc-bdc/adapter.d.ts +1 -7
  12. package/out/src/adapters/fcc-bdc/adapter.d.ts.map +1 -1
  13. package/out/src/adapters/fcc-bdc/adapter.js +4 -26
  14. package/out/src/adapters/fcc-bdc/adapter.js.map +1 -1
  15. package/out/src/adapters/geonames/adapter.d.ts +1 -1
  16. package/out/src/adapters/geonames/adapter.d.ts.map +1 -1
  17. package/out/src/adapters/geonames/adapter.js +72 -78
  18. package/out/src/adapters/geonames/adapter.js.map +1 -1
  19. package/out/src/adapters/geonames-postal/adapter.d.ts +1 -1
  20. package/out/src/adapters/geonames-postal/adapter.d.ts.map +1 -1
  21. package/out/src/adapters/geonames-postal/adapter.js +46 -51
  22. package/out/src/adapters/geonames-postal/adapter.js.map +1 -1
  23. package/out/src/adapters/gnaf/adapter.d.ts +1 -1
  24. package/out/src/adapters/gnaf/adapter.d.ts.map +1 -1
  25. package/out/src/adapters/gnaf/adapter.js +6 -9
  26. package/out/src/adapters/gnaf/adapter.js.map +1 -1
  27. package/out/src/adapters/gnaf/assemble.d.ts.map +1 -1
  28. package/out/src/adapters/gnaf/assemble.js +7 -12
  29. package/out/src/adapters/gnaf/assemble.js.map +1 -1
  30. package/out/src/adapters/index.d.ts +1 -1
  31. package/out/src/adapters/index.d.ts.map +1 -1
  32. package/out/src/adapters/index.js +1 -1
  33. package/out/src/adapters/index.js.map +1 -1
  34. package/out/src/adapters/openaddresses/adapter.d.ts +1 -1
  35. package/out/src/adapters/openaddresses/adapter.d.ts.map +1 -1
  36. package/out/src/adapters/openaddresses/adapter.js +5 -10
  37. package/out/src/adapters/openaddresses/adapter.js.map +1 -1
  38. package/out/src/adapters/overture/adapter.d.ts +1 -1
  39. package/out/src/adapters/overture/adapter.d.ts.map +1 -1
  40. package/out/src/adapters/overture/adapter.js +5 -9
  41. package/out/src/adapters/overture/adapter.js.map +1 -1
  42. package/out/src/adapters/state-hi-schools/adapter.d.ts +1 -1
  43. package/out/src/adapters/state-hi-schools/adapter.d.ts.map +1 -1
  44. package/out/src/adapters/state-hi-schools/adapter.js +57 -75
  45. package/out/src/adapters/state-hi-schools/adapter.js.map +1 -1
  46. package/out/src/adapters/state-ia-contractors/adapter.d.ts +1 -1
  47. package/out/src/adapters/state-ia-contractors/adapter.d.ts.map +1 -1
  48. package/out/src/adapters/state-ia-contractors/adapter.js +60 -78
  49. package/out/src/adapters/state-ia-contractors/adapter.js.map +1 -1
  50. package/out/src/adapters/state-ny-notaries/adapter.d.ts +1 -1
  51. package/out/src/adapters/state-ny-notaries/adapter.d.ts.map +1 -1
  52. package/out/src/adapters/state-ny-notaries/adapter.js +69 -87
  53. package/out/src/adapters/state-ny-notaries/adapter.js.map +1 -1
  54. package/out/src/adapters/state-tx-notaries/adapter.d.ts +1 -1
  55. package/out/src/adapters/state-tx-notaries/adapter.d.ts.map +1 -1
  56. package/out/src/adapters/state-tx-notaries/adapter.js +71 -89
  57. package/out/src/adapters/state-tx-notaries/adapter.js.map +1 -1
  58. package/out/src/adapters/synth-po-box/adapter.d.ts +3 -3
  59. package/out/src/adapters/synth-po-box/adapter.d.ts.map +1 -1
  60. package/out/src/adapters/synth-po-box/adapter.js +10 -18
  61. package/out/src/adapters/synth-po-box/adapter.js.map +1 -1
  62. package/out/src/adapters/tiger/adapter.d.ts +1 -1
  63. package/out/src/adapters/tiger/adapter.d.ts.map +1 -1
  64. package/out/src/adapters/tiger/adapter.js +2 -2
  65. package/out/src/adapters/tiger/adapter.js.map +1 -1
  66. package/out/src/adapters/tiger/street-decompose.d.ts.map +1 -1
  67. package/out/src/adapters/tiger/street-decompose.js +18 -27
  68. package/out/src/adapters/tiger/street-decompose.js.map +1 -1
  69. package/out/src/adapters/usgov-hrsa-fqhc/adapter.d.ts +4 -4
  70. package/out/src/adapters/usgov-hrsa-fqhc/adapter.d.ts.map +1 -1
  71. package/out/src/adapters/usgov-hrsa-fqhc/adapter.js +64 -82
  72. package/out/src/adapters/usgov-hrsa-fqhc/adapter.js.map +1 -1
  73. package/out/src/adapters/usgov-imls-pls/adapter.d.ts +1 -1
  74. package/out/src/adapters/usgov-imls-pls/adapter.d.ts.map +1 -1
  75. package/out/src/adapters/usgov-imls-pls/adapter.js +62 -84
  76. package/out/src/adapters/usgov-imls-pls/adapter.js.map +1 -1
  77. package/out/src/adapters/usgov-irs-bmf/adapter.d.ts +1 -1
  78. package/out/src/adapters/usgov-irs-bmf/adapter.d.ts.map +1 -1
  79. package/out/src/adapters/usgov-irs-bmf/adapter.js +59 -66
  80. package/out/src/adapters/usgov-irs-bmf/adapter.js.map +1 -1
  81. package/out/src/adapters/usgov-nad/adapter.d.ts +1 -1
  82. package/out/src/adapters/usgov-nad/adapter.d.ts.map +1 -1
  83. package/out/src/adapters/usgov-nad/adapter.js +6 -9
  84. package/out/src/adapters/usgov-nad/adapter.js.map +1 -1
  85. package/out/src/adapters/usgov-nppes/adapter.d.ts +1 -1
  86. package/out/src/adapters/usgov-nppes/adapter.d.ts.map +1 -1
  87. package/out/src/adapters/usgov-nppes/adapter.js +60 -78
  88. package/out/src/adapters/usgov-nppes/adapter.js.map +1 -1
  89. package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.d.ts +1 -1
  90. package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.d.ts.map +1 -1
  91. package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.js +53 -73
  92. package/out/src/adapters/usgov-samhsa-treatment-locator/adapter.js.map +1 -1
  93. package/out/src/{adapter.d.ts → adapters/utils/index.d.ts} +43 -2
  94. package/out/src/adapters/utils/index.d.ts.map +1 -0
  95. package/out/src/{adapter.js → adapters/utils/index.js} +47 -2
  96. package/out/src/adapters/utils/index.js.map +1 -0
  97. package/out/src/adapters/wof-admin-jp/adapter.d.ts +1 -1
  98. package/out/src/adapters/wof-admin-jp/adapter.d.ts.map +1 -1
  99. package/out/src/adapters/wof-admin-jp/adapter.js +24 -6
  100. package/out/src/adapters/wof-admin-jp/adapter.js.map +1 -1
  101. package/out/src/adapters/wof-admin-json/adapter.d.ts +2 -2
  102. package/out/src/adapters/wof-admin-json/adapter.d.ts.map +1 -1
  103. package/out/src/adapters/wof-admin-json/adapter.js +2 -2
  104. package/out/src/adapters/wof-admin-json/adapter.js.map +1 -1
  105. package/out/src/adapters/wof-postalcode-json/adapter.d.ts +2 -2
  106. package/out/src/adapters/wof-postalcode-json/adapter.d.ts.map +1 -1
  107. package/out/src/adapters/wof-postalcode-json/adapter.js +2 -2
  108. package/out/src/adapters/wof-postalcode-json/adapter.js.map +1 -1
  109. package/out/src/build.d.ts +3 -4
  110. package/out/src/build.d.ts.map +1 -1
  111. package/out/src/build.js +6 -8
  112. package/out/src/build.js.map +1 -1
  113. package/out/src/index.d.ts +3 -17
  114. package/out/src/index.d.ts.map +1 -1
  115. package/out/src/index.js +3 -17
  116. package/out/src/index.js.map +1 -1
  117. package/out/src/name-prone-us-suffixes.d.ts +12 -0
  118. package/out/src/name-prone-us-suffixes.d.ts.map +1 -0
  119. package/out/src/name-prone-us-suffixes.js +12 -0
  120. package/out/src/name-prone-us-suffixes.js.map +1 -0
  121. package/out/src/parquet-wrapper/reader.d.ts +8 -0
  122. package/out/src/parquet-wrapper/reader.d.ts.map +1 -1
  123. package/out/src/parquet-wrapper/reader.js +16 -0
  124. package/out/src/parquet-wrapper/reader.js.map +1 -1
  125. package/out/src/parquet-wrapper/schema.d.ts +4 -0
  126. package/out/src/parquet-wrapper/schema.d.ts.map +1 -1
  127. package/out/src/parquet-wrapper/schema.js.map +1 -1
