def scrape_adverbs(filename):
"""Gerunds (lyap) and infinitives (tum).
'tvA' gerunds come from SL_adverbs.xml.
"""
labels = ['name', 'root', 'pos', 'modification']
format_str = ','.join('{%s}' % x for x in labels)
output = []
output.append(','.join(labels))
for xml in scrape_utils.iter_xml(filename):
vu = xml.find('vu')
if vu is None:
continue
name = xml.attrib['form']
root = xml.find('s').attrib['stem']
modification = trans[vu.find('cj')[0].tag]
huet_pos = vu.find('iv')[0].tag
if huet_pos == 'abs':
pos = 'gerund'
elif huet_pos == 'inf':
pos = 'infinitive'
else:
raise Exception("Unknown POS %s" % huet_pos)
output.append(format_str.format(**{
'name': name,
'root': root,
'pos': pos,
'modification': modification or '',
}))
return '\n'.join(output)
def scrape_adverbs(filename):
"""Gerunds (ktvA).
Infinitives (tum) come from SL_final.xml, and other indeclinables
come from the MW data.
"""
labels = ['name', 'root', 'pos', 'modification']
format_str = ','.join('{%s}' % x for x in labels)
output = []
output.append(','.join(labels))
for xml in scrape_utils.iter_xml(filename):
# Only gerunds
ab = xml.find('ab')
if not ab:
continue
name = xml.attrib['form']
root = xml.find('s').attrib['stem']
modification = trans[ab.find('cj')[0].tag]
# Filter out e.g. "Asam"
if name[-2:] not in ('vA', 'ya'):
continue
output.append(
format_str.format(
**{
'name': name,
'root': root,
'pos': 'gerund',
'modification': modification or '',
}))
return '\n'.join(output)
def scrape_adverbs(filename):
"""Gerunds (ktvA).
Infinitives (tum) come from SL_final.xml, and other indeclinables
come from the MW data.
"""
labels = ['name', 'root', 'pos', 'modification']
format_str = ','.join('{%s}' % x for x in labels)
output = []
output.append(','.join(labels))
for xml in scrape_utils.iter_xml(filename):
# Only gerunds
ab = xml.find('ab')
if not ab:
continue
name = xml.attrib['form']
root = xml.find('s').attrib['stem']
modification = trans[ab.find('cj')[0].tag]
# Filter out e.g. "Asam"
if name[-2:] not in ('vA', 'ya'):
continue
output.append(format_str.format(**{
'name': name,
'root': root,
'pos': 'gerund',
'modification': modification or '',
}))
return '\n'.join(output)
def scrape(filename):
"""Inflected verbs"""
labels = ['form', 'root', 'class', 'person', 'number', 'mode', 'voice',
'modification']
rows = []
num_written = 0
for xml in scrape_utils.iter_xml(filename):
v = xml.find('v')
cj = v.find('cj')
_sys = v.find('sys')
tense = _sys[0]
np = v.find('np')
s = xml.find('s')
# Present system (present, imperfect, imperative, optative)
if tense.tag == 'prs':
vclass = tense.attrib.get('k', None)
mode = trans[tense.find('md')[0].tag]
voice = trans[tense[1].tag]
# "Tense paradigm" (future, aorist, conditional, perfect,
# injunctive, benedictive)
elif tense.tag == 'tp':
vclass = None
mode = trans[tense[0].tag]
voice = trans[tense[1].tag]
# Passive
elif tense.tag == 'pas':
vclass = None
mode = trans[tense.find('md')[0].tag]
voice = trans[tense.tag]
# Periphrastic future
elif tense.tag == 'pef':
vclass = None
mode = trans[tense.tag]
voice = trans[tense[0].tag]
else:
print ET.tostring(xml)
name = xml.attrib['form']
root = s.attrib['stem']
person = trans[np[1].tag]
number = trans[np[0].tag]
modification = trans[cj[0].tag]
# Denominative
if vclass == '11':
vclass = 'denom'
# For non-classed verb forms.
vclass = vclass or ''
rows.append((name, root, vclass, person, number, mode, voice,
modification))
num_written += 1
return labels, rows
def scrape(parts_file):
"""Participles."""
