# pylint: disable=missing-module-docstring,invalid-name
from typing import Optional
from itertools import count
from collections import OrderedDict as odict
import click
from ocrd import Processor, OcrdPageResult
from ocrd.decorators import ocrd_cli_options, ocrd_cli_wrap_processor
from ocrd_modelfactory import page_from_file
from ocrd_models import OcrdPage
from ocrd_models.ocrd_page import (
BorderType,
CoordsType,
ReadingOrderType,
UnorderedGroupType,
)
from ocrd_utils import bbox_from_points
_SEGTYPES = [
"NoiseRegion",
"LineDrawingRegion",
"AdvertRegion",
"ImageRegion",
"ChartRegion",
"MusicRegion",
"GraphicRegion",
"UnknownRegion",
"CustomRegion",
"SeparatorRegion",
"MathsRegion",
"TextRegion",
"MapRegion",
"ChemRegion",
"TableRegion",
"TextLine",
"Word",
"Glyph"
]
[docs]
def get_border_bbox(pcgts):
if pcgts.Page.Border is None:
return [0, 0, pcgts.Page.imageWidth, pcgts.Page.imageHeight]
return bbox_from_points(pcgts.Page.Border.Coords.points)
[docs]
def rename_segments(pcgts, start=1):
renamed = {}
rodict = pcgts.Page.get_ReadingOrderGroups()
# get everything that has an identifier
nodes = pcgts.xpath("//*[@id]")
# filter segments
segments = [segment for segment in map(pcgts.revmap.get, nodes)
# get PAGE objects from matching etree nodes
# but allow only hierarchy segments
if segment.__class__.__name__.replace('Type', '') in _SEGTYPES]
# count segments and rename them
# fixme: or perhaps better to have each segment type named and counted differently?
num = 0
regions = []
for num, segment in zip(count(start=start), segments):
segtype = segment.original_tagname_
#parent = segment.parent_object_
newname = "seg%011d" % num
assert not segment.id in renamed
if segtype.endswith('Region') and segment.id in rodict:
# update reading order
roelem = rodict[segment.id]
roelem.regionRef = newname
renamed[segment.id] = newname
segment.id = newname
return num
[docs]
class MergeProcessor(Processor):
[docs]
def process_page_pcgts(self, *input_pcgts: Optional[OcrdPage], page_id: Optional[str] = None) -> OcrdPageResult:
"""
Merge PAGE segment hierarchy elements from all input file groups.
For each page, open and deserialise PAGE input files. Rename all elements
of the segment hierarchy to new (clash-free) identifers. Redefine the
`Border` coordinates as the convex hull of all input borders. Then add all
regions from all input files, concatenating them into a single `ReadingOrder`
in the order of input file groups.
Produce a new PAGE output file by serialising the resulting hierarchy.
"""
actual_pcgts = list(filter(None, input_pcgts))
assert len(set(pcgts.Page.imageFilename for pcgts in actual_pcgts)) == 1, \
"input files must all reference the same @imageFilename"
# create new PAGE for image
result = OcrdPageResult(page_from_file(actual_pcgts[0].Page.imageFilename))
# unify Border
borders = [get_border_bbox(pcgts) for pcgts in actual_pcgts]
minx, miny, maxx, maxy = zip(*borders)
minx = min(minx)
miny = min(miny)
maxx = max(maxx)
maxy = max(maxy)
result.pcgts.Page.set_Border(
BorderType(CoordsType(
points=f"{minx},{miny} {maxx},{miny} {maxx},{maxy} {minx},{maxy}")))
# rename all segments
num = 1
for pcgts in actual_pcgts:
num = rename_segments(pcgts, num)
# concatenate all regions
ug = UnorderedGroupType(id="merged")
result.pcgts.Page.set_ReadingOrder(ReadingOrderType(UnorderedGroup=ug))
for pcgts in actual_pcgts:
for region in pcgts.Page.get_AllRegions():
adder = getattr(result.pcgts.Page, 'add_' + region.original_tagname_)
adder(region)
if pcgts.Page.ReadingOrder:
group = pcgts.Page.ReadingOrder.OrderedGroup or pcgts.Page.ReadingOrder.UnorderedGroup
adder = getattr(ug, 'add_' + group.original_tagname_)
adder(group)
return result
@property
def metadata_filename(self):
return 'processor/builtin/dummy/ocrd-tool.json'
@property
def executable(self):
return 'ocrd-merge'
@click.command()
@ocrd_cli_options
def cli(*args, **kwargs):
return ocrd_cli_wrap_processor(MergeProcessor, *args, **kwargs)
