universal-parcel-scraper 0.735.2-main.3769 → 0.736.0-main.3771

package/dist/carriers/hermes-de/adapter.js

@@ -72,27 +72,30 @@ const DEPOT_PLACES = [
 ];
 const COUNTRY_PLACES = [
     /\barrived in the destination country (.+?)(?: and is prepared\b|\.?$)/i,
-    /\bprepared for further shipping \(([^()]+)\)/i,
+    /\bfor further shipping \(([^()]+)\)/i,
 ];
 // The town after the postcode closing a ParcelShop's address, before the
 // collection deadline the ready-for-pickup wording appends.
 const SHOP_TOWN = /,\s*\d{5}\s+([^\d,]+?)(?:\s+max\.\s.*|\s+abgegeben)?\.?\s*$/;
 /**
  * Where a row's wording puts the parcel. Hermes's own depots and ParcelShops
- * are in Germany; a country abroad is named alone.
+ * are in Germany. A partner's scan (`INTERNATIONAL_…`) names the country after
+ * its depot, "prepared in EXEMPLEVILLE for further shipping (France)",
+ * or the country alone.
  */
 function wordedPlace(text, code) {
     const place = (pattern) => clean(pattern.exec(text)?.[1], 120);
+    const country = COUNTRY_PLACES.map(place).find(Boolean) ?? '';
+    const partner = code.startsWith('INTERNATIONAL_');
     for (const pattern of DEPOT_PLACES) {
         const depot = place(pattern);
+        if (depot && partner)
+            return country ? `${depot}, ${country}` : depot;
         if (depot)
             return `${depot}, Germany`;
     }
-    for (const pattern of COUNTRY_PLACES) {
-        const country = place(pattern);
-        if (country)
-            return country;
-    }
+    if (country)
+        return country;
     const town = code.startsWith('PARCELSHOP_') ? place(SHOP_TOWN) : '';
     return town ? `${town}, Germany` : null;
 }

package/dist/core/adapter/track.js

@@ -4,6 +4,7 @@ import { normalizeCarrierResult } from '../result/index.js';
 import { runSteps } from '../runner/index.js';
 import { NOOP_RECORDER } from '../telemetry/index.js';
 import { dateLocalClocks } from '../../places/clocks.js';
+import { restoreLostLetters } from '../../places/index.js';
 /**
  * One carrier lookup, leaving provider selection and scheduling to the consumer.
  * The adapter gets the normalized number, as from `createTracker()`: a number
@@ -25,7 +26,7 @@ export async function trackCarrier(carrier, typed, options) {
                 : await runSteps({ carrier, budgetMs: options.budgetMs ?? 120_000, signal: options.signal,
                     recorder: options.recorder ?? NOOP_RECORDER }, [{ id: registered.steps[0] ?? 'direct',
                         run: ({ signal, remainingMs }) => registered.track(input, ...(options.signal || options.budgetMs !== undefined ? [{ signal, budgetMs: context.budgetMs === undefined ? undefined : Math.min(context.budgetMs, remainingMs) }] : [])) }]);
-            return dateLocalClocks(carrier, normalizeCarrierResult(result));
+            return dateLocalClocks(carrier, restoreLostLetters(normalizeCarrierResult(result)));
         }
         catch (error) {
             // A caller's cancellation is the caller's reason, not a carrier failure.

package/dist/places/index.d.ts

@@ -1,3 +1,4 @@
+import type { CarrierResult } from '../core/result/index.js';
 export { carrierHomeZone, dateLocalClocks, type ZoneBasis } from './clocks.js';
 /** Where a scan happened, as precisely as its free-text location allows. */
 export interface EventPlace {
@@ -47,6 +48,11 @@ export declare function placeTimeZone(place: EventPlace): string | null;
  * place we can be confident about. A wrong dot is worse than no dot.
  */
 export declare function locatePlace(location: string | null | undefined, hints?: PlaceHints): EventPlace | null;
+/**
+ * The result with the letters its scans' "Town, Country" places lost restored
+ * (`restoredTownName`), in the place and wherever the wording names it.
+ */
+export declare function restoreLostLetters(result: CarrierResult): CarrierResult;
 /**
  * Places for a parcel's scans, oldest first. A scan that names no country
  * borrows one from the scans around it, then the destination and the