labels = ['stem', 'root', 'class', 'mode', 'voice', 'modification']
rows = []
num_written = 0
for xml in scrape_utils.iter_xml(parts_file):
form = xml.attrib['form']
root = xml.find('s').attrib['stem']
for pa in xml.findall('pa'):
# Inflectional info
na = pa.find('na')
case = trans[na[0].tag]
number = trans[na[1].tag]
gender = trans[na[2].tag]
if (gender, case, number) != ('m', '1', 's'):
continue
# Morphological info (stem)
modification = trans[pa.find('cj')[0].tag]
mode_elem = pa.find('no')[0]
mode, voice = trans[mode_elem.tag]
if (mode, voice) == ('pres', 'active'):
vclass = mode_elem[0].text
voice = trans[mode_elem[1].tag]
elif (mode, voice) in [('fut', 'active'), ('perf', 'active')]:
vclass = None
voice = trans[mode_elem[0].tag]
else:
vclass = None
# voice = default
# '11', '12', and '13' refer to "modified" verb classes. We can
# just discard these.
if vclass and modification is not None:
vclass = None
# Construct stem
if form[-1] == 's':
stem = form[:-1]
elif form.endswith('an'):
stem = form[:-1] + 't' # -an -> -at
elif mode == 'perf':
stem = form[:-2] + 'as' # -vAn -> -vas
elif mode == 'past':
stem = form[:-2] + 'at' # -vAn -> -vat
else:
# Encoding error, but high recall is OK.
stem = form
rows.append((stem, root, vclass, mode, voice, modification))
num_written += 1
return labels, rows
def scrape(parts_file):
"""Participles."""
labels = ['stem', 'root', 'class', 'mode', 'voice', 'modification']
rows = []
num_written = 0
for xml in scrape_utils.iter_xml(parts_file):
form = xml.attrib['form']
root = xml.find('s').attrib['stem']
for pa in xml.findall('pa'):
# Inflectional info
na = pa.find('na')
case = trans[na[0].tag]
number = trans[na[1].tag]
gender = trans[na[2].tag]
if (gender, case, number) != ('m', '1', 's'):
continue
# Morphological info (stem)
modification = trans[pa.find('cj')[0].tag]
mode_elem = pa.find('no')[0]
mode, voice = trans[mode_elem.tag]
if (mode, voice) == ('pres', 'active'):
vclass = mode_elem[0].text
voice = trans[mode_elem[1].tag]
elif (mode, voice) in [('fut', 'active'), ('perf', 'active')]:
vclass = None
voice = trans[mode_elem[0].tag]
else:
vclass = None
# voice = default
# '11', '12', and '13' refer to "modified" verb classes. We can
# just discard these.
if vclass and modification is not None:
vclass = None
# Construct stem
if form[-1] == 's':
stem = form[:-1]
elif form.endswith('an'):
stem = form[:-1] + 't' # -an -> -at
elif mode == 'perf':
stem = form[:-2] + 'as' # -vAn -> -vas
elif mode == 'past':
stem = form[:-2] + 'at' # -vAn -> -vat
else:
# Encoding error, but high recall is OK.
stem = form
rows.append((stem, root, vclass, mode, voice, modification))
num_written += 1
def scrape_adverbs(filename):
"""Gerunds (lyap) and infinitives (tum).
'tvA' gerunds come from SL_adverbs.xml.
"""
labels = ['name', 'root', 'pos', 'modification']
format_str = ','.join('{%s}' % x for x in labels)
output = []
output.append(','.join(labels))
for xml in scrape_utils.iter_xml(filename):
vu = xml.find('vu')
if vu is None:
continue
name = xml.attrib['form']
root = xml.find('s').attrib['stem']
modification = trans[vu.find('cj')[0].tag]
huet_pos = vu.find('iv')[0].tag
if huet_pos == 'abs':
pos = 'gerund'
elif huet_pos == 'inf':
pos = 'infinitive'
else:
raise Exception("Unknown POS %s" % huet_pos)
output.append(
format_str.format(
**{
'name': name,
'root': root,
'pos': pos,
'modification': modification or '',
}))
return '\n'.join(output)