  128. package/out/src/runner.d.ts +2 -2
  129. package/out/src/runner.d.ts.map +1 -1
  130. package/out/src/runner.js +1 -1
  131. package/out/src/runner.js.map +1 -1
  132. package/out/src/shard-recipes/anchor-absorption.d.ts.map +1 -1
  133. package/out/src/shard-recipes/anchor-absorption.js +5 -4
  134. package/out/src/shard-recipes/anchor-absorption.js.map +1 -1
  135. package/out/src/shard-recipes/boundary-stress.d.ts.map +1 -1
  136. package/out/src/shard-recipes/boundary-stress.js +2 -2
  137. package/out/src/shard-recipes/boundary-stress.js.map +1 -1
  138. package/out/src/shard-recipes/country-balanced.d.ts.map +1 -1
  139. package/out/src/shard-recipes/country-balanced.js +75 -75
  140. package/out/src/shard-recipes/country-balanced.js.map +1 -1
  141. package/out/src/shard-recipes/fr-admin-split.js +2 -2
  142. package/out/src/shard-recipes/fr-admin-split.js.map +1 -1
  143. package/out/src/shard-recipes/fr-fragment.d.ts.map +1 -1
  144. package/out/src/shard-recipes/fr-fragment.js +8 -5
  145. package/out/src/shard-recipes/fr-fragment.js.map +1 -1
  146. package/out/src/shard-recipes/fr-lieudit.js +2 -2
  147. package/out/src/shard-recipes/fr-lieudit.js.map +1 -1
  148. package/out/src/shard-recipes/fr-order.d.ts.map +1 -1
  149. package/out/src/shard-recipes/fr-order.js +18 -72
  150. package/out/src/shard-recipes/fr-order.js.map +1 -1
  151. package/out/src/shard-recipes/german.d.ts.map +1 -1
  152. package/out/src/shard-recipes/german.js +17 -70
  153. package/out/src/shard-recipes/german.js.map +1 -1
  154. package/out/src/shard-recipes/house-venue.d.ts.map +1 -1
  155. package/out/src/shard-recipes/house-venue.js +1 -1
  156. package/out/src/shard-recipes/house-venue.js.map +1 -1
  157. package/out/src/shard-recipes/index.d.ts.map +1 -1
  158. package/out/src/shard-recipes/index.js +4 -1
  159. package/out/src/shard-recipes/index.js.map +1 -1
  160. package/out/src/shard-recipes/intersection.d.ts +2 -2
  161. package/out/src/shard-recipes/intersection.d.ts.map +1 -1
  162. package/out/src/shard-recipes/intersection.js +28 -72
  163. package/out/src/shard-recipes/intersection.js.map +1 -1
  164. package/out/src/shard-recipes/locale.d.ts +5 -5
  165. package/out/src/shard-recipes/locale.d.ts.map +1 -1
  166. package/out/src/shard-recipes/locale.js +25 -35
  167. package/out/src/shard-recipes/locale.js.map +1 -1
  168. package/out/src/shard-recipes/no-fragment.d.ts.map +1 -1
  169. package/out/src/shard-recipes/no-fragment.js +8 -5
  170. package/out/src/shard-recipes/no-fragment.js.map +1 -1
  171. package/out/src/shard-recipes/no-street-led.d.ts.map +1 -1
  172. package/out/src/shard-recipes/no-street-led.js +8 -5
  173. package/out/src/shard-recipes/no-street-led.js.map +1 -1
  174. package/out/src/shard-recipes/no-street.d.ts.map +1 -1
  175. package/out/src/shard-recipes/no-street.js +1 -1
  176. package/out/src/shard-recipes/no-street.js.map +1 -1
  177. package/out/src/shard-recipes/po-box-cedex.d.ts +3 -3
  178. package/out/src/shard-recipes/po-box-cedex.d.ts.map +1 -1
  179. package/out/src/shard-recipes/po-box-cedex.js +101 -127
  180. package/out/src/shard-recipes/po-box-cedex.js.map +1 -1
  181. package/out/src/shard-recipes/po-box.d.ts.map +1 -1
  182. package/out/src/shard-recipes/po-box.js +1 -1
  183. package/out/src/shard-recipes/po-box.js.map +1 -1
  184. package/out/src/shard-recipes/scaffold.d.ts +58 -9
  185. package/out/src/shard-recipes/scaffold.d.ts.map +1 -1
  186. package/out/src/shard-recipes/scaffold.js +87 -29
  187. package/out/src/shard-recipes/scaffold.js.map +1 -1
  188. package/out/src/shard-recipes/street-affix.d.ts +54 -0
  189. package/out/src/shard-recipes/street-affix.d.ts.map +1 -1
  190. package/out/src/shard-recipes/street-affix.js +253 -111
  191. package/out/src/shard-recipes/street-affix.js.map +1 -1
  192. package/out/src/shard-recipes/street-bare.d.ts.map +1 -1
  193. package/out/src/shard-recipes/street-bare.js +3 -3
  194. package/out/src/shard-recipes/street-bare.js.map +1 -1
  195. package/out/src/shard-recipes/street.d.ts.map +1 -1
  196. package/out/src/shard-recipes/street.js +2 -2
  197. package/out/src/shard-recipes/street.js.map +1 -1
  198. package/out/src/shard-recipes/sub-venue-sources.d.ts +228 -0
  199. package/out/src/shard-recipes/sub-venue-sources.d.ts.map +1 -0
  200. package/out/src/shard-recipes/sub-venue-sources.js +664 -0
  201. package/out/src/shard-recipes/sub-venue-sources.js.map +1 -0
  202. package/out/src/shard-recipes/sub-venue.d.ts +269 -0
  203. package/out/src/shard-recipes/sub-venue.d.ts.map +1 -0
  204. package/out/src/shard-recipes/sub-venue.js +709 -0
  205. package/out/src/shard-recipes/sub-venue.js.map +1 -0
  206. package/out/src/shard-recipes/unit.d.ts +1 -1
  207. package/out/src/shard-recipes/unit.d.ts.map +1 -1
  208. package/out/src/shard-recipes/unit.js +26 -74
  209. package/out/src/shard-recipes/unit.js.map +1 -1
  210. package/out/src/{synthesize-anchor-absorption.d.ts → synthesizers/anchor-absorption.d.ts} +1 -1
  211. package/out/src/synthesizers/anchor-absorption.d.ts.map +1 -0
  212. package/out/src/{synthesize-anchor-absorption.js → synthesizers/anchor-absorption.js} +1 -1
  213. package/out/src/synthesizers/anchor-absorption.js.map +1 -0
  214. package/out/src/{synthesize-boundary-stress.d.ts → synthesizers/boundary-stress.d.ts} +3 -3
  215. package/out/src/synthesizers/boundary-stress.d.ts.map +1 -0
  216. package/out/src/{synthesize-boundary-stress.js → synthesizers/boundary-stress.js} +2 -2
  217. package/out/src/synthesizers/boundary-stress.js.map +1 -0
  218. package/out/src/{synthesize-german.d.ts → synthesizers/german.d.ts} +2 -2
  219. package/out/src/synthesizers/german.d.ts.map +1 -0
  220. package/out/src/{synthesize-german.js → synthesizers/german.js} +2 -2
  221. package/out/src/synthesizers/german.js.map +1 -0
  222. package/out/src/{synthesize-house-venue.d.ts → synthesizers/house-venue.d.ts} +4 -4
  223. package/out/src/synthesizers/house-venue.d.ts.map +1 -0
  224. package/out/src/{synthesize-house-venue.js → synthesizers/house-venue.js} +4 -4
  225. package/out/src/synthesizers/house-venue.js.map +1 -0
  226. package/out/src/{synthesize-intersection.d.ts → synthesizers/intersection.d.ts} +2 -2
  227. package/out/src/synthesizers/intersection.d.ts.map +1 -0
  228. package/out/src/{synthesize-intersection.js → synthesizers/intersection.js} +1 -1
  229. package/out/src/synthesizers/intersection.js.map +1 -0
  230. package/out/src/{synthesize-no-street.d.ts → synthesizers/no-street.d.ts} +3 -3
  231. package/out/src/synthesizers/no-street.d.ts.map +1 -0
  232. package/out/src/{synthesize-no-street.js → synthesizers/no-street.js} +2 -2
  233. package/out/src/synthesizers/no-street.js.map +1 -0
  234. package/out/src/{synthesize-po-box.d.ts → synthesizers/po-box.d.ts} +2 -2
  235. package/out/src/synthesizers/po-box.d.ts.map +1 -0
  236. package/out/src/{synthesize-po-box.js → synthesizers/po-box.js} +1 -1
  237. package/out/src/synthesizers/po-box.js.map +1 -0
  238. package/out/src/{synthesize-street.d.ts → synthesizers/street.d.ts} +2 -2
  239. package/out/src/synthesizers/street.d.ts.map +1 -0
  240. package/out/src/{synthesize-street.js → synthesizers/street.js} +2 -2
