def add_table_with_columns(
table: str,
column_names: t.Set[str],
data: pycldf.Dataset) -> None:
"""Add a table with the given columns to the dataset.
If such a table already exists, only add the columns that do not exist
yet.
"""
delete = True
try:
data[table]
delete = False
except KeyError:
data.add_component(table)
columns = data[table].tableSchema.columns
for c in range(len(columns) - 1, -1, -1):
column = columns[c]
expected_name = "cldf_{}".format(
column.propertyUrl.uri.split("#")[-1].lower())
if expected_name not in column_names and delete:
del columns[c]
else:
column_names.remove(expected_name)
for column_name in column_names:
data.add_columns(
table,
column_name.replace(
"cldf_", "http://cldf.clld.org/v1.0/terms.rdf#"))
def from_stream(cls, stream, spec=None):
from csvw.metadata import TableGroup
cldf = Dataset(TableGroup(fname=pathlib.Path('tmp.json')))
cldf.add_component('ExampleTable')
spec = spec or CorpusSpec()
cols = cls.get_column_names(cldf)
igts = [
IGT(
id=igt[cols.id],
gloss=igt[cols.gloss].split('\\t'),
phrase=igt[cols.phrase].split('\\t'),
language=igt.get(cols.language),
properties=igt,
spec=spec,
)
for igt in reader(stream.read().splitlines(), dicts=True)]
return cls(igts, spec=spec)