Merge pull request #151 from kba/replace-xmllint

replace xmllint callout with lxml solution, fix #72

.gitignore

@@ -123,3 +123,4 @@ test/assets/00000001.jpg
 test/assets/test.ocrd.zip
 /foo
 sanders*
+ws1

CHANGELOG.md

@@ -5,6 +5,10 @@ Versioned according to [Semantic Versioning](http://semver.org/).
 
 ## Unreleased
 
+Changed:
+
+  * Remove dependency on `xmllint` command line tool, #72, #151
+
 ## [0.7.0] - 2018-07-25
 
 Changed:

Makefile

@@ -41,8 +41,7 @@ deps-ubuntu:
 	sudo apt install -y \
 		python3 \
 		python3-pip \
-		libimage-exiftool-perl \
-		libxml2-utils
+		libimage-exiftool-perl
 
 # Install python deps via pip
 deps-pip:

README.rst

@@ -22,7 +22,6 @@ To bootstrap the tool, you'll need installed (Ubuntu packages):
 * Python (``python``)
 * pip (``python-pip``)
 * exiftool (``libimage-exiftool-perl``)
-* libxml2-utils for xmllint (``libxml2-utils``)
 
 To install system-wide:
 

ocrd/utils.py

@@ -12,7 +12,6 @@
 
 xywh is what tesserocr expects/produces.
 """
-import subprocess
 import logging
 import re
 import sys
@@ -72,15 +71,14 @@ def polygon_from_points(points):
         polygon.append([float(x_y[0]), float(x_y[1])])
     return polygon
 
-# https://stackoverflow.com/a/10133365/201318
 def xmllint_format(xml):
-    proc = subprocess.Popen(
-        ['xmllint', '--format', '/dev/stdin'],
-        stdin=subprocess.PIPE,
-        stdout=subprocess.PIPE,
-    )
-    output, _ = proc.communicate(xml)
-    return output
+    from lxml import etree as ET
+    parser = ET.XMLParser(resolve_entities=False, strip_cdata=False, remove_blank_text=True)
+    document = ET.fromstring(xml, parser)
+    return ('%s\n%s' % (
+        '<?xml version="1.0" encoding="UTF-8"?>',
+        ET.tostring(document, pretty_print=True)
+    )).encode('utf-8')
 
 def is_string(val):
     # pylint: disable=undefined-variable

test/test_utils.py

@@ -1,14 +1,15 @@
-from test.base import TestCase, main
+from test.base import TestCase, main, assets
 from ocrd.utils import (
     points_from_xywh,
     is_string,
     concat_padded,
     points_from_x0y0x1y1,
     xywh_from_points,
-    polygon_from_points
+    polygon_from_points,
+    xmllint_format
 )
 
-class TestResolver(TestCase):
+class TestUtils(TestCase):
 
     #  def runTest(self):
 
@@ -44,5 +45,12 @@ def test_is_string(self):
         self.assertTrue(is_string('x'))
         self.assertTrue(is_string(u'x'))
 
+    #  def test_xmllint(self):
+    #      with open(assets.path_to('page-with-glyphs.xml'), 'rb') as f:
+    #          xml_as_str = f.read()
+    #          pretty_xml = xmllint_format(xml_as_str)
+    #          self.assertEqual(xml_as_str, pretty_xml)
+
+
 if __name__ == '__main__':
     main()