v1.16.0

changelog.md

@@ -1,5 +1,28 @@
 # Changelog
 
+## 2026-01-20 - 1.16.0 - feat(invoices)
+add line_items extraction and normalization for invoice parsing
+
+- Introduce ILineItem interface and add line_items array to IInvoice.
+- Add extractLineItems helper to normalize item fields (position, product, description, quantity, unit_price, total_price).
+- Include line_items in parsed invoice output and sample JSON in test, defaulting to [] when absent.
+- Update logging to include extracted line item count.
+- Clarify test instructions to extract items from invoice tables and skip subtotal/total rows.
+
+## 2026-01-20 - 1.15.3 - fix(tests(nanonets))
+allow '/' when normalizing invoice strings in tests
+
+- Adjust regex in test/test.invoices.nanonets.ts to preserve forward slashes when cleaning invoice values
+- Changed pattern from [^A-Z0-9-] to [^A-Z0-9\/-] to prevent accidental removal of '/' characters in invoice identifiers
+
+## 2026-01-20 - 1.15.2 - fix(dev-deps)
+bump devDependencies @push.rocks/smartagent to ^1.6.2 and @push.rocks/smartai to ^0.13.3
+
+- Bumped @push.rocks/smartagent from ^1.5.4 to ^1.6.2 in devDependencies
+- Bumped @push.rocks/smartai from ^0.13.2 to ^0.13.3 in devDependencies
+- Updated test/test.invoices.nanonets.ts JSON extraction prompt: instruct not to omit special characters in invoice_number and to use the json validate tool
+- No breaking changes; only dev dependency updates and test prompt adjustments
+
 ## 2026-01-20 - 1.15.1 - fix(tests)
 enable progress events in invoice tests and bump @push.rocks/smartagent devDependency to ^1.5.4
 

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@host.today/ht-docker-ai",
-  "version": "1.15.1",
+  "version": "1.16.0",
   "type": "module",
   "private": false,
   "description": "Docker images for AI vision-language models including MiniCPM-V 4.5",
@@ -15,8 +15,8 @@
   "devDependencies": {
     "@git.zone/tsrun": "^2.0.1",
     "@git.zone/tstest": "^3.1.5",
-    "@push.rocks/smartagent": "^1.5.4",
-    "@push.rocks/smartai": "^0.13.2"
+    "@push.rocks/smartagent": "^1.6.2",
+    "@push.rocks/smartai": "^0.13.3"
   },
   "repository": {
     "type": "git",

pnpm-lock.yaml

@@ -19,11 +19,11 @@ importers:
         specifier: ^3.1.5
         version: 3.1.6(socks@2.8.7)(typescript@5.9.3)
       '@push.rocks/smartagent':
-        specifier: ^1.5.4
-        version: 1.5.4(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)
+        specifier: ^1.6.2
+        version: 1.6.2(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)
       '@push.rocks/smartai':
-        specifier: ^0.13.2
-        version: 0.13.2(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)
+        specifier: ^0.13.3
+        version: 0.13.3(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)
 
 packages:
 
@@ -868,11 +868,11 @@ packages:
   '@push.rocks/qenv@6.1.3':
     resolution: {integrity: sha512-+z2hsAU/7CIgpYLFqvda8cn9rUBMHqLdQLjsFfRn5jPoD7dJ5rFlpkbhfM4Ws8mHMniwWaxGKo+q/YBhtzRBLg==}
 
-  '@push.rocks/smartagent@1.5.4':
-    resolution: {integrity: sha512-W5DoG0MUY6oAvxMcMC0M0ZmMmjs6zs8WcDnz5jVcg74vD3oaSguJjcokjB4L4Tv1dzEhq0ArDOnaC2C0Lf3z7w==}
+  '@push.rocks/smartagent@1.6.2':
+    resolution: {integrity: sha512-JaYZ7tRbmS0fVrF73Z+RF9plJ/Va0H+81zvEACT8YZRf+WhhIT+P7kKh7IcTRrgudtA7aw6eVXUmOGCMeszm3Q==}
 
-  '@push.rocks/smartai@0.13.2':
-    resolution: {integrity: sha512-FqnHh31tU0Nkr/g25UMAjjE7gZVN8cuioRlSq1xo19rS9kyMiux+UzpylO2tgXF5S+lTsw5cGtfP0BUfrxlTGg==}
+  '@push.rocks/smartai@0.13.3':
+    resolution: {integrity: sha512-VDZzHs101hpGMmUaectuLfcME4kHpuOS7o5ffuGk5lYl383foyAN71+5v441jpk/gLDNf2KhDACR/d2O4n90Ag==}
 
   '@push.rocks/smartarchive@5.2.1':
     resolution: {integrity: sha512-TNv5q6QuBRX7jrzffiyb6A8AALNAr0kyAcJswa0l3ahBP1Q6zszNo9xOVXmW2gKX2KShtO/Y+Cn0i46n8lbnaQ==}
@@ -5206,9 +5206,9 @@ snapshots:
       '@push.rocks/smartlog': 3.1.10
       '@push.rocks/smartpath': 6.0.0
 
-  '@push.rocks/smartagent@1.5.4(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)':
+  '@push.rocks/smartagent@1.6.2(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)':
     dependencies:
-      '@push.rocks/smartai': 0.13.2(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)
+      '@push.rocks/smartai': 0.13.3(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)
       '@push.rocks/smartbrowser': 2.0.8(typescript@5.9.3)
       '@push.rocks/smartdeno': 1.2.0
       '@push.rocks/smartfs': 1.3.1
@@ -5230,7 +5230,7 @@ snapshots:
       - ws
       - zod
 
