Merge pull request #48 from bertsky/fix-coords

ensure valid coordinates by intersection with parent…

qurator/sbb_textline_detector/ocrd_cli.py

@@ -7,17 +7,17 @@ import ocrd_models.ocrd_page
 from ocrd import Processor
 from ocrd.decorators import ocrd_cli_options, ocrd_cli_wrap_processor
 from ocrd_modelfactory import page_from_file
-from ocrd_models import OcrdFile
+from ocrd_models.ocrd_page_generateds import CoordsType, PageType
-from ocrd_models.ocrd_page_generateds import MetadataItemType, LabelsType, LabelType, \
-    CoordsType
 from ocrd_utils import (
     assert_file_grp_cardinality,
     getLogger,
     make_file_id,
-    MIMETYPE_PAGE,
     coordinates_for_segment,
     polygon_from_points, points_from_polygon,
 )
+import numpy as np
+from shapely.geometry import Polygon, asPolygon
+from shapely.ops import unary_union
 
 from pkg_resources import resource_string
 
@@ -63,8 +63,8 @@ class OcrdSbbTextlineDetectorRecognize(Processor):
             page = pcgts.get_Page()
             page_image, page_coords, page_image_info = \
                 self.workspace.image_from_page(
-                        page, page_id,
+                    page, page_id,
-                        feature_filter='cropped,binarized,grayscale_normalized'
+                    feature_filter='cropped,binarized,grayscale_normalized'
                 )
 
             with tempfile.TemporaryDirectory() as tmp_dirname:
@@ -84,64 +84,134 @@ class OcrdSbbTextlineDetectorRecognize(Processor):
 
             # Create a new PAGE file from the input file
             pcgts.set_pcGtsId(file_id)
-            page = pcgts.get_Page()
 
             # Merge results → PAGE file
 
             # 1. Border
             if page.get_Border():
-                log.warning("Page already contained a border")
+                log.warning("Removing existing page border")
+            page.set_Border(None)
             # We need to translate the coordinates:
-            text_border = tmp_page.get_Border()
+            text_border = adapt_coords(tmp_page.get_Border(), page, page_coords)
-            coords = text_border.get_Coords().get_points()
+            if text_border is None:
-            polygon = polygon_from_points(coords)
+                # intersection is empty (border outside of rotated original image)
-            polygon_new = coordinates_for_segment(polygon, page_image, page_coords)
+                log.warning("new border would be empty, skipping")
-            points_new = points_from_polygon(polygon_new)
+            else:
-            coords_new = CoordsType(points=points_new)
+                page.set_Border(text_border)
-            text_border.set_Coords(coords_new)
-            page.set_Border(text_border)
 
             # 2. ReadingOrder
             if page.get_ReadingOrder():
-                log.warning("Page already contained a reading order")
+                log.warning("Removing existing regions' reading order")
             page.set_ReadingOrder(tmp_page.get_ReadingOrder())
 
             # 3. TextRegion
+            # FIXME: what about table and image regions?
             if page.get_TextRegion():
-                log.warning("Page already contained text regions")
+                log.warning("Removing existing text regions")
             # We need to translate the coordinates:
             text_regions_new = []
             for text_region in tmp_page.get_TextRegion():
-                coords = text_region.get_Coords().get_points()
+                text_region = adapt_coords(text_region, page, page_coords)
-                polygon = polygon_from_points(coords)
+                if text_region is None:
-                polygon_new = coordinates_for_segment(polygon, page_image, page_coords)
+                    # intersection is empty (polygon outside of above border)
-                points_new = points_from_polygon(polygon_new)
+                    log.warning("new text region polygon would be empty, skipping")
-                coords_new = CoordsType(points=points_new)
+                    continue
-                text_region.set_Coords(coords_new)
                 text_regions_new.append(text_region)
+                text_lines_new = []
+                for text_line in text_region.get_TextLine():
+                    text_line = adapt_coords(text_line, text_region, page_coords)
+                    if text_line is None:
+                        # intersection is empty (polygon outside of region)
+                        log.warning("new text line polygon would be empty, skipping")
+                        continue
+                    text_lines_new.append(text_line)
+                text_region.set_TextLine(text_lines_new)
             page.set_TextRegion(text_regions_new)
 
