feat(ocr): PDF 文本层可用时跳过 worker 调用并补装 poppler-data

- OcrService 提取 PDF 文本层后若有效字符达到阈值，直接构建文档并写入结果缓存，不再触发 OCR worker，仅无文本层时才解析 python_bin/worker_path 调用 worker
- _build_text_layer_document 复用 AggregatedOcrDocument 聚合文本层片段，_has_usable_pdf_text_layer 基于 meaningful_char_count 判定
- docker-compose 与 paddleocr bootstrap 脚本补装 poppler-data，保证 PDF 文本层抽取的中文编码正确
- 新增文本层直取与运行时依赖两项 ocr_service 单测

server/tests/test_ocr_service.py

@@ -8,6 +8,18 @@ from app.core.config import get_settings
 from app.services.ocr import OcrService
 
 
+def test_ocr_runtime_installers_include_poppler_cjk_data() -> None:
+    repo_root = Path(__file__).resolve().parents[2]
+    dependency_sources = [
+        repo_root / "docker-compose.yml",
+        repo_root / "server" / "scripts" / "bootstrap_paddleocr_mobile.sh",
+        repo_root / "server" / "scripts" / "bootstrap_paddleocr_gpu.sh",
+    ]
+
+    for path in dependency_sources:
+        assert "poppler-data" in path.read_text(encoding="utf-8")
+
+
 def test_ocr_service_uses_worker_runtime_and_keeps_unsupported_files_as_warnings(
     monkeypatch,
     tmp_path: Path,
@@ -220,6 +232,59 @@ def test_ocr_service_converts_pdf_to_images_and_returns_image_preview(
     assert recognized.lines[1].page_index == 1
 
 
+def test_ocr_service_uses_pdf_text_layer_without_worker_runtime(
+    monkeypatch,
+    tmp_path: Path,
+) -> None:
+    def fake_convert_pdf_to_images(self, *, pdf_path: Path, output_dir: Path) -> list[Path]:
+        page = output_dir / "page-1.png"
+        page.write_bytes(b"fake-rendered-page")
+        return [page]
+
+    def fail_resolve_python(self) -> str:
+        raise AssertionError("PDF 文本层可用时不应强制解析 OCR worker。")
+
+    def fail_invoke_worker(self, **kwargs) -> dict:
+        raise AssertionError("PDF 文本层可用时不应调用 OCR worker。")
+
+    monkeypatch.setenv("STORAGE_ROOT_DIR", str(tmp_path / "storage"))
+    monkeypatch.setattr(OcrService, "_resolve_python_bin", fail_resolve_python)
+    monkeypatch.setattr(OcrService, "_resolve_worker_path", lambda self: "worker.py")
+    monkeypatch.setattr(OcrService, "_convert_pdf_to_images", fake_convert_pdf_to_images)
+    monkeypatch.setattr(OcrService, "_invoke_worker", fail_invoke_worker)
+    monkeypatch.setattr(
+        OcrService,
+        "_extract_pdf_text_layer",
+        lambda self, pdf_path: (
+            "电子发票（铁路电子客票）\n"
+            "发票号码:26429165800002785705\n"
+            "武汉站\n"
+            "上海虹桥站\n"
+            "G458\n"
+            "票价:￥354.00\n"
+            "电子客票号:6580061086021391007342026"
+        ),
+    )
+    get_settings.cache_clear()
+    try:
+        result = OcrService().recognize_files(
+            [
+                ("train-ticket.pdf", b"%PDF-1.7 fake", "application/pdf"),
+            ]
+        )
+    finally:
+        get_settings.cache_clear()
+
+    recognized = result.documents[0]
+    assert result.success_count == 1
+    assert recognized.document_type == "train_ticket"
+    assert "电子发票（铁路电子客票）" in recognized.text
+    assert "电子客票号:6580061086021391007342026" in recognized.text
+    assert any(field.label == "金额" and field.value == "354元" for field in recognized.document_fields)
+    assert recognized.preview_kind == "image"
+    assert recognized.preview_data_url.startswith("data:image/png;base64,")
+
+
 def test_ocr_service_reuses_cached_document_for_same_content(
     monkeypatch,
     tmp_path: Path,
@@ -351,5 +416,5 @@ def test_ocr_service_prefers_pdf_text_layer_when_rendered_ocr_is_placeholder_hea
     assert "上海虹桥站" in recognized.text
     assert "□□□□" not in recognized.summary
     assert recognized.document_type == "train_ticket"
-    assert recognized.preview_kind == ""
-    assert recognized.preview_data_url == ""
+    assert recognized.preview_kind == "image"
+    assert recognized.preview_data_url.startswith("data:image/png;base64,")