-  '@push.rocks/smartai@0.13.2(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)':
+  '@push.rocks/smartai@0.13.3(typescript@5.9.3)(ws@8.19.0)(zod@3.25.76)':
     dependencies:
       '@anthropic-ai/sdk': 0.71.2(zod@3.25.76)
       '@mistralai/mistralai': 1.12.0

test/test.invoices.nanonets.ts

@@ -42,6 +42,15 @@ const smartAi = new SmartAi({
 // DualAgentOrchestrator for structured task execution
 let orchestrator: DualAgentOrchestrator;
 
+interface ILineItem {
+  position: number;
+  product: string;
+  description: string;
+  quantity: number;
+  unit_price: number;
+  total_price: number;
+}
+
 interface IInvoice {
   invoice_number: string;
   invoice_date: string;
@@ -50,6 +59,7 @@ interface IInvoice {
   net_amount: number;
   vat_amount: number;
   total_amount: number;
+  line_items: ILineItem[];
 }
 
 interface IImageData {
@@ -78,8 +88,9 @@ Page numbers should be wrapped in brackets. Ex: <page_number>14</page_number>.`;
 const JSON_EXTRACTION_PROMPT = `Extract key fields from the invoice. Return ONLY valid JSON.
 
 WHERE TO FIND DATA:
-- invoice_number, invoice_date, vendor_name: Look in the HEADER section at the TOP of PAGE 1 (near "Invoice no.", "Invoice date:", "Rechnungsnummer"). Use common sense. Btw. an invoice number might start on INV* .
+- invoice_number, invoice_date, vendor_name: Look in the HEADER section at the TOP of PAGE 1 (near "Invoice no.", "Invoice date:", "Rechnungsnummer"). Use common sense. Btw. an invoice number might start on INV* . Also be sure to not omit special chars like / - and sp on. They are part of the invoice number.
 - net_amount, vat_amount, total_amount: Look in the SUMMARY section at the BOTTOM (look for "Total", "Amount due", "Gesamtbetrag")
+- line_items: Look in the TABLE(s) with columns like Pos, Product, Description, Quantity, Unit Price, Price
 
 RULES:
 1. Use common sense.
@@ -89,11 +100,23 @@ RULES:
 5. net_amount: Total before tax
 6. vat_amount: Tax amount
 7. total_amount: Final total with tax
+8. line_items: Array of items from the invoice table. Skip subtotal/total rows.
 
-JSON only:
-{"invoice_number":"X","invoice_date":"YYYY-MM-DD","vendor_name":"X","currency":"EUR","net_amount":0,"vat_amount":0,"total_amount":0}
+JSON format:
+{
+  "invoice_number": "X",
+  "invoice_date": "YYYY-MM-DD",
+  "vendor_name": "X",
+  "currency": "EUR",
+  "net_amount": 0,
+  "vat_amount": 0,
+  "total_amount": 0,
+  "line_items": [
+    {"position": 1, "product": "X", "description": "X", "quantity": 1, "unit_price": 0, "total_price": 0}
+  ]
+}
 
-Double check for valid JSON syntax.
+Double check for valid JSON syntax. use the json validate tool.
 
 `;
 
@@ -310,7 +333,7 @@ function extractInvoiceNumber(s: string | undefined): string {
     const match = clean.match(pattern);
     if (match) return match[1];
   }
-  return clean.replace(/[^A-Z0-9-]/gi, '').trim() || clean;
+  return clean.replace(/[^A-Z0-9\/-]/gi, '').trim() || clean;
 }
 
 /**
@@ -340,6 +363,21 @@ function extractCurrency(s: string | undefined): string {
   return 'EUR';
 }
 
+/**
+ * Extract and normalize line items array
+ */
+function extractLineItems(items: unknown): ILineItem[] {
+  if (!Array.isArray(items)) return [];
+  return items.map((item: Record<string, unknown>, index: number) => ({
+    position: typeof item.position === 'number' ? item.position : index + 1,
+    product: String(item.product || '').trim(),
+    description: String(item.description || '').trim(),
+    quantity: parseAmount(item.quantity as string | number) || 1,
+    unit_price: parseAmount(item.unit_price as string | number),
+    total_price: parseAmount(item.total_price as string | number),
+  }));
+}
+
 /**
  * Try to extract valid JSON from a response string
  */
@@ -448,6 +486,7 @@ ${JSON_EXTRACTION_PROMPT}`;
       net_amount: parseAmount(jsonData.net_amount as string | number),
       vat_amount: parseAmount(jsonData.vat_amount as string | number),
       total_amount: parseAmount(jsonData.total_amount as string | number),
+      line_items: extractLineItems(jsonData.line_items),
     };
   } catch (error) {
     const elapsed = ((Date.now() - startTime) / 1000).toFixed(1);
@@ -471,6 +510,7 @@ async function extractInvoice(markdown: string, docName: string): Promise<IInvoi
       net_amount: 0,
       vat_amount: 0,
       total_amount: 0,
+      line_items: [],
     };
   }
   console.log(`    [${docName}] Extracted: ${invoice.invoice_number}`);
@@ -703,7 +743,7 @@ for (const tc of testCases) {
     const elapsedMs = Date.now() - startTime;
     processingTimes.push(elapsedMs);
 
-    console.log(`    Extracted: ${extracted.invoice_number} | ${extracted.invoice_date} | ${extracted.total_amount} ${extracted.currency}`);
+    console.log(`    Extracted: ${extracted.invoice_number} | ${extracted.invoice_date} | ${extracted.total_amount} ${extracted.currency} | ${extracted.line_items.length} items`);
 
     const result = compareInvoice(extracted, expected);