site notes pdf to json 🟥

backend/documents_parser/pdf.py

@@ -0,0 +1,5 @@
+from typing import List
+
+
+def pdf_to_text_list(pdf_bytes: bytes) -> List[str]:
+    raise NotImplementedError

backend/documents_parser/tests/test_pdf.py

@@ -0,0 +1,31 @@
+import json
+import os
+
+import pytest
+
+from backend.documents_parser.pdf import pdf_to_text_list
+
+PDF_PATH = os.path.join(os.path.dirname(__file__), "fixtures", "ExampleSiteNotes.pdf")
+FIXTURE_PATH = os.path.join(os.path.dirname(__file__), "fixtures", "site_notes_example_text.json")
+
+
+@pytest.fixture
+def pdf_bytes() -> bytes:
+    with open(PDF_PATH, "rb") as f:
+        return f.read()
+
+
+class TestPdfToTextList:
+    def test_returns_list(self, pdf_bytes: bytes) -> None:
+        result = pdf_to_text_list(pdf_bytes)
+        assert isinstance(result, list)
+
+    def test_all_elements_are_strings(self, pdf_bytes: bytes) -> None:
+        result = pdf_to_text_list(pdf_bytes)
+        assert all(isinstance(t, str) for t in result)
+
+    def test_matches_fixture(self, pdf_bytes: bytes) -> None:
+        with open(FIXTURE_PATH) as f:
+            expected = json.load(f)
+        result = pdf_to_text_list(pdf_bytes)
+        assert result == expected