Make listing discovery unit-aware

skills/web-automation/scripts/real-estate-address.js

@@ -0,0 +1,125 @@
+#!/usr/bin/env node
+
+const STREET_STOP_WORDS = new Set([
+  "st",
+  "street",
+  "rd",
+  "road",
+  "dr",
+  "drive",
+  "ave",
+  "avenue",
+  "blvd",
+  "boulevard",
+  "ln",
+  "lane",
+  "ct",
+  "court",
+  "cir",
+  "circle",
+  "way",
+  "trl",
+  "trail",
+  "pkwy",
+  "parkway",
+  "tx"
+]);
+
+const UNIT_LABEL_PATTERN = "(?:apt|apartment|unit|suite|ste|#)";
+
+function tokenize(value) {
+  return String(value || "")
+    .toLowerCase()
+    .replace(/[^a-z0-9#\s-]/g, " ")
+    .split(/[\s-]+/)
+    .map((token) => token.trim())
+    .filter(Boolean);
+}
+
+function collapseWhitespace(value) {
+  return String(value || "").replace(/\s+/g, " ").trim();
+}
+
+function escapeRegex(value) {
+  return value.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+}
+
+export function parseAddressIdentity(rawAddress) {
+  const address = collapseWhitespace(rawAddress);
+  if (!address) {
+    throw new Error("Missing address.");
+  }
+
+  const [streetPartRaw, ...restParts] = address.split(",");
+  const streetPart = collapseWhitespace(streetPartRaw);
+  const locality = collapseWhitespace(restParts.join(" "));
+
+  const unitMatch = streetPart.match(
+    new RegExp(`^(.*?)(?:\\s+(?:${UNIT_LABEL_PATTERN})\\s*([a-z0-9-]+))$`, "i")
+  );
+
+  const streetWithoutUnit = collapseWhitespace(unitMatch ? unitMatch[1] : streetPart);
+  const unitValue = unitMatch ? unitMatch[2].toLowerCase() : null;
+
+  const streetTokens = tokenize(streetWithoutUnit).filter(
+    (token) => !STREET_STOP_WORDS.has(token)
+  );
+  const localityTokens = tokenize(locality).filter((token) => !STREET_STOP_WORDS.has(token));
+
+  return {
+    raw: address,
+    streetPart,
+    streetWithoutUnit,
+    locality,
+    streetTokens,
+    localityTokens,
+    streetNumber: streetTokens[0] || null,
+    unitValue,
+    hasUnit: Boolean(unitValue)
+  };
+}
+
+export function buildZillowAddressSlug(rawAddress) {
+  const address = collapseWhitespace(rawAddress);
+  return address
+    .replace(/,/g, "")
+    .replace(/#/g, "")
+    .trim()
+    .split(/\s+/)
+    .join("-");
+}
+
+export function scoreAddressCandidate(identity, candidateText) {
+  const normalized = collapseWhitespace(candidateText).toLowerCase();
+  const normalizedPadded = ` ${normalized.replace(/[^a-z0-9#]+/g, " ")} `;
+  const allTokens = [...identity.streetTokens, ...identity.localityTokens];
+  const uniqueTokens = Array.from(new Set(allTokens));
+  const matchedTokens = uniqueTokens.filter((token) =>
+    normalizedPadded.includes(` ${token} `)
+  );
+
+  let unitMatched = true;
+  if (identity.hasUnit && identity.unitValue) {
+    const unitRegex = new RegExp(
+      `(?:^|\\s)(?:${UNIT_LABEL_PATTERN})\\s*${escapeRegex(identity.unitValue)}(?:\\s|$)`,
+      "i"
+    );
+    const hashRegex = new RegExp(`#\\s*${escapeRegex(identity.unitValue)}(?:\\b|$)`, "i");
+    const looseTokenRegex = new RegExp(`(?:^|\\s)${escapeRegex(identity.unitValue)}(?:\\s|$)`, "i");
+    unitMatched =
+      unitRegex.test(normalized) ||
+      hashRegex.test(candidateText) ||
+      looseTokenRegex.test(normalizedPadded);
+  }
+
+  const minimumCoreMatches = identity.hasUnit ? 3 : 2;
+  const matched =
+    matchedTokens.length >= Math.min(minimumCoreMatches, uniqueTokens.length) && unitMatched;
+
+  return {
+    matched,
+    score: matchedTokens.length + (unitMatched && identity.hasUnit ? 2 : 0),
+    matchedTokens,
+    unitMatched
+  };
+}