Cropping of individual exos

tests/test_crop_exercise_bottoms.py

@@ -0,0 +1,59 @@
+import tempfile
+import unittest
+from pathlib import Path
+
+import pymupdf
+
+from copienator.crop_exercise_bottoms import process_exercise_pdf
+
+
+class CropExerciseBottomsTests(unittest.TestCase):
+    def setUp(self):
+        self.temp = tempfile.TemporaryDirectory()
+        self.addCleanup(self.temp.cleanup)
+        self.root = Path(self.temp.name)
+        self.review = self.root / "review"
+
+    def make_pdf(self, name: str, height: float, text_y: float, footer=True) -> Path:
+        path = self.root / name
+        with pymupdf.open() as document:
+            page = document.new_page(width=600, height=height)
+            page.insert_text((80, text_y), "student answer", fontsize=16)
+            if footer:
+                page.insert_text((20, height - 3), "Ex 2", fontsize=10)
+            document.save(path)
+        return path
+
+    def test_short_page_is_skipped(self):
+        source = self.make_pdf("short.pdf", 200, 100)
+        destination = self.root / "out" / source.name
+        records = process_exercise_pdf(
+            source, destination, self.review, 800, dpi=150
+        )
+        self.assertEqual(records[0]["status"], "skipped-short")
+        self.assertFalse(destination.exists())
+
+    def test_footer_fragment_is_ignored_for_a_large_bottom_crop(self):
+        source = self.make_pdf("large.pdf", 400, 100)
+        destination = self.root / "out" / source.name
+        records = process_exercise_pdf(
+            source, destination, self.review, 800, dpi=150
+        )
+        self.assertEqual(records[0]["status"], "cropped")
+        self.assertGreaterEqual(records[0]["bottom_removed_lines"], 4)
+        with pymupdf.open(destination) as result:
+            self.assertLess(result[0].rect.height, 250)
+            self.assertAlmostEqual(result[0].rect.width, 600)
+
+    def test_crop_smaller_than_four_lines_is_not_written(self):
+        source = self.make_pdf("small.pdf", 400, 330, footer=False)
+        destination = self.root / "out" / source.name
+        records = process_exercise_pdf(
+            source, destination, self.review, 800, dpi=150
+        )
+        self.assertEqual(records[0]["status"], "unchanged-small-crop")
+        self.assertFalse(destination.exists())
+
+
+if __name__ == "__main__":
+    unittest.main()