qurator-spk/ocrd_calamari
1
0
Fork 0
mirror of https://github.com/mikegerber/ocrd_calamari.git synced 2025-07-23 17:09:54 +02:00
Code Releases Activity

✅ Remove GT text to not accidently check it instead of OCR text

Browse source
This commit is contained in:
Gerber, Mike 2020-02-04 19:29:56 +01:00
parent 82fe0333f1
commit 0f0bae18ba
1 changed files with 16 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

20
test/test_recognize.py
View file

@ -3,6 +3,7 @@ import shutil
import subprocess import subprocess
import urllib.request import urllib.request
from lxml import etree from lxml import etree
from glob import glob
import pytest import pytest
from ocrd.resolver import Resolver from ocrd.resolver import Resolver
@ -10,10 +11,15 @@ from ocrd.resolver import Resolver
from ocrd_calamari import CalamariRecognize from ocrd_calamari import CalamariRecognize
from .base import assets from .base import assets
METS_KANT = assets.url_of('kant_aufklaerung_1784-page-block-line-word_glyph/data/mets.xml') METS_KANT = assets.url_of('kant_aufklaerung_1784-page-block-line-word_glyph/data/mets.xml')
WORKSPACE_DIR = '/tmp/test-ocrd-calamari' WORKSPACE_DIR = '/tmp/test-ocrd-calamari'
CHECKPOINT = os.path.join(os.getcwd(), 'gt4histocr-calamari/*.ckpt.json') CHECKPOINT = os.path.join(os.getcwd(), 'gt4histocr-calamari/*.ckpt.json')
# Because XML namespace versions are so much fun, we not only use one, we use TWO!
NSMAP = { "pc": "http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15" }
NSMAP_GT = { "pc": "http://schema.primaresearch.org/PAGE/gts/pagecontent/2013-07-15" }
@pytest.fixture @pytest.fixture
def workspace(): def workspace():
@ -44,12 +50,20 @@ def workspace():
ff = os.path.join(WORKSPACE_DIR, 'OCR-D-IMG', f) ff = os.path.join(WORKSPACE_DIR, 'OCR-D-IMG', f)
subprocess.call(['convert', ff, '-threshold', '50%', ff]) subprocess.call(['convert', ff, '-threshold', '50%', ff])
# Remove GT Words and TextEquivs, to not accidently check GT text instead of the OCR text
for of in workspace.mets.find_files(fileGrp="OCR-D-GT-SEG-LINE"):
workspace.download_file(of)
for to_remove in ["//pc:Word", "//pc:TextEquiv"]:
for ff in glob(os.path.join(WORKSPACE_DIR, "OCR-D-GT-SEG-LINE", "*")):
tree = etree.parse(ff)
for e in tree.xpath(to_remove, namespaces=NSMAP_GT):
e.getparent().remove(e)
tree.write(ff, xml_declaration=True, encoding="utf-8")
return workspace return workspace
def test_recognize(workspace): def test_recognize(workspace):
# XXX Should remove GT text to really test this
CalamariRecognize( CalamariRecognize(
workspace, workspace,
input_file_grp="OCR-D-GT-SEG-LINE", input_file_grp="OCR-D-GT-SEG-LINE",
@ -81,8 +95,6 @@ def test_word_segmentation(workspace):
assert os.path.exists(page1) assert os.path.exists(page1)
tree = etree.parse(page1) tree = etree.parse(page1)
NSMAP = { "pc": "http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15" }
# The result should contain a TextLine that contains the text "December" # The result should contain a TextLine that contains the text "December"
line = tree.xpath(".//pc:TextLine[pc:TextEquiv/pc:Unicode[contains(text(),'December')]]", namespaces=NSMAP)[0] line = tree.xpath(".//pc:TextLine[pc:TextEquiv/pc:Unicode[contains(text(),'December')]]", namespaces=NSMAP)[0]
assert line assert line