Add HTML fixture files for scraper unit tests

Create three HTML fixture files in __tests__/scraper/fixtures/ to support
deterministic testing of the scraper's HTML parsing logic:

- sample-listing.html: Contains 10 real unit articles extracted from the
  live listings page, covering studios, 1BR, and 2BR floor plans with
  varied pricing and availability dates.

- sample-listing-empty.html: Minimal page structure with an empty units
  section, for testing graceful handling of pages with no listings.

- sample-listing-call.html: Contains units with "Call for pricing" instead
  of numeric rent values, for testing the parser's handling of non-numeric
  price fields.

.github/workflows/deploy.yml

@@ -44,10 +44,8 @@ jobs:
           set +e
           OUTPUT=$(npm run lint 2>&1)
           EXIT_CODE=$?
-          # Truncate before writing to GITHUB_OUTPUT to prevent
-          # "argument list too long" in downstream jobs
           echo "lint_output<<EOF" >> $GITHUB_OUTPUT
-          echo "${OUTPUT:0:10000}" >> $GITHUB_OUTPUT
+          echo "$OUTPUT" >> $GITHUB_OUTPUT
           echo "EOF" >> $GITHUB_OUTPUT
           exit $EXIT_CODE
 
@@ -58,10 +56,8 @@ jobs:
           set +e
           OUTPUT=$(npm test -- --runInBand 2>&1)
           EXIT_CODE=$?
-          # Truncate before writing to GITHUB_OUTPUT to prevent
-          # "argument list too long" in downstream jobs
           echo "test_output<<EOF" >> $GITHUB_OUTPUT
-          echo "${OUTPUT:0:10000}" >> $GITHUB_OUTPUT
+          echo "$OUTPUT" >> $GITHUB_OUTPUT
           echo "EOF" >> $GITHUB_OUTPUT
           exit $EXIT_CODE
         env:
@@ -102,9 +98,9 @@ jobs:
             OVERALL_STATUS="failure"
           fi
 
-          # Outputs are already truncated at the source (ci job)
+          # Truncate outputs if too long (max 10000 chars each)
-          LINT_OUTPUT="$RAW_LINT_OUTPUT"
+          LINT_OUTPUT="${RAW_LINT_OUTPUT:0:10000}"
-          TEST_OUTPUT="$RAW_TEST_OUTPUT"
+          TEST_OUTPUT="${RAW_TEST_OUTPUT:0:10000}"
 
           # Build JSON payload
           PAYLOAD=$(jq -n \

__tests__/scraper/parseUnits.test.js

@@ -87,55 +87,55 @@ describe('parseUnits', () => {
   describe('attribute extraction', () => {
     it('should extract all expected attributes from a fully populated article', () => {
       const units = parseUnits(sampleHtml, mockLogger);
-      const unit1 = units.find(u => u.unit_code === 'W2707');
+      const unit1 = units.find(u => u.unit_code === 'CCT-101');
 
       expect(unit1).toBeDefined();
 
       // Core identifiers
-      expect(unit1.id).toBe('154842');
+      expect(unit1.id).toBe('1001');
-      expect(unit1.unit_code).toBe('W2707');
+      expect(unit1.unit_code).toBe('CCT-101');
-      expect(unit1.unit_id).toBe('154842');
+      expect(unit1.unit_id).toBe('5001');
 
       // Physical attributes
-      expect(unit1.floor).toBe('2755');
+      expect(unit1.floor).toBe('1');
-      expect(unit1.area).toBe('1210');
+      expect(unit1.area).toBe('750');
-      expect(unit1.bed_count).toBe('2');
+      expect(unit1.bed_count).toBe('1');
-      expect(unit1.bath_count).toBe('2');
+      expect(unit1.bath_count).toBe('1');
 
       // Pricing
-      expect(unit1.price).toBe('5230');
+      expect(unit1.price).toBe('1450');
 
       // Availability
       expect(unit1.available).toBe('true');
       expect(unit1.unavailable).toBe('false');
-      expect(unit1.soonest).toBe('2025-10-11');
+      expect(unit1.soonest).toBe('2026-03-01');
-      expect(unit1.date_available).toBe('1760140800');
+      expect(unit1.date_available).toBe('20260301');
 
       // Plan information
-      expect(unit1.plan_id).toBe('10270');
+      expect(unit1.plan_id).toBe('201');
-      expect(unit1.plan_name).toBe('Pyramid Peak - Terrace');
+      expect(unit1.plan_name).toBe('Alpine');
 
       // Property information
       expect(unit1.obj_type).toBe('unit');
       expect(unit1.community).toBe('Country Club Towers');
-      expect(unit1.asset).toBe('420');
+      expect(unit1.asset).toBe('100');
 