package/dist/places/index.js

@@ -1,7 +1,7 @@
 import { readFileSync } from 'node:fs';
 import { brotliDecompressSync } from 'node:zlib';
 import { CARRIER_DEFINITIONS } from '../core/catalog/definitions.js';
-import { timeZoneCountry } from '../core/time/index.js';
+import { countryCode, timeZoneCountry } from '../core/time/index.js';
 import { ambiguousAddressCodes, militaryPostOffice, trackingPlace } from './trackingLocation.js';
 import facilityList from './facilities.json' with { type: 'json' };
 import hubList from './hubs.json' with { type: 'json' };
@@ -511,6 +511,70 @@ export function locatePlace(location, hints = {}) {
     // Alone, a name two towns of a country share alike says neither.
     return found && !found.twins ? found.place : null;
 }
+const LOST = '\uFFFD';
+/**
+ * A town's name with the letters a carrier lost on the way, each a replacement
+ * character, filled from the one spelling of a town of `country` that fits:
+ * "SAINT-\uFFFDTIENNE" is Saint-Étienne's. A letter inside a word takes the
+ * case of the letters beside it, one starting a word the town's own. Null when
+ * no letter is lost or no single spelling fits.
+ */
+function restoredTownName(town, country) {
+    if (!town.includes(LOST))
+        return null;
+    const chars = [...town];
+    const source = chars.map((char) => char === LOST ? '(\\p{L})' : /[\s-]/u.test(char) ? '[\\s-]'
+        : char.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')).join('');
+    const pattern = new RegExp(`^${source}$`, 'iu');
+    const data = gazetteer();
+    const fills = new Set();
+    for (const index of countryPlaces(country)) {
+        const match = pattern.exec(data.names[index]);
+        if (match)
+            fills.add(match.slice(1).join(' '));
+    }
+    if (new Set([...fills].map((fill) => fill.toLowerCase())).size !== 1)
+        return null;
+    const letters = [...fills][0].split(' ');
+    const letter = (char) => char !== undefined && char !== LOST && /\p{L}/u.test(char);
+    return chars.map((char, at) => {
+        if (char !== LOST)
+            return char;
+        const filled = letters.shift();
+        if (!letter(chars[at - 1]))
+            return filled;
+        const beside = [chars[at - 1], chars[at + 1]].filter(letter);
+        return beside.every((other) => other === other.toUpperCase()) ? filled.toUpperCase() : filled.toLowerCase();
+    }).join('');
+}
+/**
+ * The result with the letters its scans' "Town, Country" places lost restored
+ * (`restoredTownName`), in the place and wherever the wording names it.
+ */
+export function restoreLostLetters(result) {
+    if (!result.events?.some((event) => event.location?.includes(LOST)))
+        return result;
+    const restored = new Map();
+    for (const event of result.events) {
+        const [, town, country] = /^(.+), ([^,]+)$/.exec(event.location ?? '') ?? [];
+        const code = country ? countryCode(country) : null;
+        const name = town && code && !restored.has(town) ? restoredTownName(town, code) : null;
+        if (name)
+            restored.set(town, name);
+    }
+    if (!restored.size)
+        return result;
+    const spell = (text) => [...restored].reduce((spelled, [lost, name]) => spelled.replaceAll(lost, name), text);
+    return {
+        ...result,
+        ...(result.last_status_text ? { last_status_text: spell(result.last_status_text) } : {}),
+        events: result.events.map((event) => ({
+            ...event,
+            ...(event.location ? { location: spell(event.location) } : {}),
+            ...(event.description ? { description: spell(event.description) } : {}),
+        })),
+    };
+}
 /** `split` reads a country the text names apart from its town; without it the whole text is the town's. */
 function locate(location, hints, split = true) {
     const text = location?.trim();

package/dist/server/openapi.json

@@ -2,7 +2,7 @@
   "openapi": "3.1.0",
   "info": {
     "title": "Universal Parcel Scraper",
-    "version": "0.735.2"
+    "version": "0.736.0"
   },
   "paths": {
     "/health": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "universal-parcel-scraper",
-  "version": "0.735.2-main.3769",
+  "version": "0.736.0-main.3771",
   "description": "Self-hosted parcel tracking: carrier detection, dedicated scrapers and optional universal providers.",
   "type": "module",
   "license": "Apache-2.0",