def margin(span):
"""Enrich a margin match."""
data = {
'subpart': 'margin',
'margin': squash([t.lower_ for t in span
if t.ent_type_ in {'margin_term', 'surface'}])}
return data
def convert(token):
"""Build a collector trait"""
names = regex.split(r'\s*(?:and|with|[,&])\s*',
token.group.get('col_name'))
traits = []
for name, suffix in zip_longest(names, names[1:], fillvalue=''):
name = regex.sub(r'\.{3,}.*', '', name)
if len(name) < MIN_LEN:
continue
trait = Trait(start=token.start, end=token.end)
trait.col_name = name
if suffix.lower() in name_parts.SUFFIXES:
trait.col_name = f'{name} {suffix}'
if name.lower() not in name_parts.SUFFIXES:
traits.append(trait)
if not traits:
return None
if token.group.get('collector_no'):
col_no = token.group['collector_no']
# Temp hack
if col_no[-1] in ('m', 'M'):
return None
traits[0].col_no = col_no
return squash(traits)
def surface(span):
"""Enrich a surface match."""
data = {}
fields = {
'present': set(),
'subpart': set(),
'surface': [],
}
for token in span:
label = token.ent_type_
if label in ('part', 'location'):
data[label] = REPLACE.get(token.lower_, token.lower_)
elif label == 'surface':
fields['surface'].append(REPLACE.get(token.lower_, token.lower_))
elif label == 'subpart':
fields['subpart'].add(REPLACE.get(token.lower_, token.lower_))
elif token.lower_ in PRESENT:
fields['present'].add(PRESENCE.get(token.lower_, False))
if len(fields['subpart']) > 1:
fields['subpart'] -= {'surface'}
fields = {k: squash(v) for k, v in fields.items() if fields[k]}
data = {**data, **fields}
return data
def as_value(token, trait, value_field="number", unit_field="units"):
"""Convert token values and units to trait fields."""
units = as_list(token.group.get(unit_field, []))
trait.units = squash(units) if units else None
values = []
for i, val in enumerate(as_list(token.group.get(value_field, []))):
val = to_positive_float(val)
if val is None:
return False
if i < len(units):
unit = units[i]
else:
unit = units[-1] if units else None
values.append(convert_units(val, unit))
if not values:
return False
trait.value = squash(values)
trait.units_inferred = not bool(trait.units)
return True
def shape(span):
"""Enrich a shape match."""
data = {
'shape':
squash([
REPLACE.get(t.lower_, t.lower_) for t in span
if t.ent_type_ == 'shape'
])
}
if field := [t.lower_ for t in span if t.ent_type_ == 'part']:
data['part'] = field[0]
def compound(token):
"""Handle a pattern like: 4 ft 9 in."""
trait = Trait(start=token.start, end=token.end)
trait.units = [token.group["feet"], token.group["inches"]]
trait.units_inferred = False
trait.is_flag_missing(token, "key", rename="ambiguous_key")
fts = convert_units(to_positive_float(token.group["ft"]), "ft")
ins = [
convert_units(to_positive_float(i), "in")
for i in as_list(token.group["in"])
]
value = [round(fts + i, 2) for i in ins]
trait.value = squash(value)
add_flags(token, trait)
return trait
def compound(token):
"""Convert a compound weight like: 2 lbs. 3.1 - 4.5 oz."""
trait = Trait(start=token.start, end=token.end)
trait.units = [token.group["pounds"], token.group["ounces"]]
trait.units_inferred = False
trait.is_flag_missing(token, "key", rename="ambiguous_key")
lbs = convert_units(to_positive_float(token.group["lbs"]), "lbs")
ozs = [
convert_units(to_positive_float(oz), "ozs")
for oz in as_list(token.group["ozs"])
]
value = [round(lbs + oz, 2) for oz in ozs]
trait.value = squash(value)
add_flags(token, trait)
return trait
def transfer(self, token, names):
"""Move fields from a token to the trait if they exist in the token."""
for name in names:
if name in token.group:
values = [v.lower() for v in as_list(token.group[name])]
setattr(self, name, squash(values))