  241. package/out/src/synthesizers/street.js.map +1 -0
  242. package/out/src/{synthesize.d.ts → synthesizers/utils.d.ts} +3 -3
  243. package/out/src/synthesizers/utils.d.ts.map +1 -0
  244. package/out/src/{synthesize.js → synthesizers/utils.js} +4 -15
  245. package/out/src/synthesizers/utils.js.map +1 -0
  246. package/out/src/tools/align-shard.d.ts.map +1 -1
  247. package/out/src/tools/align-shard.js +4 -8
  248. package/out/src/tools/align-shard.js.map +1 -1
  249. package/out/src/tools/audit.d.ts.map +1 -1
  250. package/out/src/tools/audit.js +4 -4
  251. package/out/src/tools/audit.js.map +1 -1
  252. package/out/src/tools/corpus-stats.d.ts +2 -2
  253. package/out/src/tools/corpus-stats.d.ts.map +1 -1
  254. package/out/src/tools/corpus-stats.js +18 -31
  255. package/out/src/tools/corpus-stats.js.map +1 -1
  256. package/out/src/tools/fetch/download.d.ts.map +1 -1
  257. package/out/src/tools/fetch/download.js +5 -6
  258. package/out/src/tools/fetch/download.js.map +1 -1
  259. package/out/src/tools/fetch/imls-pls.d.ts.map +1 -1
  260. package/out/src/tools/fetch/imls-pls.js +3 -15
  261. package/out/src/tools/fetch/imls-pls.js.map +1 -1
  262. package/out/src/tools/fetch/index.d.ts +14 -1
  263. package/out/src/tools/fetch/index.d.ts.map +1 -1
  264. package/out/src/tools/fetch/index.js +14 -1
  265. package/out/src/tools/fetch/index.js.map +1 -1
  266. package/out/src/tools/fetch/nad.d.ts +1 -1
  267. package/out/src/tools/fetch/nad.js +1 -1
  268. package/out/src/tools/fetch/nppes.d.ts.map +1 -1
  269. package/out/src/tools/fetch/nppes.js +8 -14
  270. package/out/src/tools/fetch/nppes.js.map +1 -1
  271. package/out/src/tools/fetch/openaddresses.d.ts +1 -1
  272. package/out/src/tools/fetch/openaddresses.js +1 -1
  273. package/out/src/tools/fetch/ourairports.d.ts +45 -0
  274. package/out/src/tools/fetch/ourairports.d.ts.map +1 -0
  275. package/out/src/tools/fetch/ourairports.js +123 -0
  276. package/out/src/tools/fetch/ourairports.js.map +1 -0
  277. package/out/src/tools/fetch/wikidata-subvenue.d.ts +171 -0
  278. package/out/src/tools/fetch/wikidata-subvenue.d.ts.map +1 -0
  279. package/out/src/tools/fetch/wikidata-subvenue.js +275 -0
  280. package/out/src/tools/fetch/wikidata-subvenue.js.map +1 -0
  281. package/out/src/tools/golden-expand.d.ts.map +1 -1
  282. package/out/src/tools/golden-expand.js +18 -16
  283. package/out/src/tools/golden-expand.js.map +1 -1
  284. package/out/src/tools/golden-promote.d.ts.map +1 -1
  285. package/out/src/tools/golden-promote.js +12 -6
  286. package/out/src/tools/golden-promote.js.map +1 -1
  287. package/out/src/tools/golden-relabel-street.d.ts +196 -0
  288. package/out/src/tools/golden-relabel-street.d.ts.map +1 -0
  289. package/out/src/tools/golden-relabel-street.js +440 -0
  290. package/out/src/tools/golden-relabel-street.js.map +1 -0
  291. package/out/src/tools/index.d.ts +4 -0
  292. package/out/src/tools/index.d.ts.map +1 -1
  293. package/out/src/tools/index.js +4 -0
  294. package/out/src/tools/index.js.map +1 -1
  295. package/out/src/tools/jsonl-to-parquet.d.ts.map +1 -1
  296. package/out/src/tools/jsonl-to-parquet.js +3 -2
  297. package/out/src/tools/jsonl-to-parquet.js.map +1 -1
  298. package/out/src/tools/lint-shard-vocab.d.ts.map +1 -1
  299. package/out/src/tools/lint-shard-vocab.js +3 -2
  300. package/out/src/tools/lint-shard-vocab.js.map +1 -1
  301. package/out/src/tools/lint-shard.d.ts +3 -3
  302. package/out/src/tools/lint-shard.d.ts.map +1 -1
  303. package/out/src/tools/lint-shard.js +23 -32
  304. package/out/src/tools/lint-shard.js.map +1 -1
  305. package/out/src/tools/overlay-manifest.d.ts.map +1 -1
  306. package/out/src/tools/overlay-manifest.js +2 -1
  307. package/out/src/tools/overlay-manifest.js.map +1 -1
  308. package/out/src/tools/overture-subvenue.d.ts +112 -0
  309. package/out/src/tools/overture-subvenue.d.ts.map +1 -0
  310. package/out/src/tools/overture-subvenue.js +144 -0
  311. package/out/src/tools/overture-subvenue.js.map +1 -0
  312. package/out/src/tools/shard-kryptonite.d.ts +1 -1
  313. package/out/src/tools/shard-kryptonite.d.ts.map +1 -1
  314. package/out/src/tools/shard-kryptonite.js +6 -6
  315. package/out/src/tools/shard-kryptonite.js.map +1 -1
  316. package/out/src/tools/shard-translit.d.ts +6 -3
  317. package/out/src/tools/shard-translit.d.ts.map +1 -1
  318. package/out/src/tools/shard-translit.js +9 -8
  319. package/out/src/tools/shard-translit.js.map +1 -1
  320. package/out/src/tools/sub-venue-lexicon.d.ts +507 -0
  321. package/out/src/tools/sub-venue-lexicon.d.ts.map +1 -0
  322. package/out/src/tools/sub-venue-lexicon.js +817 -0
  323. package/out/src/tools/sub-venue-lexicon.js.map +1 -0
  324. package/out/src/tools/sub-venue-promotions.d.ts +94 -0
  325. package/out/src/tools/sub-venue-promotions.d.ts.map +1 -0
  326. package/out/src/tools/sub-venue-promotions.js +266 -0
  327. package/out/src/tools/sub-venue-promotions.js.map +1 -0
  328. package/out/src/{align.d.ts → utils/align.d.ts} +1 -1
  329. package/out/src/utils/align.d.ts.map +1 -0
  330. package/out/src/utils/align.js.map +1 -0
  331. package/out/src/{golden.d.ts → utils/golden.d.ts} +5 -0
  332. package/out/src/utils/golden.d.ts.map +1 -0
  333. package/out/src/{golden.js → utils/golden.js} +24 -10
  334. package/out/src/utils/golden.js.map +1 -0
  335. package/out/src/utils/index.d.ts +14 -0
  336. package/out/src/utils/index.d.ts.map +1 -0
  337. package/out/src/utils/index.js +14 -0
  338. package/out/src/utils/index.js.map +1 -0
  339. package/out/src/utils/license.d.ts.map +1 -0
  340. package/out/src/utils/license.js.map +1 -0
  341. package/out/src/{parquet.d.ts → utils/parquet.d.ts} +2 -2
  342. package/out/src/utils/parquet.d.ts.map +1 -0
  343. package/out/src/{parquet.js → utils/parquet.js} +3 -15
  344. package/out/src/utils/parquet.js.map +1 -0
  345. package/out/src/{split.d.ts → utils/split.d.ts} +8 -1
  346. package/out/src/utils/split.d.ts.map +1 -0
  347. package/out/src/{split.js → utils/split.js} +7 -0
  348. package/out/src/utils/split.js.map +1 -0
  349. package/out/src/utils/tokenize.d.ts.map +1 -0
  350. package/out/src/utils/tokenize.js.map +1 -0
  351. package/out/src/{wof-json.d.ts → utils/wof-json.d.ts} +2 -2
  352. package/out/src/utils/wof-json.d.ts.map +1 -0
  353. package/out/src/{wof-json.js → utils/wof-json.js} +5 -8
  354. package/out/src/utils/wof-json.js.map +1 -0
  355. package/package.json +66 -20
  356. package/src/adapters/ban/adapter.ts +58 -68
  357. package/src/adapters/ban/street-decompose.ts +17 -25
  358. package/src/adapters/fcc-bdc/adapter.ts +5 -36
  359. package/src/adapters/geonames/adapter.ts +67 -76
  360. package/src/adapters/geonames-postal/adapter.ts +45 -54
  361. package/src/adapters/gnaf/adapter.ts +7 -11
  362. package/src/adapters/gnaf/assemble.ts +7 -11
  363. package/src/adapters/index.ts +3 -2
  364. package/src/adapters/openaddresses/adapter.ts +6 -12
  365. package/src/adapters/overture/adapter.ts +6 -10
  366. package/src/adapters/state-hi-schools/adapter.ts +53 -77
  367. package/src/adapters/state-ia-contractors/adapter.ts +55 -80
  368. package/src/adapters/state-ny-notaries/adapter.ts +64 -89
  369. package/src/adapters/state-tx-notaries/adapter.ts +63 -88
