test: add unified probe unit tests (8 Rust + 6 Python), fix pre-existing test compilation errors

scripts/test_probe_file.py

@@ -0,0 +1,92 @@
+#!/opt/homebrew/bin/python3.11
+"""Unit tests for probe_file.py"""
+
+import sys
+import json
+import os
+import unittest
+
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+import probe_file  # noqa: E402
+
+
+class TestProbePDF(unittest.TestCase):
+    def setUp(self):
+        self.path = "/tmp/_test_probe.pdf"
+        if not os.path.exists(self.path):
+            from PyPDF2 import PdfWriter
+            w = PdfWriter()
+            w.add_blank_page(210, 297)
+            with open(self.path, 'wb') as f:
+                w.write(f)
+
+    def test_pdf_has_pages(self):
+        result = json.loads(probe_file.probe(self.path))
+        self.assertIn("document", result)
+        self.assertGreater(result["document"]["pages"], 0)
+
+    def test_pdf_has_streams(self):
+        result = json.loads(probe_file.probe(self.path))
+        self.assertEqual(result["streams"], [])
+
+
+class TestProbeDOCX(unittest.TestCase):
+    def setUp(self):
+        self.path = "/tmp/_test_probe.docx"
+        if not os.path.exists(self.path):
+            from docx import Document
+            d = Document()
+            d.add_paragraph("Test paragraph.")
+            d.save(self.path)
+
+    def test_docx_has_paragraphs(self):
+        result = json.loads(probe_file.probe(self.path))
+        self.assertIn("document", result)
+        self.assertGreater(result["document"]["paragraphs"], 0)
+
+
+class TestProbeXLSX(unittest.TestCase):
+    def setUp(self):
+        self.path = "/tmp/_test_probe.xlsx"
+        if not os.path.exists(self.path):
+            import openpyxl
+            wb = openpyxl.Workbook()
+            ws = wb.active
+            ws.title = "Sheet1"
+            ws.append(["A", 1])
+            wb.save(self.path)
+
+    def test_xlsx_has_sheets(self):
+        result = json.loads(probe_file.probe(self.path))
+        self.assertIn("spreadsheet", result)
+        self.assertGreater(result["spreadsheet"]["sheet_count"], 0)
+
+
+class TestProbePPTX(unittest.TestCase):
+    def setUp(self):
+        self.path = "/tmp/_test_probe.pptx"
+        if not os.path.exists(self.path):
+            from pptx import Presentation
+            prs = Presentation()
+            slide = prs.slides.add_slide(prs.slide_layouts[0])
+            slide.shapes.title.text = "Test"
+            prs.save(self.path)
+
+    def test_pptx_has_slides(self):
+        result = json.loads(probe_file.probe(self.path))
+        self.assertIn("presentation", result)
+        self.assertGreater(result["presentation"]["slide_count"], 0)
+
+
+class TestProbeUnknown(unittest.TestCase):
+    def test_unknown_extension(self):
+        path = "/tmp/_test_probe.xyz"
+        with open(path, 'w') as f:
+            f.write("test")
+        result = json.loads(probe_file.probe(path))
+        os.remove(path)
+        self.assertEqual(result["streams"], [])
+
+
+if __name__ == "__main__":
+    unittest.main()