utils/image: add polygon_for_parent w/ a test

ocrd_utils/ocrd_utils/__init__.py

@@ -97,6 +97,7 @@
     points_from_x0y0x1y1,
     points_from_xywh,
     points_from_y0x0y1x1,
+    polygon_for_parent,
     polygon_from_bbox,
     polygon_from_points,
     polygon_from_x0y0x1y1,

ocrd_utils/ocrd_utils/image.py

@@ -2,6 +2,7 @@
 
 import numpy as np
 from PIL import Image, ImageStat, ImageDraw, ImageChops
+from shapely.geometry import Polygon
 
 from .logging import getLogger
 from .introspect import membername
@@ -20,6 +21,7 @@
     'points_from_x0y0x1y1',
     'points_from_xywh',
     'points_from_y0x0y1x1',
+    'polygon_for_parent',
     'polygon_from_bbox',
     'polygon_from_points',
     'polygon_from_x0y0x1y1',
@@ -541,6 +543,33 @@ def points_from_x0y0x1y1(xyxy):
         x0, y1
     )
 
+def polygon_for_parent(polygon, parent):
+    """
+    Clip polygon to parent polygon range.
+    """
+    childp = Polygon(polygon)
+    # XXX better test would be
+    # if isinstance(parent, PageType):
+    # but that would require ocrd_utils to circularly depened on ocrd_utils
+    if parent.__class__.__name__ == 'PageType':
+        if parent.get_Border():
+            parentp = Polygon(polygon_from_points(parent.get_Border().get_Coords().points))
+        else:
+            parentp = Polygon([[0,0], [0,parent.get_imageHeight()],
+                               [parent.get_imageWidth(),parent.get_imageHeight()],
+                               [parent.get_imageWidth(),0]])
+    else:
+        parentp = Polygon(polygon_from_points(parent.get_Coords().points))
+    if childp.within(parentp):
+        return polygon
+    interp = childp.intersection(parentp)
+    if interp.is_empty:
+        # FIXME: we need a better strategy against this
+        raise Exception("intersection of would-be segment with parent is empty")
+    if interp.type == 'MultiPolygon':
+        interp = interp.convex_hull
+    return interp.exterior.coords[:-1] # keep open
+
 def polygon_from_bbox(minx, miny, maxx, maxy):
     """Construct polygon coordinates in numeric list representation from a numeric list representing a bounding box."""
     return [[minx, miny], [maxx, miny], [maxx, maxy], [minx, maxy]]

ocrd_utils/requirements.txt

@@ -1,2 +1,3 @@
 Pillow >= 7.2.0
 numpy >= 1.17.0
+shapely

ocrd_validators/requirements.txt

@@ -3,4 +3,3 @@ bagit_profile >= 1.3.0
 click >=7
 jsonschema
 pyyaml
-shapely

tests/utils/test_image.py

@@ -0,0 +1,27 @@
+from tests.base import TestCase, main, assets
+from ocrd import Resolver
+from ocrd_modelfactory import page_from_file
+from ocrd_utils import (
+    polygon_for_parent,
+    pushd_popd,
+
+    MIMETYPE_PAGE
+)
+class TestImageUtils(TestCase):
+
+    def test_polygon_for_parent(self):
+        resolver = Resolver()
+        with pushd_popd(assets.path_to('gutachten/data')):
+            ws = resolver.workspace_from_url(assets.path_to('gutachten/data/mets.xml'))
+            input_file = ws.download_file(ws.mets.find_files(mimetype=MIMETYPE_PAGE)[0])
+            pcgts = page_from_file(input_file)
+            page = pcgts.get_Page()
+            page_image, page_coords, page_image_info = ws.image_from_page(page, 'f')
+            width = page_image_info.width
+            height = page_image_info.height
+            p = polygon_for_parent([(-10, -10), (width + 10, -10), (width + 10, height + 10), (-10, height + 10)], page)
+            self.assertEqual(p, [(0, 0), (0, height), (width, height), (width, 0)])
+
+
+if __name__ == '__main__':
+    main(__file__)