  370. package/src/adapters/synth-po-box/adapter.ts +12 -23
  371. package/src/adapters/tiger/adapter.ts +4 -3
  372. package/src/adapters/tiger/street-decompose.ts +21 -31
  373. package/src/adapters/usgov-hrsa-fqhc/adapter.ts +64 -88
  374. package/src/adapters/usgov-imls-pls/adapter.ts +57 -86
  375. package/src/adapters/usgov-irs-bmf/adapter.ts +55 -66
  376. package/src/adapters/usgov-nad/adapter.ts +7 -11
  377. package/src/adapters/usgov-nppes/adapter.ts +54 -79
  378. package/src/adapters/usgov-samhsa-treatment-locator/adapter.ts +48 -75
  379. package/src/{adapter.ts → adapters/utils/index.ts} +59 -3
  380. package/src/adapters/wof-admin-jp/adapter.ts +35 -9
  381. package/src/adapters/wof-admin-json/adapter.ts +3 -4
  382. package/src/adapters/wof-postalcode-json/adapter.ts +3 -4
  383. package/src/build.ts +12 -11
  384. package/src/index.ts +3 -17
  385. package/src/name-prone-us-suffixes.ts +12 -0
  386. package/src/parquet-wrapper/reader.ts +19 -0
  387. package/src/parquet-wrapper/schema.ts +4 -0
  388. package/src/runner.ts +7 -2
  389. package/src/shard-recipes/anchor-absorption.ts +5 -4
  390. package/src/shard-recipes/boundary-stress.ts +6 -2
  391. package/src/shard-recipes/country-balanced.ts +77 -91
  392. package/src/shard-recipes/fr-admin-split.ts +3 -3
  393. package/src/shard-recipes/fr-fragment.ts +10 -7
  394. package/src/shard-recipes/fr-lieudit.ts +3 -3
  395. package/src/shard-recipes/fr-order.ts +18 -83
  396. package/src/shard-recipes/german.ts +17 -80
  397. package/src/shard-recipes/house-venue.ts +2 -1
  398. package/src/shard-recipes/index.ts +4 -1
  399. package/src/shard-recipes/intersection.ts +37 -74
  400. package/src/shard-recipes/locale.ts +31 -39
  401. package/src/shard-recipes/no-fragment.ts +10 -7
  402. package/src/shard-recipes/no-street-led.ts +10 -7
  403. package/src/shard-recipes/no-street.ts +2 -1
  404. package/src/shard-recipes/po-box-cedex.ts +127 -161
  405. package/src/shard-recipes/po-box.ts +6 -1
  406. package/src/shard-recipes/scaffold.ts +125 -27
  407. package/src/shard-recipes/street-affix.ts +335 -134
  408. package/src/shard-recipes/street-bare.ts +4 -3
  409. package/src/shard-recipes/street.ts +3 -2
  410. package/src/shard-recipes/sub-venue-sources.ts +895 -0
  411. package/src/shard-recipes/sub-venue.ts +982 -0
  412. package/src/shard-recipes/unit.ts +27 -85
  413. package/src/{synthesize-boundary-stress.ts → synthesizers/boundary-stress.ts} +2 -2
  414. package/src/{synthesize-german.ts → synthesizers/german.ts} +2 -2
  415. package/src/{synthesize-house-venue.ts → synthesizers/house-venue.ts} +4 -4
  416. package/src/{synthesize-intersection.ts → synthesizers/intersection.ts} +1 -1
  417. package/src/{synthesize-no-street.ts → synthesizers/no-street.ts} +2 -2
  418. package/src/{synthesize-po-box.ts → synthesizers/po-box.ts} +1 -1
  419. package/src/{synthesize-street.ts → synthesizers/street.ts} +2 -2
  420. package/src/{synthesize.ts → synthesizers/utils.ts} +4 -18
  421. package/src/tools/align-shard.ts +4 -7
  422. package/src/tools/audit.ts +6 -5
  423. package/src/tools/corpus-stats.ts +25 -33
  424. package/src/tools/fetch/download.ts +7 -5
  425. package/src/tools/fetch/imls-pls.ts +4 -18
  426. package/src/tools/fetch/index.ts +14 -1
  427. package/src/tools/fetch/nad.ts +1 -1
  428. package/src/tools/fetch/nppes.ts +9 -16
  429. package/src/tools/fetch/openaddresses.ts +1 -1
  430. package/src/tools/fetch/ourairports.ts +166 -0
  431. package/src/tools/fetch/wikidata-subvenue.ts +386 -0
  432. package/src/tools/golden-expand.ts +18 -13
  433. package/src/tools/golden-promote.ts +15 -6
  434. package/src/tools/golden-relabel-street.ts +685 -0
  435. package/src/tools/index.ts +4 -0
  436. package/src/tools/jsonl-to-parquet.ts +3 -2
  437. package/src/tools/lint-shard-vocab.ts +3 -2
  438. package/src/tools/lint-shard.ts +33 -36
  439. package/src/tools/overlay-manifest.ts +2 -1
  440. package/src/tools/overture-subvenue.ts +215 -0
  441. package/src/tools/shard-kryptonite.ts +8 -9
  442. package/src/tools/shard-translit.ts +22 -11
  443. package/src/tools/sub-venue-lexicon.ts +1250 -0
  444. package/src/tools/sub-venue-promotions.ts +330 -0
  445. package/src/{align.ts → utils/align.ts} +1 -1
  446. package/src/{golden.ts → utils/golden.ts} +25 -11
  447. package/src/utils/index.ts +14 -0
  448. package/src/{parquet.ts → utils/parquet.ts} +5 -19
  449. package/src/{split.ts → utils/split.ts} +8 -2
  450. package/src/{wof-json.ts → utils/wof-json.ts} +5 -8
  451. package/out/src/adapter.d.ts.map +0 -1
  452. package/out/src/adapter.js.map +0 -1
  453. package/out/src/align.d.ts.map +0 -1
  454. package/out/src/align.js.map +0 -1
  455. package/out/src/format.d.ts +0 -14
  456. package/out/src/format.d.ts.map +0 -1
  457. package/out/src/format.js +0 -14
  458. package/out/src/format.js.map +0 -1
  459. package/out/src/golden.d.ts.map +0 -1
  460. package/out/src/golden.js.map +0 -1
  461. package/out/src/license.d.ts.map +0 -1
  462. package/out/src/license.js.map +0 -1
  463. package/out/src/parquet.d.ts.map +0 -1
  464. package/out/src/parquet.js.map +0 -1
  465. package/out/src/split.d.ts.map +0 -1
  466. package/out/src/split.js.map +0 -1
  467. package/out/src/synthesize-anchor-absorption.d.ts.map +0 -1
  468. package/out/src/synthesize-anchor-absorption.js.map +0 -1
  469. package/out/src/synthesize-boundary-stress.d.ts.map +0 -1
  470. package/out/src/synthesize-boundary-stress.js.map +0 -1
  471. package/out/src/synthesize-german.d.ts.map +0 -1
  472. package/out/src/synthesize-german.js.map +0 -1
  473. package/out/src/synthesize-house-venue.d.ts.map +0 -1
  474. package/out/src/synthesize-house-venue.js.map +0 -1
  475. package/out/src/synthesize-intersection.d.ts.map +0 -1
  476. package/out/src/synthesize-intersection.js.map +0 -1
  477. package/out/src/synthesize-no-street.d.ts.map +0 -1
  478. package/out/src/synthesize-no-street.js.map +0 -1
  479. package/out/src/synthesize-po-box.d.ts.map +0 -1
  480. package/out/src/synthesize-po-box.js.map +0 -1
  481. package/out/src/synthesize-street.d.ts.map +0 -1
  482. package/out/src/synthesize-street.js.map +0 -1
  483. package/out/src/synthesize.d.ts.map +0 -1
  484. package/out/src/synthesize.js.map +0 -1
  485. package/out/src/tokenize.d.ts.map +0 -1
  486. package/out/src/tokenize.js.map +0 -1
  487. package/out/src/wof-json.d.ts.map +0 -1
  488. package/out/src/wof-json.js.map +0 -1
  489. package/src/format.ts +0 -14
  490. /package/out/src/{align.js → utils/align.js} +0 -0
  491. /package/out/src/{license.d.ts → utils/license.d.ts} +0 -0
  492. /package/out/src/{license.js → utils/license.js} +0 -0
  493. /package/out/src/{tokenize.d.ts → utils/tokenize.d.ts} +0 -0
  494. /package/out/src/{tokenize.js → utils/tokenize.js} +0 -0
  495. /package/src/{synthesize-anchor-absorption.ts → synthesizers/anchor-absorption.ts} +0 -0
  496. /package/src/{license.ts → utils/license.ts} +0 -0
  497. /package/src/{tokenize.ts → utils/tokenize.ts} +0 -0
@@ -8,6 +8,9 @@
8
8
  * canonical → `alignRow` → `LabeledRow` JSONL emit step. A recipe ({@link ShardRecipe}) supplies
9
9
  * only its synthesis + filter; the `mailwoman corpus shard <recipe>` command supplies the I/O.
