fix(berths): CM-2 — robust purchase-price extraction (clean-token + magnitude floor)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

src/lib/services/berth-pdf-parser.ts

@@ -357,10 +357,14 @@ export function extractFromOcrText(rawText: string): {
     };
   }
 
-  // Purchase price: "PURCHASE PRICE:\nFEE SIMPLE OR STRATA LOT\n3,880,800 USD"
-  const priceMatch = text.match(/PURCHASE\s+PRICE[\s\S]{0,80}?([0-9][0-9,]+)\s*USD/i);
-  if (priceMatch) {
-    out.price = { value: Number(priceMatch[1]!.replace(/,/g, '')), confidence: 0.7, engine: 'ocr' };
+  // Purchase price: the single clean comma-grouped currency figure. The rates
+  // on the same sheet are letter-spaced (garble) and below the floor, so they
+  // never collide with the main price. See extractPurchasePrice().
+  const priceResult = extractPurchasePrice(text);
+  if (priceResult.value != null) {
+    out.price = { value: priceResult.value, confidence: priceResult.confidence, engine: 'ocr' };
+  } else if (priceResult.warning) {
+    warnings.push(priceResult.warning);
   }
 
   // Pricing validity: "ALL PRICES ABOVE ARE CONFIRMED THROUGH UNTIL SEPTEMBER 15TH, 2025"
@@ -507,6 +511,62 @@ function coerceFieldValue(key: keyof ExtractedBerthFields, raw: string): string
   return numeric < 0 ? null : numeric;
 }
 
+/**
+ * Floor that separates a 6–7-figure purchase price from the ≤~12k weekly/daily
+ * lease rates printed on the same sheet. Observed prices: 277,200 … 5,433,120;
+ * observed weekly highs ≤ 11,341. A wide-margin separator.
+ */
+export const PURCHASE_PRICE_FLOOR = 50_000;
+
+/**
+ * Strict clean comma-grouped currency token. On the real spec sheets the
+ * purchase price is the one figure rendered WITHOUT letter-spacing (the large
+ * bold number); the weekly/daily rates ARE letter-spaced and garble in text
+ * extraction, so they never match this pattern. The floor is a second guard
+ * for clean/synthetic PDFs where rates would also extract cleanly.
+ */
+const PRICE_TOKEN_RE = /\b(\d{1,3}(?:,\d{3})+)\s?(USD|EUR|GBP)\b/gi;
+
+/**
+ * Extract the single main purchase price from raw PDF text. Returns
+ * `value: null` (with a warning) when zero above-floor tokens are found, or
+ * when two or more DISTINCT above-floor values appear (genuinely ambiguous —
+ * flag for human review rather than guess).
+ */
+export function extractPurchasePrice(rawText: string): {
+  value: number | null;
+  currency: string | null;
+  confidence: number;
+  warning?: string;
+} {
+  const candidates: Array<{ value: number; currency: string }> = [];
+  for (const m of rawText.matchAll(PRICE_TOKEN_RE)) {
+    const value = Number(m[1]!.replace(/,/g, ''));
+    if (Number.isFinite(value) && value >= PURCHASE_PRICE_FLOOR) {
+      candidates.push({ value, currency: m[2]!.toUpperCase() });
+    }
+  }
+  if (candidates.length === 0) {
+    return {
+      value: null,
+      currency: null,
+      confidence: 0,
+      warning: 'No purchase-price token found (no clean figure ≥ floor).',
+    };
+  }
+  const distinct = [...new Set(candidates.map((c) => c.value))];
+  if (distinct.length > 1) {
+    return {
+      value: null,
+      currency: null,
+      confidence: 0,
+      warning: `Multiple purchase-price candidates (${distinct.join(', ')}) — needs review.`,
+    };
+  }
+  const best = candidates[0]!;
+  return { value: best.value, currency: best.currency, confidence: 0.95 };
+}
+
 /** Parse a human date like "September 15 2025" → "2025-09-15". */
 export function parseHumanDate(raw: string): string | null {
   const cleaned = raw.replace(/(\d+)(st|nd|rd|th)/i, '$1').trim();

tests/unit/services/berth-pdf-parser.test.ts

@@ -13,9 +13,11 @@ import { describe, expect, it } from 'vitest';
 
 import {
   extractFromOcrText,
+  extractPurchasePrice,
   isPdfMagic,
   parseFeetInches,
   parseHumanDate,
+  PURCHASE_PRICE_FLOOR,
   shouldOfferAiTier,
 } from '@/lib/services/berth-pdf-parser';
 
@@ -191,3 +193,43 @@ describe('shouldOfferAiTier', () => {
     ).toBe(false);
   });
 });
+
+describe('extractPurchasePrice', () => {
+  it('isolates the single clean main price among letter-spaced rate garble', () => {
+    // Real-sheet shape: rates are letter-spaced (so they never match the strict
+    // token); the main price renders clean.
+    const text =
+      'W E E K H I G H / LO W : 1 1 , 3 4 1 U S D / 8 , 1 0 0 U S D 3,880,800 USD ' +
+      'DAY H I G H / LO W : 1 , 8 9 0 U S D / 1 , 3 5 0 U S D';
+    const r = extractPurchasePrice(text);
+    expect(r.value).toBe(3880800);
+    expect(r.currency).toBe('USD');
+    expect(r.confidence).toBeGreaterThanOrEqual(0.9);
+  });
+
+  it('excludes clean rate tokens below the floor (synthetic clean sheet)', () => {
+    const text = '3,880,800 USD WEEK HIGH / LOW: 11,341 USD / 8,100 USD';
+    expect(extractPurchasePrice(text).value).toBe(3880800);
+  });
+
+  it('returns null + warning when no price-magnitude token is present', () => {
+    const r = extractPurchasePrice('no prices here, just 12 USD of nothing');
+    expect(r.value).toBeNull();
+    expect(r.warning).toMatch(/no purchase-price/i);
+  });
+
+  it('flags ambiguity when two DISTINCT above-floor tokens appear', () => {
+    const r = extractPurchasePrice('3,880,800 USD and also 1,247,400 USD');
+    expect(r.value).toBeNull();
+    expect(r.warning).toMatch(/multiple/i);
+  });
+
+  it('treats a repeated identical price as unambiguous', () => {
+    const r = extractPurchasePrice('720,720 USD ... header ... 720,720 USD');
+    expect(r.value).toBe(720720);
+  });
+
+  it('exposes the floor constant', () => {
+    expect(PURCHASE_PRICE_FLOOR).toBe(50_000);
+  });
+});