completed local setup with compose

schemas/kg_schema.json

@@ -0,0 +1,202 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "title": "Tax Knowledge Graph Schema",
+  "definitions": {
+    "temporal_properties": {
+      "type": "object",
+      "properties": {
+        "valid_from": { "type": "string", "format": "date-time" },
+        "valid_to": { "type": "string", "format": "date-time" },
+        "asserted_at": { "type": "string", "format": "date-time" },
+        "retracted_at": { "type": ["string", "null"], "format": "date-time" },
+        "source": { "type": "string" },
+        "extractor_version": { "type": "string" }
+      },
+      "required": ["valid_from", "asserted_at", "source", "extractor_version"]
+    },
+    "provenance": {
+      "type": "object",
+      "properties": {
+        "doc_id": { "type": "string" },
+        "page": { "type": "integer", "minimum": 1 },
+        "bbox": {
+          "type": "object",
+          "properties": {
+            "x": { "type": "number" },
+            "y": { "type": "number" },
+            "width": { "type": "number" },
+            "height": { "type": "number" }
+          },
+          "required": ["x", "y", "width", "height"]
+        },
+        "text_hash": { "type": "string" },
+        "ocr_confidence": { "type": "number", "minimum": 0, "maximum": 1 }
+      },
+      "required": ["doc_id", "page", "text_hash"]
+    }
+  },
+  "oneOf": [
+    {
+      "title": "TaxpayerProfile",
+      "type": "object",
+      "properties": {
+        "node_type": { "const": "TaxpayerProfile" },
+        "taxpayer_id": { "type": "string" },
+        "type": { "enum": ["Individual", "Partnership", "Company"] },
+        "residence": { "type": "string" },
+        "contact": {
+          "type": "object",
+          "properties": {
+            "email": { "type": "string", "format": "email" },
+            "phone": { "type": "string" },
+            "address": { "type": "string" }
+          }
+        },
+        "tax_years": { "type": "array", "items": { "type": "string" } },
+        "utr": { "type": "string", "pattern": "^[0-9]{10}$" },
+        "ni_number": {
+          "type": "string",
+          "pattern": "^[A-CEGHJ-PR-TW-Z]{2}\\d{6}[A-D]$"
+        }
+      },
+      "allOf": [{ "$ref": "#/definitions/temporal_properties" }],
+      "required": ["node_type", "taxpayer_id", "type"]
+    },
+    {
+      "title": "TaxYear",
+      "type": "object",
+      "properties": {
+        "node_type": { "const": "TaxYear" },
+        "label": { "type": "string" },
+        "start_date": { "type": "string", "format": "date" },
+        "end_date": { "type": "string", "format": "date" },
+        "jurisdiction_ref": { "type": "string" }
+      },
+      "allOf": [{ "$ref": "#/definitions/temporal_properties" }],
+      "required": [
+        "node_type",
+        "label",
+        "start_date",
+        "end_date",
+        "jurisdiction_ref"
+      ]
+    },
+    {
+      "title": "Document",
+      "type": "object",
+      "properties": {
+        "node_type": { "const": "Document" },
+        "doc_id": { "type": "string" },
+        "kind": {
+          "enum": [
+            "bank_statement",
+            "invoice",
+            "receipt",
+            "p_and_l",
+            "balance_sheet",
+            "payslip",
+            "dividend_voucher",
+            "property_statement",
+            "prior_return",
+            "letter",
+            "certificate"
+          ]
+        },
+        "source": { "type": "string" },
+        "mime": { "type": "string" },
+        "date_range": {
+          "type": "object",
+          "properties": {
+            "start": { "type": "string", "format": "date" },
+            "end": { "type": "string", "format": "date" }
+          }
+        },
+        "checksum": { "type": "string" },
+        "file_size": { "type": "integer" },
+        "pages": { "type": "integer", "minimum": 1 }
+      },
+      "allOf": [{ "$ref": "#/definitions/temporal_properties" }],
+      "required": ["node_type", "doc_id", "kind", "source", "checksum"]
+    },
+    {
+      "title": "Evidence",
+      "type": "object",
+      "properties": {
+        "node_type": { "const": "Evidence" },
+        "snippet_id": { "type": "string" },
+        "doc_ref": { "type": "string" },
+        "page": { "type": "integer", "minimum": 1 },
+        "bbox": {
+          "type": "object",
+          "properties": {
+            "x": { "type": "number" },
+            "y": { "type": "number" },
+            "width": { "type": "number" },
+            "height": { "type": "number" }
+          },
+          "required": ["x", "y", "width", "height"]
+        },
+        "text_hash": { "type": "string" },
+        "ocr_confidence": { "type": "number", "minimum": 0, "maximum": 1 },
+        "extracted_text": { "type": "string" }
+      },
+      "allOf": [{ "$ref": "#/definitions/temporal_properties" }],
+      "required": [
+        "node_type",
+        "snippet_id",
+        "doc_ref",
+        "page",
+        "bbox",
+        "text_hash"
+      ]
+    },
+    {
+      "title": "IncomeItem",
+      "type": "object",
+      "properties": {
+        "node_type": { "const": "IncomeItem" },
+        "type": {
+          "enum": [
+            "employment",
+            "self_employment",
+            "property",
+            "dividend",
+            "interest",
+            "other"
+          ]
+        },
+        "gross": { "type": "number" },
+        "net": { "type": "number" },
+        "tax_withheld": { "type": "number" },
+        "period_start": { "type": "string", "format": "date" },
+        "period_end": { "type": "string", "format": "date" },
+        "currency": { "type": "string", "pattern": "^[A-Z]{3}$" },
+        "description": { "type": "string" }
+      },
+      "allOf": [
+        { "$ref": "#/definitions/temporal_properties" },
+        { "$ref": "#/definitions/provenance" }
+      ],
+      "required": ["node_type", "type", "gross", "currency"]
+    },
+    {
+      "title": "ExpenseItem",
+      "type": "object",
+      "properties": {
+        "node_type": { "const": "ExpenseItem" },
+        "type": { "enum": ["business", "property", "capital", "personal"] },
+        "amount": { "type": "number" },
+        "category": { "type": "string" },
+        "capitalizable_flag": { "type": "boolean" },
+        "currency": { "type": "string", "pattern": "^[A-Z]{3}$" },
+        "description": { "type": "string" },
+        "allowable": { "type": "boolean" }
+      },
+      "allOf": [
+        { "$ref": "#/definitions/temporal_properties" },
+        { "$ref": "#/definitions/provenance" }
+      ],
+      "required": ["node_type", "type", "amount", "currency"]
+    }
+  ]
+}