10
10
  */
11
+ import type { PathBuilderLike } from "path-ts";
12
+ import type { AsyncChunkIterator, AsyncDataResource } from "spliterator";
13
+ import { AsyncSequence } from "spliterator";
11
14
  /**
12
15
  * {@link stableSourceID}, but accepting arbitrary disambiguator keys (e.g. a variant index `v`) that aren't
13
16
  * `ComponentTag`s. `stableSourceID` sorts + hashes EVERY key it's given, so passing extra keys is how the legacy
@@ -25,20 +28,42 @@ export interface ShardTuple {
25
28
  [k: string]: unknown;
26
29
  }
27
30
  /**
28
- * The deterministic LCG (`s = s*1664525 + 1013904223 mod 2^32`) the street/po-box/anchor builders used. A recipe whose
29
- * legacy `.mjs` seeded this must create it here so `--seed N` is byte-reproducible.
31
+ * The two seeded generators the legacy `build-*-shard` scripts used, re-exported from their home in
32
+ * `@mailwoman/core/utils` so a recipe keeps importing everything it needs from this one scaffold module.
33
+ *
34
+ * - `makeLcg` (`s = s*1664525 + 1013904223 mod 2^32`) — what the street/po-box/anchor builders seeded.
35
+ * - `makeMulberry32` — what the MAJORITY of them used (german, locale, boundary-stress, unit, fr-order, country-balanced,
36
+ * intersection, fr-admin-split, street-affix, street-bare, po-box-cedex).
37
+ *
38
+ * A recipe must seed the same one its `.mjs` did, the same way it did (usually `seed`, but some derive a per-stream
39
+ * seed), or `--seed N` stops being byte-reproducible.
30
40
  */
31
- export declare function makeLcg(seed: number): () => number;
41
+ export { makeLcg, mulberry32 as makeMulberry32, makeLcg as makeRandom } from "@mailwoman/core/utils";
32
42
  /**
33
- * Back-compat alias for {@link makeLcg}.
43
+ * One CSV record, keyed by its lower-cased header name, every value trimmed. A column the header does not declare reads
44
+ * as `undefined`; a column the header declares but this record does not reach reads as `""`.
34
45
  */
35
- export declare const makeRandom: typeof makeLcg;
46
+ export type CSVRecord = Record<string, string | undefined>;
36
47
  /**
37
- * Mulberry32 — the PRNG the MAJORITY of the legacy `build-*-shard` scripts used (german, locale, boundary-stress, unit,
38
- * fr-order, country-balanced, intersection, fr-admin-split, street-affix, street-bare, po-box-cedex). A recipe must
39
- * seed it EXACTLY as its `.mjs` did (usually `seed`, but some derive a per-stream seed) to stay byte-reproducible.
48
+ * Read a CSV as header-keyed records.
49
+ *
50
+ * Returns the spliterator's own {@linkcode AsyncSequence}, so a caller composes `take`/`drop`/`filter` onto it — those
51
+ * ops fuse into one pull loop, and a `take` that is satisfied closes the source's file handle on the way out. Wrapping
52
+ * this in an `async function*` costs an async frame per row and takes those ops away; don't.
53
+ *
54
+ * A source at or below the spliterator's 128 KiB bulk threshold is read whole and parsed by the synchronous engine, so
55
+ * this is also the right reader for small sources — there is no buffered variant to reach for.
40
56
  */
41
- export declare function makeMulberry32(seed: number): () => number;
57
+ export declare function readCSVRecords(source: AsyncDataResource | AsyncChunkIterator): AsyncSequence<CSVRecord>;
58
+ /**
59
+ * {@link readCSVRecords} over one member of a zip archive — what every recipe reading a cached OA source wants.
60
+ *
61
+ * A source a checkout has not cached yields nothing, after saying so. A lab holds the archives for the countries it has
62
+ * built, so a recipe naming ten sources routinely finds three, and the `unzip -p` subprocesses these replaced behaved
63
+ * the same way by accident — a non-zero exit warned and returned no rows. A recipe that ends up with NO tuples at all
64
+ * still throws; that is the case where the cache, not the recipe, is the problem.
65
+ */
66
+ export declare function readZippedCSVRecords(archivePath: PathBuilderLike, entryName: string): AsyncSequence<CSVRecord>;
42
67
  /**
43
68
  * Stream-parse a tuples JSONL file, yielding each parsed object (blank/invalid lines skipped).
44
69
  */
@@ -117,6 +142,30 @@ export interface ShardRecipeOpts {
117
142
  * `no-fragment` knob 3: minimum digit count for a number to count as "long" and be boosted. Default 3.
118
143
  */
119
144
  longNumberMinDigits?: number;
145
+ /**
146
+ * `sub-venue`: the sub-venue lexicon JSON. Default = the committed `corpus/data/sub-venue-lexicon.json`, resolved
147
+ * through the package manifest so it works from the source tree and from `out/`.
148
+ */
149
+ lexicon?: string;
150
+ /**
151
+ * `sub-venue`: directory of `sub-venue-extract` JSONLs, one per region. Default
152
+ * `$MAILWOMAN_DATA_ROOT/sub-venue/extracts`.
153
+ */
154
+ extractsDir?: string;
155
+ /**
156
+ * `sub-venue`: the `poi.db` spatial layer, read for the en-US and fr-FR venue + confound pools (the two of poi.db's
157
+ * four countries this shard has legs for). Default `$MAILWOMAN_DATA_ROOT/poi/poi.db`.
158
+ */
159
+ poiDb?: string;
160
+ /**
161
+ * `sub-venue`: GB/US/FR address-context tuples JSONL. Default the house-venue v3 tuples
162
+ * (`$MAILWOMAN_DATA_ROOT/corpus/intermediate/house-venue-tuples-v3.jsonl`); DE and ES read OpenAddresses directly.
163
+ */
164
+ subVenueTuples?: string;
165
+ /**
166
+ * `sub-venue`: share of emitted rows that are confound NEGATIVES. Default 0.3.
167
+ */
168
+ negativeFraction?: number;
120
169
  }
121
170
  /**
122
171
  * Tally a recipe returns.
