markup: matching works

markup/utils.py

@@ -0,0 +1,66 @@
+def pdf_rect(rect, container_height):
+    x1 = min(rect[0], rect[2])
+    y1 = max(rect[1], rect[3])
+    x2 = max(rect[0], rect[2])
+    y2 = min(rect[1], rect[3])
+    # and convert from pdf to image coords
+    return Rect(x1, container_height - y1, x2, container_height - y2)
+
+
+def cv2_rect(l, t, w, h):
+    return Rect(l, t, l + w, t + h)
+
+
+def overlaps(r1, r2, threshold):
+    A = r1.to_dict()
+    B = r2.to_dict()
+
+    # https://stackoverflow.com/questions/9324339/how-much-do-two-rectangles-overlap
+    SA = A['w'] * A['h']
+    SB = B['w'] * B['h']
+    SI = max([0, 1 + min([A['x2'], B['x2']]) - max([A['x1'], B['x1']])]) * max([0, 1 + min([A['y2'], B['y2']]) - max([A['y1'], B['y1']])])
+    SU = SA + SB - SI
+    overlap = float(SI) / float(SU)
+
+    print('overlap: {}'.format(int(overlap * 100)))
+    return overlap > threshold
+
+
+class Rect(object):
+
+    def __init__(self, l, t, r, b):
+        self.left = l
+        self.top = t
+        self.right = r
+        self.bottom = b
+
+    def translate(self, x, y):
+        self.left += x
+        self.top += y
+        self.right += x
+        self.bottom += y
+        return self
+
+    def scale(self, x, y):
+        self.left *= x
+        self.top *= y
+        self.right *= x
+        self.bottom *= y
+        return self
+
+    def p1(self, page_height):
+        return (self.left, self.top)
+
+    def p2(self, page_height):
+        return (self.right, self.bottom)
+
+    def to_dict(self):
+        return {'x1': self.left,
+                'y1': self.top,
+                'x2': self.right,
+                'y2': self.bottom,
+                'w': self.right - self.left,
+                'h': self.bottom - self.top }
+
+    def __repr__(self):
+        return 'Rect[l={}, t={}, r={}, b={}]'.format(int(self.left), int(self.top), int(self.right), int(self.bottom))