def mock_get_ft(*args, **kwargs):
data = open(EPMC_FT, "r").read()
xml = etree.fromstring(data)
l = xml.xpath("//license")
l[0].getparent().remove(l[0])
s = etree.tostring(xml)
return epmc.EPMCFullText(s)
def mock_get_ft(*args, **kwargs):
data = open(EPMC_FT, "r").read()
xml = etree.fromstring(data)
aids = xml.xpath("//article-id[@pub-id-type='manuscript']")
aids[0].getparent().remove(aids[0])
s = etree.tostring(xml)
return epmc.EPMCFullText(s)
def test_07_ft_info(self):
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
data = open(EPMC_FT, "r").read()
ft = epmc.EPMCFullText(data)
workflow.extract_fulltext_info(msg, ft)
assert record.has_ft_xml is True
assert len(record.provenance) == 2
assert record.aam is True
assert record.aam_from_xml is True
def mock_get_ft(*args, **kwargs):
data = open(EPMC_FT, "r").read()
return epmc.EPMCFullText(data)
def test_08_ft_licence(self):
data = open(EPMC_FT, "r").read()
xml = etree.fromstring(data)
l = xml.xpath("//license")
lp = l[0].find("license-p")
# licence in type attribute
l[0].set(
"license-type", "cc by"
) # note the missing "-"; to test the licence representation variations at the same time
l[0].set("{http://www.w3.org/1999/xlink}href", "http://random.url")
lp.clear()
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type == "cc-by"
assert record.licence_source == "epmc_xml"
assert len(record.provenance) == 1
# licence in href attribute
l[0].set("license-type", "open access")
l[0].set("{http://www.w3.org/1999/xlink}href",
"http://creativecommons.org/licenses/by-nd/3.0")
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type == "cc-by-nd"
assert record.licence_source == "epmc_xml"
assert len(record.provenance) == 1
# licence in text
l[0].set("license-type", "open access")
l[0].set("{http://www.w3.org/1999/xlink}href", "http://random.url")
lp.text = "licence is <a href='http://creativecommons.org/licenses/by-nc-nd/3.0'>http://creativecommons.org/licenses/by-nc-nd/3.0</a>"
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type == "cc-by-nc-nd"
assert record.licence_source == "epmc_xml"
assert len(record.provenance) == 1
# licence in /second/ licence paragraph
lp.text = "some waffle"
lp2 = etree.SubElement(l[0], "license-p")
lp2.text = "licence is <a href='http://creativecommons.org/licenses/by/3.0'>http://creativecommons.org/licenses/by/3.0</a>"
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type == "cc-by"
assert record.licence_source == "epmc_xml"
assert len(record.provenance) == 1
# licence in words in text
l[0].set("license-type", "open access")
l[0].set("{http://www.w3.org/1999/xlink}href", "http://random.url")
lp.text = "This is a Creative Commons Attribution-NonCommercial licenced article"
l[0].remove(lp2)
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type == "cc-by-nc"
assert record.licence_source == "epmc_xml"
assert len(record.provenance) == 1
# licence present but unrecognised
lp.text = "wibble wibble wobble"
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type == "non-standard-licence"
assert record.licence_source == "epmc_xml"
assert len(record.provenance) == 1
# no licence element present
p = l[0].getparent()
p.remove(l[0])
s = etree.tostring(xml)
ft = epmc.EPMCFullText(s)
record = models.Record()
msg = workflow.WorkflowMessage(record=record)
workflow.extract_fulltext_licence(msg, ft)
assert record.licence_type is None
assert record.licence_source is None
assert len(record.provenance) == 0
