Automatic page splitting

tests/test_fast_page_split.py

@@ -0,0 +1,88 @@
+import unittest
+
+import cv2
+import numpy as np
+import pymupdf
+
+from copienator.auto_page_split import _analyse_page_full as analyse_page
+from copienator.fast_page_split import analyse_page_fast, analyse_pdf_page_fast, quick_page_content
+
+
+class FastPageSplitTests(unittest.TestCase):
+    def test_direct_pdf_strip_respects_rotation_and_crop_offsets(self):
+        image = np.full((1400, 2000, 3), 255, np.uint8)
+        for y in range(150, 650, 70):
+            for x in (150, 1300):
+                cv2.putText(image, 'answer x y z', (x, y),
+                            cv2.FONT_HERSHEY_SIMPLEX, 1, (20, 20, 20), 2)
+        cv2.putText(image, 'x + y = z', (870, 1000),
+                    cv2.FONT_HERSHEY_SIMPLEX, 1, (20, 20, 20), 2)
+        for rotation in (0, 90, 180, 270):
+            with self.subTest(rotation=rotation), pymupdf.open() as doc:
+                source = np.ascontiguousarray(np.rot90(image, rotation//90))
+                h,w = source.shape[:2]
+                page = doc.new_page(width=w*72/150+20,height=h*72/150+30)
+                box = pymupdf.Rect(10,15,w*72/150+10,h*72/150+15)
+                pix = pymupdf.Pixmap(pymupdf.csRGB,w,h,source.tobytes(),False)
+                page.insert_image(box,pixmap=pix)
+                page.set_cropbox(box)
+                page.set_rotation(rotation)
+                rendered = page.get_pixmap(dpi=150,alpha=False)
+                full = np.frombuffer(rendered.samples,np.uint8).reshape(rendered.height,rendered.width,3)
+                reference = analyse_page(full,150,'a3')
+                fast,overview = analyse_pdf_page_fast(page,150,'a3')
+                self.assertIsNone(fast['analysis']['fallback'])
+                self.assertEqual(fast['size_px'],reference['size_px'])
+                self.assertLess(overview.size,full.size)
+                self.assertEqual([o['keep'] for o in fast['outputs']],[True,True])
+                for a,b in zip(fast['outputs'],reference['outputs']):
+                    self.assertLessEqual(max(abs(x-y) for x,y in zip(a['rect_px'],b['rect_px'])),2)
+
+    def test_plain_blank_and_sparse_pencil_use_full_detector(self):
+        for pencil in (False, True):
+            with self.subTest(pencil=pencil):
+                image = np.full((1200, 850, 3), 255, np.uint8)
+                if pencil:
+                    cv2.putText(image, 'pencil', (100, 600),
+                                cv2.FONT_HERSHEY_SIMPLEX, 1, (195, 195, 195), 2)
+                mask, early = quick_page_content(image, 150)
+                self.assertFalse(early)
+                expected = analyse_page(image, 150, 'a4')
+                actual = analyse_page_fast(image, 150, 'a4')
+                self.assertEqual(actual['outputs'][0]['keep'], expected['outputs'][0]['keep'])
+                self.assertEqual(actual['outputs'][0]['rect_px'], [0, 0, 850, 1200])
+                self.assertEqual(bool(mask.any()), pencil)
+
+    def test_compact_roi_maps_bounds_back_to_full_spread(self):
+        image = np.full((1400, 2000, 3), 255, np.uint8)
+        for y in range(150, 650, 70):
+            for x in (150, 1300):
+                cv2.putText(image, 'answer x y z', (x, y),
+                            cv2.FONT_HERSHEY_SIMPLEX, 1, (20, 20, 20), 2)
+        cv2.putText(image, 'x + y = z', (870, 1000),
+                    cv2.FONT_HERSHEY_SIMPLEX, 1, (20, 20, 20), 2)
+        for mirrored in (False, True):
+            with self.subTest(mirrored=mirrored):
+                rgb = image[:, ::-1].copy() if mirrored else image
+                reference = analyse_page(rgb, 150, 'a3')
+                fast = analyse_page_fast(rgb, 150, 'a3')
+                self.assertIsNone(fast['analysis']['fallback'])
+                self.assertEqual(fast['size_px'], [2000, 1400])
+                self.assertEqual(fast['extensions_mm'], reference['extensions_mm'])
+                self.assertEqual([o['rect_px'] for o in fast['outputs']],
+                                 [o['rect_px'] for o in reference['outputs']])
+                self.assertTrue(all(o['keep'] for o in fast['outputs']))
+
+    def test_ambiguous_or_wide_writing_falls_back_without_truncation(self):
+        image = np.full((700, 1000, 3), 255, np.uint8)
+        image[200:230, 300:650] = 20
+        reference = analyse_page(image, 100, 'a3')
+        fast = analyse_page_fast(image, 100, 'a3')
+        self.assertEqual(fast['analysis']['fallback'], 'full-spread')
+        self.assertEqual(fast['outputs'], reference['outputs'])
+        self.assertEqual(fast['crossings'], reference['crossings'])
+
+    def test_unknown_geometry_keeps_reference_behaviour(self):
+        image = np.full((400, 500, 3), 255, np.uint8)
+        self.assertEqual(analyse_page_fast(image, 100, 'unknown'),
+                         analyse_page(image, 100, 'unknown'))