def test_should_return_one_page(self):
lxml_root = E.DOCUMENT(
E.PAGE(
)
)
svg_pages = list(iter_svg_pages_for_lxml(lxml_root))
assert len(svg_pages) == 1
def test_should_add_background_rect(self):
lxml_root = E.DOCUMENT(E.PAGE(width='600', height='800'))
svg_pages = list(iter_svg_pages_for_lxml(lxml_root))
assert len(svg_pages) == 1
background_rect = svg_pages[0].xpath('svg:rect[@class="background"]',
namespaces={'svg': SVG_NS})
assert len(background_rect) == 1
def test_should_create_text_node_with_common_attributes(self):
lxml_root = E.DOCUMENT(
E.PAGE(
E.TEXT(
E.TOKEN(
SOME_TEXT,
COMMON_LXML_TOKEN_ATTRIBS
)
)
)
)
svg_pages = list(iter_svg_pages_for_lxml(lxml_root))
assert len(svg_pages) == 1
first_page = svg_pages[0]
svg_text = first_page.find('.//' + SVG_TEXT)
assert svg_text is not None
assert svg_text.text == SOME_TEXT
assert float(svg_text.attrib[SVG.X]) == float(SOME_X)
assert float(svg_text.attrib[SVG.Y]) == float(SOME_Y)
assert float(svg_text.attrib[SVG.FONT_SIZE]) == float(SOME_FONT_SIZE)
assert svg_text.attrib[SVG.FONT_FAMILY] == SOME_FONT_FAMILY
assert svg_text.attrib[SVG.FILL] == SOME_FONT_COLOR
assert parse_bounding_box(svg_text.attrib.get(SVG.BOUNDING_BOX)) == BoundingBox(
float(COMMON_LXML_TOKEN_ATTRIBS[LXML.X]),
float(COMMON_LXML_TOKEN_ATTRIBS[LXML.Y]),
float(COMMON_LXML_TOKEN_ATTRIBS[LXML.WIDTH]),
float(COMMON_LXML_TOKEN_ATTRIBS[LXML.HEIGHT])
)
def test_should_set_svg_dimensions(self):
lxml_root = E.DOCUMENT(
E.PAGE(
width='600',
height='800'
)
)
svg_pages = list(iter_svg_pages_for_lxml(lxml_root))
assert len(svg_pages) == 1
assert svg_pages[0].attrib.get('viewBox') == '0 0 600 800'
def test_should_use_base_as_y_in_svg_if_available(self):
lxml_root = E.DOCUMENT(
E.PAGE(
E.TEXT(
E.TOKEN(
SOME_TEXT,
dict_extend(COMMON_LXML_TOKEN_ATTRIBS,
{LXML.BASE: SOME_BASE})))))
svg_pages = list(iter_svg_pages_for_lxml(lxml_root))
assert len(svg_pages) == 1
first_page = svg_pages[0]
svg_text = first_page.find('.//' + SVG_TEXT)
assert float(svg_text.attrib[SVG.Y]) == float(SOME_BASE)
def test_should_keep_text_block_structure_without_block(self):
lxml_root = E.DOCUMENT(
E.PAGE(
E.TEXT(
E.TOKEN(
SOME_TEXT,
dict_extend(COMMON_LXML_TOKEN_ATTRIBS,
{LXML.BASE: SOME_BASE})))))
svg_pages = list(iter_svg_pages_for_lxml(lxml_root))
assert len(svg_pages) == 1
first_page = svg_pages[0]
svg_text = first_page.find('.//' + SVG_TEXT)
assert svg_text is not None
assert svg_text.getparent().tag == SVG_G
assert svg_text.getparent().getparent().tag == SVG_DOC
def convert_and_annotate_lxml_content(lxml_content, xml_content, xml_mapping, name=None):
stop_watch_recorder = StopWatchRecorder()
stop_watch_recorder.start('parse lxml')
lxml_root = etree.fromstring(lxml_content)
# use a more lenient way to parse xml as xml errors are not uncomment
stop_watch_recorder.start('parse xml')
xml_root = xml_from_string_with_recover(xml_content)
stop_watch_recorder.start('extract target annotations')
target_annotations = xml_root_to_target_annotations(
xml_root,
xml_mapping
)
stop_watch_recorder.stop()
annotators = DEFAULT_ANNOTATORS + [MatchingAnnotator(
target_annotations,
use_tag_begin_prefix=True
)]
annotator = Annotator(annotators)
stop_watch_recorder.start('convert to svg')
svg_roots = list(iter_svg_pages_for_lxml(lxml_root))
stop_watch_recorder.start('annotate svg')
annotator.annotate(SvgStructuredDocument(svg_roots))
stop_watch_recorder.start('add visualisation')
svg_roots = [
visualize_svg_annotations(svg_root)
for svg_root in svg_roots
]
stop_watch_recorder.stop()
get_logger().info(
'processed: name=%s, lxml size=%s, xml size=%s, timings=[%s] (native align impl=%s)',
name, format(len(lxml_content), ','), format(len(xml_content), ','),
stop_watch_recorder, align_native_enabled
)
return svg_roots