@@ -1 +1 @@
1
- {"version":3,"file":"scaffold.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/scaffold.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAKH;;;;GAIG;AACH,wBAAgB,aAAa,CAAC,SAAS,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,SAAS,CAAC,GAAG,MAAM,CAElG;AAED;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,CAAC,CAAC,EAAE,MAAM,GAAG,OAAO,CAAA;CACpB;AAED;;;GAGG;AACH,wBAAgB,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,MAAM,CAQlD;AAED;;GAEG;AACH,eAAO,MAAM,UAAU,gBAAU,CAAA;AAEjC;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,MAAM,CAUzD;AAED;;GAEG;AACH,wBAAuB,UAAU,CAAC,KAAK,EAAE,MAAM,GAAG,cAAc,CAAC,UAAU,CAAC,CAgB3E;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IACjC,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC,OAAO,EAAE,MAAM,CAAA;IACf,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,MAAM,EAAE,MAAM,CAAA;IACd,SAAS,EAAE,MAAM,CAAA;IACjB,cAAc,CAAC,EAAE,MAAM,CAAA;IACvB,OAAO,CAAC,EAAE,MAAM,CAAA;CAChB;AAED;;;GAGG;AACH,wBAAgB,aAAa,CAC5B,KAAK,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,EAC7B,SAAS,EAAE,iBAAiB,EAC5B,WAAW,EAAE,MAAM,EACnB,WAAW,GAAE,MAAM,GAAG,IAAW,GAC/B,OAAO,CAOT;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC/B,MAAM,EAAE,MAAM,CAAA;IACd,IAAI,EAAE,MAAM,CAAA;IACZ,QAAQ,EAAE,MAAM,CAAA;IAChB,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,OAAO,CAAA;IAChB,UAAU,CAAC,EAAE,MAAM,CAAA;IAEnB,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,YAAY,CAAC,EAAE,MAAM,CAAA;IACrB;;OAEG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB;;;;;;OAMG;IACH,kBAAkB,CAAC,EAAE,OAAO,CAAA;IAC5B,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB;;OAEG;IACH,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB;;;OAGG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB;;OAEG;IACH,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB;;;OAGG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB;;OAEG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAA;CAC5B;AAED;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,OAAO,EAAE,MAAM,CAAA;IACf,OAAO,EAAE,MAAM,CAAA;IACf;;;;OAIG;IACH,YAAY,CAAC,EAAE,MAAM,CAAA;CACrB;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IACjC,IAAI,EAAE,MAAM,CAAA;IACZ,WAAW,EAAE,MAAM,CAAA;CACnB;AAED;;GAEG;AACH,MAAM,WAAW,WAAW;IAC3B;;OAEG;IACH,IAAI,EAAE,MAAM,CAAA;IACZ;;OAEG;IACH,WAAW,EAAE,MAAM,CAAA;IACnB;;OAEG;IACH,IAAI,EAAE,QAAQ,GAAG,UAAU,CAAA;IAC3B;;OAEG;IACH,OAAO,CAAC,EAAE,iBAAiB,EAAE,CAAA;IAC7B;;;OAGG;IACH,GAAG,CAAC,IAAI,EAAE,eAAe,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,CAAA;CAC9E"}
1
+ {"version":3,"file":"scaffold.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/scaffold.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAQH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAA;AAC9C,OAAO,KAAK,EAAE,kBAAkB,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAA;AACxE,OAAO,EAAE,aAAa,EAAkB,MAAM,aAAa,CAAA;AAE3D;;;;GAIG;AACH,wBAAgB,aAAa,CAAC,SAAS,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,SAAS,CAAC,GAAG,MAAM,CAElG;AAED;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,CAAC,CAAC,EAAE,MAAM,GAAG,OAAO,CAAA;CACpB;AAED;;;;;;;;;;GAUG;AACH,OAAO,EAAE,OAAO,EAAE,UAAU,IAAI,cAAc,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,uBAAuB,CAAA;AAEpG;;;GAGG;AACH,MAAM,MAAM,SAAS,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,SAAS,CAAC,CAAA;AA4C1D;;;;;;;;;GASG;AACH,wBAAgB,cAAc,CAAC,MAAM,EAAE,iBAAiB,GAAG,kBAAkB,GAAG,aAAa,CAAC,SAAS,CAAC,CAYvG;AAED;;;;;;;GAOG;AACH,wBAAgB,oBAAoB,CAAC,WAAW,EAAE,eAAe,EAAE,SAAS,EAAE,MAAM,GAAG,aAAa,CAAC,SAAS,CAAC,CAQ9G;AAED;;GAEG;AACH,wBAAuB,UAAU,CAAC,KAAK,EAAE,MAAM,GAAG,cAAc,CAAC,UAAU,CAAC,CAgB3E;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IACjC,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC,OAAO,EAAE,MAAM,CAAA;IACf,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,MAAM,EAAE,MAAM,CAAA;IACd,SAAS,EAAE,MAAM,CAAA;IACjB,cAAc,CAAC,EAAE,MAAM,CAAA;IACvB,OAAO,CAAC,EAAE,MAAM,CAAA;CAChB;AAED;;;GAGG;AACH,wBAAgB,aAAa,CAC5B,KAAK,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,EAC7B,SAAS,EAAE,iBAAiB,EAC5B,WAAW,EAAE,MAAM,EACnB,WAAW,GAAE,MAAM,GAAG,IAAW,GAC/B,OAAO,CAOT;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC/B,MAAM,EAAE,MAAM,CAAA;IACd,IAAI,EAAE,MAAM,CAAA;IACZ,QAAQ,EAAE,MAAM,CAAA;IAChB,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,OAAO,CAAA;IAChB,UAAU,CAAC,EAAE,MAAM,CAAA;IAEnB,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,YAAY,CAAC,EAAE,MAAM,CAAA;IACrB;;OAEG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB;;;;;;OAMG;IACH,kBAAkB,CAAC,EAAE,OAAO,CAAA;IAC5B,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB;;OAEG;IACH,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB;;;OAGG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB;;OAEG;IACH,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB;;;OAGG;IACH,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB;;OAEG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAA;IAC5B;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB;;;OAGG;IACH,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB;;;OAGG;IACH,KAAK,CAAC,EAAE,MAAM,CAAA;IACd;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAA;IACvB;;OAEG;IACH,gBAAgB,CAAC,EAAE,MAAM,CAAA;CACzB;AAED;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,OAAO,EAAE,MAAM,CAAA;IACf,OAAO,EAAE,MAAM,CAAA;IACf;;;;OAIG;IACH,YAAY,CAAC,EAAE,MAAM,CAAA;CACrB;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IACjC,IAAI,EAAE,MAAM,CAAA;IACZ,WAAW,EAAE,MAAM,CAAA;CACnB;AAED;;GAEG;AACH,MAAM,WAAW,WAAW;IAC3B;;OAEG;IACH,IAAI,EAAE,MAAM,CAAA;IACZ;;OAEG;IACH,WAAW,EAAE,MAAM,CAAA;IACnB;;OAEG;IACH,IAAI,EAAE,QAAQ,GAAG,UAAU,CAAA;IAC3B;;OAEG;IACH,OAAO,CAAC,EAAE,iBAAiB,EAAE,CAAA;IAC7B;;;OAGG;IACH,GAAG,CAAC,IAAI,EAAE,eAAe,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,CAAA;CAC9E"}
@@ -8,8 +8,12 @@
8
8
  * canonical → `alignRow` → `LabeledRow` JSONL emit step. A recipe ({@link ShardRecipe}) supplies
9
9
  * only its synthesis + filter; the `mailwoman corpus shard <recipe>` command supplies the I/O.
10
10
  */
11
- import { stableSourceID } from "../adapter.js";
12
- import { alignRow } from "../align.js";
11
+ import { existsSync } from "node:fs";
12
+ import { readZipEntry } from "@mailwoman/core/fs/zip";
13
+ import { tryParsingJSON } from "@mailwoman/core/objects";
14
+ import { stableSourceID } from "@mailwoman/corpus/adapters/utils";
15
+ import { alignRow } from "@mailwoman/corpus/utils";
16
+ import { AsyncSequence, CSVSpliterator } from "spliterator";
13
17
  /**
14
18
  * {@link stableSourceID}, but accepting arbitrary disambiguator keys (e.g. a variant index `v`) that aren't
15
19
  * `ComponentTag`s. `stableSourceID` sorts + hashes EVERY key it's given, so passing extra keys is how the legacy
@@ -19,50 +23,104 @@ export function shardSourceID(adapterID, parts) {
19
23
  return stableSourceID(adapterID, parts);
20
24
  }
21
25
  /**
22
- * The deterministic LCG (`s = s*1664525 + 1013904223 mod 2^32`) the street/po-box/anchor builders used. A recipe whose
23
- * legacy `.mjs` seeded this must create it here so `--seed N` is byte-reproducible.
26
+ * The two seeded generators the legacy `build-*-shard` scripts used, re-exported from their home in
27
+ * `@mailwoman/core/utils` so a recipe keeps importing everything it needs from this one scaffold module.
28
+ *
29
+ * - `makeLcg` (`s = s*1664525 + 1013904223 mod 2^32`) — what the street/po-box/anchor builders seeded.
30
+ * - `makeMulberry32` — what the MAJORITY of them used (german, locale, boundary-stress, unit, fr-order, country-balanced,
31
+ * intersection, fr-admin-split, street-affix, street-bare, po-box-cedex).
32
+ *
33
+ * A recipe must seed the same one its `.mjs` did, the same way it did (usually `seed`, but some derive a per-stream
34
+ * seed), or `--seed N` stops being byte-reproducible.
35
+ */
36
+ export { makeLcg, mulberry32 as makeMulberry32, makeLcg as makeRandom } from "@mailwoman/core/utils";
37
+ /**
38
+ * Quote handling spans the ROW split, not only the column split: a newline inside a quoted field belongs to a single
39
+ * record, so the record boundaries can only be found by a scanner that already knows where the quotes are. Find the
40
+ * lines first and unquote afterwards and such a record splits in two — the first half short by however many columns
41
+ * followed the newline, which is how a `street` value comes to be read as `city`.
42
+ *
43
+ * `header: false` keeps the first record in the stream, so this module owns the lower-casing rather than
44
+ * `normalizeKeys` — see {@link toHeader}.
24
45
  */
25
- export function makeLcg(seed) {
26
- let s = seed >>> 0;
27
- return () => {
28
- s = (s * 1_664_525 + 1_013_904_223) % 4_294_967_296;
29
- return s / 4_294_967_296;
30
- };
46
+ const CSV_INIT = { header: false, enableQuoteHandling: true };
47
+ /**
48
+ * Header names are lower-cased on the way in. `normalizeKeys` will not do it: it leaves an ALL CAPS header alone, and
49
+ * OpenAddresses ships `LON,LAT,NUMBER,STREET` while other extracts ship the same names lower-case. A recipe names its
50
+ * columns in lower case either way.
51
+ */
52
+ function toHeader(cells) {
53
+ return cells.map((name) => name.trim().toLowerCase());
31
54
  }
32
55
  /**
33
- * Back-compat alias for {@link makeLcg}.