             # Save metadata about this operation
-            metadata = pcgts.get_Metadata()
+            self.add_metadata(pcgts)
-            metadata.add_MetadataItem(
-                MetadataItemType(type_="processingStep",
-                                 name=self.ocrd_tool['steps'][0],
-                                 value=TOOL,
-                                 Labels=[LabelsType(
-                                     externalModel="ocrd-tool",
-                                     externalId="parameters",
-                                     Label=[LabelType(type_=name, value=self.parameter[name])
-                                            for name in self.parameter.keys()])]))
 
             self.workspace.add_file(
-                 ID=file_id,
+                ID=file_id,
-                 file_grp=self.output_file_grp,
+                file_grp=self.output_file_grp,
-                 pageId=page_id,
+                pageId=page_id,
-                 mimetype='application/vnd.prima.page+xml',
+                mimetype='application/vnd.prima.page+xml',
-                 local_filename=os.path.join(self.output_file_grp, file_id) + '.xml',
+                local_filename=os.path.join(self.output_file_grp, file_id) + '.xml',
-                 content=ocrd_models.ocrd_page.to_xml(pcgts)
+                content=ocrd_models.ocrd_page.to_xml(pcgts)
             )
 
 
+def adapt_coords(segment, parent, transform):
+    points = segment.get_Coords().get_points()
+    polygon = polygon_from_points(points)
+    # polygon absolute coords (after transforming back from page coords, e.g. deskewing)
+    polygon_new = coordinates_for_segment(polygon, None, transform)
+    # intersection with parent polygon
+    polygon_new = polygon_for_parent(polygon_new, parent)
+    if polygon_new is None:
+        return None
+    points_new = points_from_polygon(polygon_new)
+    segment.set_Coords(CoordsType(points=points_new))
+    return segment
+
+# from ocrd_tesserocr, to be integrated into core (somehow)...
+def polygon_for_parent(polygon, parent):
+    """Clip polygon to parent polygon range.
+
+    (Should be moved to ocrd_utils.coordinates_for_segment.)
+    """
+    childp = Polygon(polygon)
+    if isinstance(parent, 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))
+    # check if clipping is necessary
+    if childp.within(parentp):
+        return polygon
+    # ensure input coords have valid paths (without self-intersection)
+    # (this can happen when shapes valid in floating point are rounded)
+    childp = make_valid(childp)
+    parentp = make_valid(parentp)
+    # clip to parent
+    interp = childp.intersection(parentp)
+    if interp.is_empty or interp.area == 0.0:
+        # this happens if Tesseract "finds" something
+        # outside of the valid Border of a deskewed/cropped page
+        # (empty corners created by masking); will be ignored
+        return None
+    if interp.type == 'GeometryCollection':
+        # heterogeneous result: filter zero-area shapes (LineString, Point)
+        interp = unary_union([geom for geom in interp.geoms if geom.area > 0])
+    if interp.type == 'MultiPolygon':
+        # homogeneous result: construct convex hull to connect
+        # FIXME: construct concave hull / alpha shape
+        interp = interp.convex_hull
+    if interp.minimum_clearance < 1.0:
+        # follow-up calculations will necessarily be integer;
+        # so anticipate rounding here and then ensure validity
+        interp = asPolygon(np.round(interp.exterior.coords))
+        interp = make_valid(interp)
+    return interp.exterior.coords[:-1] # keep open
+
+# from ocrd_tesserocr, to be integrated into core (somehow)...
+def make_valid(polygon):
+    for split in range(1, len(polygon.exterior.coords)-1):
+        if polygon.is_valid or polygon.simplify(polygon.area).is_valid:
+            break
+        # simplification may not be possible (at all) due to ordering
+        # in that case, try another starting point
+        polygon = Polygon(polygon.exterior.coords[-split:]+polygon.exterior.coords[:-split])
+    for tolerance in range(1, int(polygon.area)):
+        if polygon.is_valid:
+            break
+        # simplification may require a larger tolerance
+        polygon = polygon.simplify(tolerance)
+    return polygon
+
 if __name__ == '__main__':
     ocrd_sbb_textline_detector()

requirements.txt

@@ -9,4 +9,5 @@ scikit-learn
 numpy == 1.18.*  # XXX for tensorflow-gpu 1.15
 tensorflow-gpu ~=1.15.2
 scipy
-ocrd >= 2.0.0
+ocrd >= 2.18.0
+shapely >= 1.7.1