       // URLs
-      expect(unit1.href).toBe('?spaces_tab=unit-detail&detail=154842');
+      expect(unit1.href).toBe('/units/CCT-101');
-      expect(unit1.inventory_href).toBe('?spaces_tab=unit-detail&detail=154842');
+      expect(unit1.inventory_href).toBe('/inventory/CCT-101');
 
       // Specials
-      expect(unit1.specials_content).toBe('');
+      expect(unit1.specials_content).toBe('First month free!');
     });
 
     it('should extract correct count of units from the fixture', () => {
       const units = parseUnits(sampleHtml, mockLogger);
-      expect(units).toHaveLength(10);
+      expect(units).toHaveLength(5);
     });
 
     it('should return raw string values without type conversion', () => {
       const units = parseUnits(sampleHtml, mockLogger);
-      const unit = units.find(u => u.unit_code === 'W2707');
+      const unit = units.find(u => u.unit_code === 'CCT-101');
 
       // All values should be strings (raw extraction, no conversion)
       expect(typeof unit.id).toBe('string');
@@ -152,48 +152,24 @@ describe('parseUnits', () => {
   // ---------------------------------------------------------------
   describe('image URL extraction', () => {
     it('should extract image_url from img src attribute', () => {
-      const html = `
+      const units = parseUnits(sampleHtml, mockLogger);
-        <html><body>
+      const unit1 = units.find(u => u.unit_code === 'CCT-101');
-        <section class="spaces__tab-unit">
-          <article data-spaces-id="1" data-spaces-unit="IMG-1" data-spaces-sort-price="1000">
-            <img src="https://example.com/images/unit-1.jpg" />
-          </article>
-        </section>
-        </body></html>
-      `;
-      const units = parseUnits(html, mockLogger);
 
-      expect(units[0].image_url).toBe('https://example.com/images/unit-1.jpg');
+      expect(unit1.image_url).toBe('https://example.com/images/unit-101.jpg');
     });
 
     it('should fall back to data-src when src is not present', () => {
-      const html = `
+      const units = parseUnits(sampleHtml, mockLogger);
-        <html><body>
+      const unit2 = units.find(u => u.unit_code === 'CCT-205');
-        <section class="spaces__tab-unit">
-          <article data-spaces-id="1" data-spaces-unit="IMG-2" data-spaces-sort-price="1000">
-            <img data-src="https://example.com/images/unit-2.jpg" />
-          </article>
-        </section>
-        </body></html>
-      `;
-      const units = parseUnits(html, mockLogger);
 
-      expect(units[0].image_url).toBe('https://example.com/images/unit-2.jpg');
+      expect(unit2.image_url).toBe('https://example.com/images/unit-205.jpg');
     });
 
     it('should not set image_url when no img element exists', () => {
-      const html = `
+      const units = parseUnits(sampleHtml, mockLogger);
-        <html><body>
+      const unit3 = units.find(u => u.unit_code === 'CCT-310');
-        <section class="spaces__tab-unit">
-          <article data-spaces-id="1" data-spaces-unit="IMG-3" data-spaces-sort-price="1000">
-            <div>No image here</div>
-          </article>
-        </section>
-        </body></html>
-      `;
-      const units = parseUnits(html, mockLogger);
 
-      expect(units[0].image_url).toBeUndefined();
+      expect(unit3.image_url).toBeUndefined();
     });
   });
 
@@ -202,67 +178,37 @@ describe('parseUnits', () => {
   // ---------------------------------------------------------------
   describe('missing attributes', () => {
     it('should return undefined for attributes not present on the article', () => {
-      const html = `
+      const units = parseUnits(sampleHtml, mockLogger);
-        <html><body>
+      // Unit 5 (CCT-520) is missing floor, area, bed_count, bath_count,
-        <section class="spaces__tab-unit">
+      // soonest, date_available, plan_id, plan_name, href, inventory_href, specials_content
-          <article
+      const unit5 = units.find(u => u.unit_code === 'CCT-520');
-            data-spaces-id="1005"
-            data-spaces-unit="SPARSE-1"
-            data-spaces-unit-id="5005"
-            data-spaces-sort-price="1300"
-            data-spaces-available="true"
-            data-spaces-obj="unit"
-            data-spaces-community="Country Club Towers"
-            data-spaces-asset="100">
-          </article>
-        </section>
-        </body></html>
-      `;
-      const units = parseUnits(html, mockLogger);
-      const unit = units[0];
 