56
+ * Line breaks inside a value become single spaces. A quote-aware parse is the first thing here able to return a value
57
+ * CONTAINING one — `us/ia/statewide.csv` has 12, all unit designators like `"#2\n#2"` — and every consumer synthesizes
58
+ * one-line address text from these cells with no guard, because until that parse landed no value could carry one.
59
+ * Collapsing keeps the record (the address is fine; the source's line break is not part of it) without emitting a
60
+ * training row with a newline inside it.
61
+ *
62
+ * Only `\r` and `\n`, deliberately — NOT `\s`. Runs of spaces and tabs pass through exactly as the source wrote them
63
+ * (OA's IA extract writes `NORTH`, three spaces, `MAIN STREET`), because those could always appear and every shard
64
+ * built to date contains them. Widening this to `\s+` silently rewrites values on rows with no line break at all.
65
+ * `scaffold.test.ts` pins both halves, against both readers.
34
66
  */
35
- export const makeRandom = makeLcg;
67
+ function toRecord(header, cells) {
68
+ const record = {};
69
+ for (let i = 0; i < header.length; i++) {
70
+ record[header[i]] = (cells[i] ?? "").replaceAll(/[\r\n]+/g, " ").trim();
71
+ }
72
+ return record;
73
+ }
36
74
  /**
37
- * Mulberry32 — the PRNG the MAJORITY of the legacy `build-*-shard` scripts used (german, locale, boundary-stress, unit,
38
- * fr-order, country-balanced, intersection, fr-admin-split, street-affix, street-bare, po-box-cedex). A recipe must
39
- * seed it EXACTLY as its `.mjs` did (usually `seed`, but some derive a per-stream seed) to stay byte-reproducible.
75
+ * Read a CSV as header-keyed records.
76
+ *
77
+ * Returns the spliterator's own {@linkcode AsyncSequence}, so a caller composes `take`/`drop`/`filter` onto it — those
78
+ * ops fuse into one pull loop, and a `take` that is satisfied closes the source's file handle on the way out. Wrapping
79
+ * this in an `async function*` costs an async frame per row and takes those ops away; don't.
80
+ *
81
+ * A source at or below the spliterator's 128 KiB bulk threshold is read whole and parsed by the synchronous engine, so
82
+ * this is also the right reader for small sources — there is no buffered variant to reach for.
40
83
  */
41
- export function makeMulberry32(seed) {
42
- let a = seed >>> 0;
43
- return () => {
44
- a = (a + 0x6d_2b_79_f5) | 0;
45
- let t = Math.imul(a ^ (a >>> 15), 1 | a);
46
- t = (t + Math.imul(t ^ (t >>> 7), 61 | t)) ^ t;
47
- return ((t ^ (t >>> 14)) >>> 0) / 4_294_967_296;
48
- };
84
+ export function readCSVRecords(source) {
85
+ let header = null;
86
+ return CSVSpliterator.fromAsync(source, CSV_INIT)
87
+ .filter((cells) => {
88
+ if (header)
89
+ return true;
90
+ header = toHeader(cells);
91
+ return false;
92
+ })
93
+ .map((cells) => toRecord(header, cells));
94
+ }
95
+ /**
96
+ * {@link readCSVRecords} over one member of a zip archive — what every recipe reading a cached OA source wants.
97
+ *
98
+ * A source a checkout has not cached yields nothing, after saying so. A lab holds the archives for the countries it has
99
+ * built, so a recipe naming ten sources routinely finds three, and the `unzip -p` subprocesses these replaced behaved
100
+ * the same way by accident — a non-zero exit warned and returned no rows. A recipe that ends up with NO tuples at all
101
+ * still throws; that is the case where the cache, not the recipe, is the problem.
102
+ */
103
+ export function readZippedCSVRecords(archivePath, entryName) {
104
+ if (!existsSync(String(archivePath))) {
105
+ console.error(` WARN: ${archivePath} is not cached — skipping ${entryName}`);
106
+ return AsyncSequence.from([]);
107
+ }
108
+ return readCSVRecords(readZipEntry(archivePath, entryName));
49
109
  }
50
110
  /**
51
111
  * Stream-parse a tuples JSONL file, yielding each parsed object (blank/invalid lines skipped).
52
112
  */
53
113
  export async function* readTuples(input) {
54
114
  // TextSpliterator (not JSONSpliterator) so a malformed line is SKIPPED, not thrown — the
55
- // per-line try/catch below is the tolerance this reader has always had.
115
+ // per-line tryParsingJSON below is the tolerance this reader has always had.
56
116
  const { TextSpliterator } = await import("spliterator");
57
117
  for await (const line of TextSpliterator.fromAsync(input)) {
58
118
  const trimmed = line.trim();
59
119
  if (!trimmed)
60
120
  continue;
61
- try {
62
- yield JSON.parse(trimmed);
63
- }
64
- catch {
65
- // skip malformed line
121
+ const tuple = tryParsingJSON(trimmed);
122
+ if (tuple !== null) {
123
+ yield tuple;
66
124
  }
67
125
  }
68
126
  }
@@ -1 +1 @@
1
- {"version":3,"file":"scaffold.js","sourceRoot":"","sources":["../../../src/shard-recipes/scaffold.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,eAAe,CAAA;AAC9C,OAAO,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAA;AAEtC;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAAC,SAAiB,EAAE,KAAyC;IACzF,OAAO,cAAc,CAAC,SAAS,EAAE,KAAwD,CAAC,CAAA;AAC3F,CAAC;AAaD;;;GAGG;AACH,MAAM,UAAU,OAAO,CAAC,IAAY;IACnC,IAAI,CAAC,GAAG,IAAI,KAAK,CAAC,CAAA;IAElB,OAAO,GAAG,EAAE;QACX,CAAC,GAAG,CAAC,CAAC,GAAG,SAAS,GAAG,aAAa,CAAC,GAAG,aAAa,CAAA;QAEnD,OAAO,CAAC,GAAG,aAAa,CAAA;IACzB,CAAC,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,UAAU,GAAG,OAAO,CAAA;AAEjC;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,IAAY;IAC1C,IAAI,CAAC,GAAG,IAAI,KAAK,CAAC,CAAA;IAElB,OAAO,GAAG,EAAE;QACX,CAAC,GAAG,CAAC,CAAC,GAAG,aAAa,CAAC,GAAG,CAAC,CAAA;QAC3B,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAA;QACxC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;QAE9C,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,aAAa,CAAA;IAChD,CAAC,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,SAAS,CAAC,CAAC,UAAU,CAAC,KAAa;IAC9C,yFAAyF;IACzF,wEAAwE;IACxE,MAAM,EAAE,eAAe,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,CAAA;IAEvD,IAAI,KAAK,EAAE,MAAM,IAAI,IAAI,eAAe,CAAC,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAA;QAE3B,IAAI,CAAC,OAAO;YAAE,SAAQ;QAEtB,IAAI,CAAC;YACJ,MAAM,IAAI,CAAC,KAAK,CAAC,OAAO,CAAe,CAAA;QACxC,CAAC;QAAC,MAAM,CAAC;YACR,sBAAsB;QACvB,CAAC;IACF,CAAC;AACF,CAAC;AAgBD;;;GAGG;AACH,MAAM,UAAU,aAAa,CAC5B,KAA6B,EAC7B,SAA4B,EAC5B,WAAmB,EACnB,cAA6B,IAAI;IAEjC,MAAM,OAAO,GAAG,QAAQ,CAAC,SAA2C,CAAC,CAAA;IAErE,IAAI,CAAC,OAAO,CAAC,GAAG;QAAE,OAAO,KAAK,CAAA;IAC9B,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,OAAO,CAAC,GAAG,EAAE,YAAY,EAAE,WAAW,EAAE,aAAa,EAAE,WAAW,EAAE,CAAC,GAAG,IAAI,CAAC,CAAA;IAEvG,OAAO,IAAI,CAAA;AACZ,CAAC"}
1
