Allow signaling issues in label verification

tests/test_pdf_supplements.py

@@ -0,0 +1,110 @@
+import tempfile
+import unittest
+from pathlib import Path
+from types import SimpleNamespace
+from unittest.mock import patch
+
+from PIL import Image
+from reportlab.pdfgen import canvas
+
+from copienator.commands.reading_grouped_annotations import (
+    context_content_key, get_extra_pdfs_as_images, save_paginated_pdf,
+)
+
+
+class PdfSupplementTests(unittest.TestCase):
+    def setUp(self):
+        temporary = tempfile.TemporaryDirectory()
+        self.addCleanup(temporary.cleanup)
+        self.root = Path(temporary.name)
+        (self.root / "Text2").mkdir()
+        (self.root / "Sol2").mkdir()
+        self.labels = ["Q1", "Q2", "Q3"]
+        self.rendered = []
+
+        def render(path):
+            self.rendered.append(Path(path).name)
+            return Image.new("RGB", (100, 30), "white"), 30, 100
+
+        self.renderer = SimpleNamespace(make_base_image=render)
+
+    def pdf(self, folder, name, text):
+        path = self.root / folder / name
+        document = canvas.Canvas(str(path))
+        document.drawString(20, 800, text)
+        document.save()
+        return path
+
+    def extras(self, label, seen):
+        return get_extra_pdfs_as_images(
+            self.root, label, self.renderer, self.labels, seen_contexts=seen,
+        )
+
+    def test_content_deduplication_across_ranges_and_copy_reset(self):
+        self.pdf("Text2", "CTXT Q1 -> Q2.pdf", "Bob   collects coupons: N >= 1")
+        self.pdf("Text2", "CTXT Q2 -> Q3.pdf", "Bob collects coupons: N >= 1")
+        self.pdf("Text2", "CTXT Q2 -> Q2.pdf", "Bob collects coupons: N >= 2")
+        seen = set()
+        self.assertEqual(len(self.extras("Q1", seen)), 1)
+        self.assertEqual(len(self.extras("Q2", seen)), 1)
+        self.assertEqual(len(self.extras("Q3", seen)), 0)
+        self.assertEqual(len(seen), 2)
+        self.assertEqual(len(self.extras("Q3", set())), 1)
+
+    def test_first_retained_question_gets_context_and_semantic_borders(self):
+        self.pdf("Text2", "CTXT Q1 -> Q3.pdf", "Context")
+        self.pdf("Text2", "CTXT Q2 -> Q3.pdf", "Other context")
+        self.pdf("Text2", "Q2.pdf", "Question")
+        self.pdf("Sol2", "Q2.pdf", "Solution")
+        # Q1 was filtered out; it never consumes the shared context.
+        images = self.extras("Q2", set())
+        self.assertEqual([i.getpixel((0, 0)) for i in images],
+                         [(255, 255, 255), (255, 255, 255), (0, 0, 0), (0, 0, 255)])
+        self.assertTrue(all(i.getpixel((50, 15)) == (255, 255, 255) for i in images))
+
+    def test_missing_question_does_not_shift_solution_border(self):
+        self.pdf("Text2", "CTXT Q1 -> Q3.pdf", "Context")
+        self.pdf("Sol2", "Q2.pdf", "Solution")
+        images = self.extras("Q2", set())
+        self.assertEqual([i.getpixel((0, 0)) for i in images],
+                         [(255, 255, 255), (0, 0, 255)])
+
+    def test_unrendered_context_is_not_consumed(self):
+        self.pdf("Text2", "CTXT Q1 -> Q3.pdf", "Context")
+        seen = set()
+        with patch.object(self.renderer, "make_base_image", return_value=(None, 0, 0)):
+            self.assertEqual(self.extras("Q1", seen), [])
+        self.assertFalse(seen)
+        self.assertEqual(len(self.extras("Q2", seen)), 1)
+
+    def test_missing_solution_does_not_shift_border_to_student_answer(self):
+        self.pdf("Text2", "Q2.pdf", "Question")
+        images = self.extras("Q2", set())
+        answer = Image.new("RGB", (100, 200), "white")
+        captured = []
+
+        def save(page, *args, **kwargs):
+            captured.extend([page.copy(), *kwargs['append_images']])
+
+        with patch.object(Image.Image, "save", save):
+            save_paginated_pdf([images + [answer], [answer]], self.root / "out.pdf")
+        self.assertEqual(len(captured), 2)
+        self.assertEqual(captured[0].getpixel((11, 7)), (0, 0, 0))
+        self.assertEqual(captured[0].getpixel((11, 37)), (255, 255, 255))
+        self.assertEqual(captured[1].getpixel((11, 7)), (255, 255, 255))
+        self.assertEqual(answer.getpixel((0, 0)), (255, 255, 255))
+
+    def test_empty_extracted_text_does_not_merge_distinct_contexts(self):
+        paths = []
+        for name, color in [("a.pdf", "red"), ("b.pdf", "blue")]:
+            path = self.root / name
+            document = canvas.Canvas(str(path))
+            document.setFillColor(color)
+            document.rect(20, 20, 100, 100, fill=1)
+            document.save()
+            paths.append(path)
+        self.assertNotEqual(context_content_key(paths[0]), context_content_key(paths[1]))
+
+
+if __name__ == "__main__":
+    unittest.main()