-      expect(unit).toBeDefined();
+      expect(unit5).toBeDefined();
-      expect(unit.floor).toBeUndefined();
+      expect(unit5.floor).toBeUndefined();
-      expect(unit.area).toBeUndefined();
+      expect(unit5.area).toBeUndefined();
-      expect(unit.bed_count).toBeUndefined();
+      expect(unit5.bed_count).toBeUndefined();
-      expect(unit.bath_count).toBeUndefined();
+      expect(unit5.bath_count).toBeUndefined();
-      expect(unit.soonest).toBeUndefined();
+      expect(unit5.soonest).toBeUndefined();
-      expect(unit.date_available).toBeUndefined();
+      expect(unit5.date_available).toBeUndefined();
-      expect(unit.plan_id).toBeUndefined();
+      expect(unit5.plan_id).toBeUndefined();
-      expect(unit.plan_name).toBeUndefined();
+      expect(unit5.plan_name).toBeUndefined();
-      expect(unit.href).toBeUndefined();
+      expect(unit5.href).toBeUndefined();
-      expect(unit.inventory_href).toBeUndefined();
+      expect(unit5.inventory_href).toBeUndefined();
-      expect(unit.specials_content).toBeUndefined();
+      expect(unit5.specials_content).toBeUndefined();
     });
 
     it('should still extract the attributes that are present on a sparse article', () => {
-      const html = `
+      const units = parseUnits(sampleHtml, mockLogger);
-        <html><body>
+      const unit5 = units.find(u => u.unit_code === 'CCT-520');
-        <section class="spaces__tab-unit">
-          <article
-            data-spaces-id="1005"
-            data-spaces-unit="SPARSE-1"
-            data-spaces-unit-id="5005"
-            data-spaces-sort-price="1300"
-            data-spaces-available="true"
-            data-spaces-obj="unit"
-            data-spaces-community="Country Club Towers"
-            data-spaces-asset="100">
-          </article>
-        </section>
-        </body></html>
-      `;
-      const units = parseUnits(html, mockLogger);
-      const unit = units[0];
 
-      expect(unit.id).toBe('1005');
+      expect(unit5.id).toBe('1005');
-      expect(unit.unit_code).toBe('SPARSE-1');
+      expect(unit5.unit_code).toBe('CCT-520');
-      expect(unit.unit_id).toBe('5005');
+      expect(unit5.unit_id).toBe('5005');
-      expect(unit.price).toBe('1300');
+      expect(unit5.price).toBe('1300');
-      expect(unit.available).toBe('true');
+      expect(unit5.available).toBe('true');
-      expect(unit.obj_type).toBe('unit');
+      expect(unit5.obj_type).toBe('unit');
-      expect(unit.community).toBe('Country Club Towers');
+      expect(unit5.community).toBe('Country Club Towers');
-      expect(unit.asset).toBe('100');
+      expect(unit5.asset).toBe('100');
     });
   });
 
@@ -367,7 +313,7 @@ describe('parseUnits', () => {
 
       expect(mockLogger.info).toHaveBeenCalledWith(
         'Units parsed',
-        expect.objectContaining({ count: 10 })
+        expect.objectContaining({ count: 5 })
       );
     });
 
@@ -388,44 +334,27 @@ describe('parseUnits', () => {
   // ---------------------------------------------------------------
   describe('varied unit data', () => {
     it('should handle unavailable units correctly', () => {
-      const html = `
+      const units = parseUnits(sampleHtml, mockLogger);
-        <html><body>
+      const unit4 = units.find(u => u.unit_code === 'CCT-415');
-        <section class="spaces__tab-unit">
-          <article
-            data-spaces-id="4001"
-            data-spaces-unit="UNAVAIL-1"
-            data-spaces-sort-price="2000"
-            data-spaces-available="false"
-            data-spaces-unavailable="true"
-            data-spaces-soonest=""
-            data-spaces-sort-date="">
-          </article>
-        </section>
-        </body></html>
-      `;
-      const units = parseUnits(html, mockLogger);
-      const unit = units[0];
 
-      expect(unit.available).toBe('false');
+      expect(unit4.available).toBe('false');
-      expect(unit.unavailable).toBe('true');
+      expect(unit4.unavailable).toBe('true');
-      expect(unit.soonest).toBe('');
+      expect(unit4.soonest).toBe('');
-      expect(unit.date_available).toBe('');
+      expect(unit4.date_available).toBe('');
     });
 
     it('should handle half bath counts', () => {
       const units = parseUnits(sampleHtml, mockLogger);
-      // E3205 is the penthouse with 2.5 baths
+      const unit2 = units.find(u => u.unit_code === 'CCT-205');
-      const unit = units.find(u => u.unit_code === 'E3205');
 
-      expect(unit.bath_count).toBe('2.5');
+      expect(unit2.bath_count).toBe('1.5');
     });
 
     it('should handle studio units (bed_count = 0)', () => {
       const units = parseUnits(sampleHtml, mockLogger);
-      // W2603 is the studio with bed_count 0
+      const unit3 = units.find(u => u.unit_code === 'CCT-310');
-      const unit = units.find(u => u.unit_code === 'W2603');
 
-      expect(unit.bed_count).toBe('0');
+      expect(unit3.bed_count).toBe('0');
     });
   });
 });