+ {"version":3,"file":"scaffold.js","sourceRoot":"","sources":["../../../src/shard-recipes/scaffold.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAA;AAEpC,OAAO,EAAE,YAAY,EAAE,MAAM,wBAAwB,CAAA;AACrD,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAA;AACxD,OAAO,EAAE,cAAc,EAAE,MAAM,kCAAkC,CAAA;AACjE,OAAO,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAA;AAGlD,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAE3D;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAAC,SAAiB,EAAE,KAAyC;IACzF,OAAO,cAAc,CAAC,SAAS,EAAE,KAAwD,CAAC,CAAA;AAC3F,CAAC;AAaD;;;;;;;;;;GAUG;AACH,OAAO,EAAE,OAAO,EAAE,UAAU,IAAI,cAAc,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,uBAAuB,CAAA;AAQpG;;;;;;;;GAQG;AACH,MAAM,QAAQ,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,mBAAmB,EAAE,IAAI,EAAW,CAAA;AAEtE;;;;GAIG;AACH,SAAS,QAAQ,CAAC,KAAwB;IACzC,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,CAAA;AACtD,CAAC;AAED;;;;;;;;;;;GAWG;AACH,SAAS,QAAQ,CAAC,MAAyB,EAAE,KAAwB;IACpE,MAAM,MAAM,GAAc,EAAE,CAAA;IAE5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,UAAU,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IACzE,CAAC;IAED,OAAO,MAAM,CAAA;AACd,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,cAAc,CAAC,MAA8C;IAC5E,IAAI,MAAM,GAAoB,IAAI,CAAA;IAElC,OAAO,cAAc,CAAC,SAAS,CAAW,MAAM,EAAE,QAAQ,CAAC;SACzD,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE;QACjB,IAAI,MAAM;YAAE,OAAO,IAAI,CAAA;QAEvB,MAAM,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAA;QAExB,OAAO,KAAK,CAAA;IACb,CAAC,CAAC;SACD,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAO,EAAE,KAAK,CAAC,CAAC,CAAA;AAC3C,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,oBAAoB,CAAC,WAA4B,EAAE,SAAiB;IACnF,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,EAAE,CAAC;QACtC,OAAO,CAAC,KAAK,CAAC,WAAW,WAAW,6BAA6B,SAAS,EAAE,CAAC,CAAA;QAE7E,OAAO,aAAa,CAAC,IAAI,CAAY,EAAE,CAAC,CAAA;IACzC,CAAC;IAED,OAAO,cAAc,CAAC,YAAY,CAAC,WAAW,EAAE,SAAS,CAAC,CAAC,CAAA;AAC5D,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,SAAS,CAAC,CAAC,UAAU,CAAC,KAAa;IAC9C,yFAAyF;IACzF,6EAA6E;IAC7E,MAAM,EAAE,eAAe,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,CAAA;IAEvD,IAAI,KAAK,EAAE,MAAM,IAAI,IAAI,eAAe,CAAC,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAA;QAE3B,IAAI,CAAC,OAAO;YAAE,SAAQ;QAEtB,MAAM,KAAK,GAAG,cAAc,CAAa,OAAO,CAAC,CAAA;QAEjD,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;YACpB,MAAM,KAAK,CAAA;QACZ,CAAC;IACF,CAAC;AACF,CAAC;AAgBD;;;GAGG;AACH,MAAM,UAAU,aAAa,CAC5B,KAA6B,EAC7B,SAA4B,EAC5B,WAAmB,EACnB,cAA6B,IAAI;IAEjC,MAAM,OAAO,GAAG,QAAQ,CAAC,SAA2C,CAAC,CAAA;IAErE,IAAI,CAAC,OAAO,CAAC,GAAG;QAAE,OAAO,KAAK,CAAA;IAC9B,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,OAAO,CAAC,GAAG,EAAE,YAAY,EAAE,WAAW,EAAE,aAAa,EAAE,WAAW,EAAE,CAAC,GAAG,IAAI,CAAC,CAAA;IAEvG,OAAO,IAAI,CAAA;AACZ,CAAC"}
@@ -26,10 +26,64 @@
26
26
  * to keep the postcode-ORDER distribution multi-locale so a US-heavy affix shard doesn't dilute
27
27
  * FR/DE postcode (the v0.9.8 blemish).
28
28
  */
29
+ import type { ComponentTag } from "@mailwoman/core/types";
29
30
  import { type ShardRecipe } from "./scaffold.ts";
31
+ /**
32
+ * A real US skeleton tuple read from a cached OA zip.
33
+ */
34
+ interface USTuple {
35
+ house_number: string;
36
+ street: string;
37
+ locality: string;
38
+ region: string;
39
+ postcode: string;
40
+ base_source_id: string;
41
+ }
42
+ export type SuffixBoundaryClass = "terminal-only" | "terminal-contrast";
43
+ /**
44
+ * Classify a real street surface for #1569. `terminal-only` has an ambiguous suffix-eligible name word immediately
45
+ * before a different terminal type (`Blue Hill Rd`); `terminal-contrast` ends at the ambiguous word itself (`Sutton
46
+ * Hollow`). The contrast check intentionally wins when both final words are name-prone: only the LAST token is the
47
+ * suffix under the canonical rule.
48
+ */
49
+ export declare function classifySuffixBoundaryStreet(street: string): SuffixBoundaryClass | null;
50
+ /**
51
+ * Layout-shell options for {@link renderRow}. `cuts` are the cumulative random() cutoffs for [full, bare, street-only] —
52
+ * the remainder is the venue shell. Defaults reproduce the original street-affix distribution (40/25/20/15) with the
53
+ * six template venues.
54
+ */
55
+ interface RenderRowOpts {
56
+ venues?: readonly string[];
57
+ cuts?: readonly [number, number, number];
58
+ }
59
+ /**
60
+ * Embed the rendered street in a RANDOM layout so the model recognizes affixes wherever the street sits: full address,
61
+ * bare house+street, street-only (pure affix parse), or venue-prefixed.
62
+ */
63
+ export declare function renderRow(random: () => number, base: USTuple, street: string, streetComponents: Partial<Record<ComponentTag, string>>, opts?: RenderRowOpts): {
64
+ fmt: string;
65
+ raw: string;
66
+ components: Partial<Record<ComponentTag, string>>;
67
+ };
30
68
  /**
31
69
  * Shard recipe registered with the corpus builder — see the file header for the parse behaviour it exists to exercise,
32
70
  * and `description` below for the surface form it generates.
33
71
  */
34
72
  export declare const streetAffixRecipe: ShardRecipe;
73
+ /**
74
+ * #1569 root-fix shard. Both classes come from real non-Vermont OA streets and use the affix recipe's existing layout
75
+ * diversity. v4.3.1 makes terminal-only 80% of the mix: the first 40/60 run moved a 100-row TRAIN sample only 4→11
76
+ * while contrast was already 95/100 before training (93/100 after). Post-run audit found that the global affix relabel
77
+ * pass corrupts many already-decomposed target rows into double suffixes; do not retrain this recipe until relabel is
78
+ * idempotent over a decomposed street family. The 20% contrast leg remains explicit, additive to the already-strong
79
+ * base distribution, and B2 still gates it unchanged.
80
+ *
81
+ * V2 (corpus 0.19.0, 2026-08-10 recipe review): the v4.3.3 board split (rich venue-led rows 5/43 vs bare 53/65) showed
82
+ * the model separating template rows from real ones, and the venue shell was the giveaway — six fixed venue strings. v2
83
+ * draws the venue shell from thousands of REAL HRSA facility names and raises its share (venue 30%, full 35%, bare 20%,
84
+ * street-only 15%). Dose policy moved to the recipe's config side: weight ≤4 effective passes per run (Muennighoff
85
+ * 2023's repetition knee) and the source is excluded from the augmentation pool — see the v4.4.0 config.
86
+ */
87
+ export declare const suffixBoundaryRecipe: ShardRecipe;
88
+ export {};
35
89
  //# sourceMappingURL=street-affix.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"street-affix.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/street-affix.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAkBH,OAAO,EAAkB,KAAK,WAAW,EAAE,MAAM,eAAe,CAAA;AA6XhE;;;GAGG;AACH,eAAO,MAAM,iBAAiB,EAAE,WAiM/B,CAAA"}
1
+ {"version":3,"file":"street-affix.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/street-affix.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAWH,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAA;AAOzD,OAAO,EAAuE,KAAK,WAAW,EAAE,MAAM,eAAe,CAAA;AAuFrH;;GAEG;AACH,UAAU,OAAO;IAChB,YAAY,EAAE,MAAM,CAAA;IACpB,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,EAAE,MAAM,CAAA;IAChB,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,EAAE,MAAM,CAAA;IAChB,cAAc,EAAE,MAAM,CAAA;CACtB;AAmGD,MAAM,MAAM,mBAAmB,GAAG,eAAe,GAAG,mBAAmB,CAAA;AAEvE;;;;;GAKG;AACH,wBAAgB,4BAA4B,CAAC,MAAM,EAAE,MAAM,GAAG,mBAAmB,GAAG,IAAI,CAevF;AAsFD;;;;GAIG;AACH,UAAU,aAAa;IACtB,MAAM,CAAC,EAAE,SAAS,MAAM,EAAE,CAAA;IAC1B,IAAI,CAAC,EAAE,SAAS,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,CAAA;CACxC;AAED;;;GAGG;AACH,wBAAgB,SAAS,CACxB,MAAM,EAAE,MAAM,MAAM,EACpB,IAAI,EAAE,OAAO,EACb,MAAM,EAAE,MAAM,EACd,gBAAgB,EAAE,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,EACvD,IAAI,GAAE,aAAkB,GACtB;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,GAAG,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAA;CAAE,CA8BjF;AAyDD;;;GAGG;AACH,eAAO,MAAM,iBAAiB,EAAE,WAiM/B,CAAA;AAED;;;;;;;;;;;;;GAaG;AACH,eAAO,MAAM,oBAAoB,EAAE,WA0JlC,CAAA"}