From 4fae7d4884fe045b2f95dfaa4880947a12d81c81 Mon Sep 17 00:00:00 2001
From: imClumsyPanda <littlepanda0716@gmail.com>
Date: Sun, 3 Sep 2023 16:13:37 +0800
Subject: [PATCH] add unit test for RapidOCRLoader and RapidOCRPDFLoader

---
 tests/document_loader/test_imgloader.py | 21 +++++++++++++++++++++
 tests/document_loader/test_pdfloader.py | 21 +++++++++++++++++++++
 2 files changed, 42 insertions(+)
 create mode 100644 tests/document_loader/test_imgloader.py
 create mode 100644 tests/document_loader/test_pdfloader.py

diff --git a/tests/document_loader/test_imgloader.py b/tests/document_loader/test_imgloader.py
new file mode 100644
index 0000000..8bba7da
--- /dev/null
+++ b/tests/document_loader/test_imgloader.py
@@ -0,0 +1,21 @@
+import sys
+from pathlib import Path
+
+root_path = Path(__file__).parent.parent.parent
+sys.path.append(str(root_path))
+from pprint import pprint
+
+test_files = {
+    "ocr_test.pdf": str(root_path / "tests" / "samples" / "ocr_test.pdf"),
+}
+
+def test_rapidocrpdfloader():
+    pdf_path = test_files["ocr_test.pdf"]
+    from document_loaders import RapidOCRPDFLoader
+
+    loader = RapidOCRPDFLoader(pdf_path)
+    docs = loader.load()
+    pprint(docs)
+    assert isinstance(docs, list) and len(docs) > 0 and isinstance(docs[0].page_content, str)
+
+
diff --git a/tests/document_loader/test_pdfloader.py b/tests/document_loader/test_pdfloader.py
new file mode 100644
index 0000000..92460cb
--- /dev/null
+++ b/tests/document_loader/test_pdfloader.py
@@ -0,0 +1,21 @@
+import sys
+from pathlib import Path
+
+root_path = Path(__file__).parent.parent.parent
+sys.path.append(str(root_path))
+from pprint import pprint
+
+test_files = {
+    "ocr_test.jpg": str(root_path / "tests" / "samples" / "ocr_test.jpg"),
+}
+
+def test_rapidocrloader():
+    img_path = test_files["ocr_test.jpg"]
+    from document_loaders import RapidOCRLoader
+
+    loader = RapidOCRLoader(img_path)
+    docs = loader.load()
+    pprint(docs)
+    assert isinstance(docs, list) and len(docs) > 0 and isinstance(docs[0].page_content, str)
+
+