SCRAPE-5: Implement convertDataTypes() with type parsing helpers (#10)

Co-authored-by: Stephen Minakian <stephenminakian@gmail.com>
Co-committed-by: Stephen Minakian <stephenminakian@gmail.com>

services/scraperService.js

@@ -198,10 +198,189 @@ function parseUnits(html, logger) {
   return deduplicated;
 }
 
+// ============================================================
+// Data Type Conversion Helpers
+// ============================================================
+
+/**
+ * Parse string to integer, return null for invalid
+ * @param {*} value - Value to parse
+ * @returns {number|null} Parsed integer or null
+ */
+function parseInteger(value) {
+  if (value === null || value === undefined || value === '') {
+    return null;
+  }
+  const parsed = parseInt(value, 10);
+  return Number.isNaN(parsed) || !Number.isFinite(parsed) ? null : parsed;
+}
+
+/**
+ * Parse string to positive integer, return null for 0 or invalid.
+ * Used for fields like area where 0 means "not available".
+ * @param {*} value - Value to parse
+ * @returns {number|null} Parsed positive integer or null
+ */
+function parsePositiveInteger(value) {
+  const parsed = parseInteger(value);
+  return parsed === 0 ? null : parsed;
+}
+
+/**
+ * Parse value that could be integer or string identifier.
+ * Returns integer if cleanly parseable, otherwise trimmed string.
+ * @param {*} value - Value to parse
+ * @returns {number|string|null} Parsed integer, trimmed string, or null
+ */
+function parseIntegerOrString(value) {
+  if (value === null || value === undefined || value === '') {
+    return null;
+  }
+  const parsed = parseInt(value, 10);
+  // If it parses cleanly to an integer, return integer
+  if (!Number.isNaN(parsed) && String(parsed) === String(value).trim()) {
+    return parsed;
+  }
+  // Otherwise return as trimmed string
+  return String(value).trim();
+}
+
+/**
+ * Parse price, stripping non-numeric characters.
+ * Handles "$1,234" format and "Call for pricing" text.
+ * @param {*} value - Value to parse
+ * @returns {number|null} Parsed price or null
+ */
+function parsePrice(value) {
+  if (value === null || value === undefined || value === '') {
+    return null;
+  }
+
+  // Check for "Call for pricing" or similar text
+  if (typeof value === 'string' && /call|contact|inquire/i.test(value)) {
+    return null;
+  }
+
+  // Strip non-numeric characters except decimal point
+  const cleaned = String(value).replace(/[^0-9.]/g, '');
+  const parsed = parseInt(cleaned, 10);
+
+  if (Number.isNaN(parsed) || !Number.isFinite(parsed)) {
+    return null;
+  }
+
+  // Negative prices are invalid
+  if (parsed < 0) {
+    return null;
+  }
+
+  return parsed;
+}
+
+/**
+ * Parse boolean from string.
+ * Handles "true"/"false"/"1"/"0" and actual boolean values.
+ * @param {*} value - Value to parse
+ * @returns {boolean|null} Parsed boolean or null
+ */
+function parseBoolean(value) {
+  if (value === null || value === undefined || value === '') {
+    return null;
+  }
+  if (typeof value === 'boolean') {
+    return value;
+  }
+  const str = String(value).toLowerCase().trim();
+  if (str === 'true' || str === '1') return true;
+  if (str === 'false' || str === '0') return false;
+  return null;
+}
+
+/**
+ * Parse float value.
+ * Named parseFloatValue to avoid shadowing the global parseFloat.
+ * @param {*} value - Value to parse
+ * @returns {number|null} Parsed float or null
+ */
+function parseFloatValue(value) {
+  if (value === null || value === undefined || value === '') {
+    return null;
+  }
+  const parsed = Number.parseFloat(value);
+  return Number.isNaN(parsed) || !Number.isFinite(parsed) ? null : parsed;
+}
+
+/**
+ * Trim string, return null for empty.
+ * Converts non-string values to string before trimming.
+ * @param {*} value - Value to trim
+ * @returns {string|null} Trimmed string or null
+ */
+function trimString(value) {
+  if (value === null || value === undefined) {
+    return null;
+  }
+  const trimmed = String(value).trim();
+  return trimmed === '' ? null : trimmed;
+}
+
+// ============================================================
+// Main Data Type Conversion Function
+// ============================================================
+
+/**
+ * Convert unit data types from strings to proper types.
+ * Applies the appropriate parser to each field based on its expected type.
+ * @param {Object} unit - Raw unit object with string values
+ * @returns {Object} Unit object with converted types
+ */
+function convertDataTypes(unit) {
+  return {
+    // Integer fields
+    id: parseInteger(unit.id),
+    unit_id: parseIntegerOrString(unit.unit_id),
+    floor: parseInteger(unit.floor),
+    area: parsePositiveInteger(unit.area),
+    bed_count: parseInteger(unit.bed_count),
+    plan_id: parseInteger(unit.plan_id),
+    asset: parseInteger(unit.asset),
+    date_available: parseInteger(unit.date_available),
+
+    // Float fields
+    bath_count: parseFloatValue(unit.bath_count),
+
+    // Price field (special handling)
+    price: parsePrice(unit.price),
+
+    // Boolean fields
+    available: parseBoolean(unit.available),
+    unavailable: parseBoolean(unit.unavailable),
+
+    // String fields (trim and preserve)
+    unit_code: trimString(unit.unit_code),
+    plan_name: trimString(unit.plan_name),
+    soonest: trimString(unit.soonest),
+    obj_type: trimString(unit.obj_type),
+    community: trimString(unit.community),
+    href: trimString(unit.href),
+    inventory_href: trimString(unit.inventory_href),
+    image_url: trimString(unit.image_url),
+    specials_content: trimString(unit.specials_content)
+  };
+}
+
 module.exports = {
   fetchPage,
   parseUnits,
-  // Export for testing
+  convertDataTypes,
+  // Export helpers for testing
+  parseInteger,
+  parsePositiveInteger,
+  parseIntegerOrString,
+  parsePrice,
+  parseBoolean,
+  parseFloatValue,
+  trimString,
   isRetryableError,
   sleep
 };