feat: add OCR result and transport types

2026-07-21 22:24:51 -04:00
parent 88127a6e8b
commit 2fdd14f64b
2 changed files with 120 additions and 0 deletions
@@ -154,3 +154,61 @@ class TestAuxTypes:
            raw={"id": "x"},
        )
        assert s.raw["id"] == "x"
+
+
+class TestOcrTypes:
+    """M3 OCR 五类型(设计 §3.1): 可构造、frozen、全必填无默认。"""
+
+    def _element(self):
+        from polygateway.types import OcrLayoutElement
+
+        return OcrLayoutElement(type="table", bbox=(41.0, 48.0, 218.0, 282.0), page_index=0)
+
+    def test_layout_element_fields(self):
+        el = self._element()
+        assert el.type == "table" and el.bbox == (41.0, 48.0, 218.0, 282.0) and el.page_index == 0
+        with pytest.raises(dataclasses.FrozenInstanceError):
+            el.type = "text"
+
+    def test_text_result_provenance(self):
+        from polygateway.types import OcrTextResult
+
+        r = OcrTextResult(
+            text="TOSHIBA",
+            source_name="monkey_1",
+            usage=Usage(0, 0),
+            latency_ms=3200,
+            call_id="cid",
+            raw={"success": True},
+        )
+        assert r.text == "TOSHIBA" and r.usage.prompt_tokens == 0
+        with pytest.raises(dataclasses.FrozenInstanceError):
+            r.text = "x"
+
+    def test_layout_result_provenance(self):
+        from polygateway.types import OcrLayoutResult
+
+        r = OcrLayoutResult(
+            elements=[self._element()],
+            page_sizes=[(759.0, 540.0)],
+            source_name="monkey_1",
+            usage=Usage(0, 0),
+            latency_ms=8000,
+            call_id="cid",
+            raw={"has_table": True},
+        )
+        assert r.elements[0].type == "table" and r.page_sizes[0] == (759.0, 540.0)
+
+    def test_transport_results(self):
+        from polygateway.types import OcrLayoutTransportResult, OcrTextTransportResult
+
+        t = OcrTextTransportResult(text="", raw={})
+        assert t.text == ""  # 空串 = 合法"无文字"
+        lt = OcrLayoutTransportResult(elements=[], page_sizes=[(1.0, 1.0)], raw={})
+        assert lt.elements == []  # 空 elements = 合法"无元素"
+
+    def test_all_fields_required(self):
+        from polygateway.types import OcrTextResult
+
+        with pytest.raises(TypeError):
+            OcrTextResult(text="x")  # 溯源件不可省略