def _dataconvert(args):
# What is the type of input file?
if args.format:
intype = args.format
else:
intype = guess_type(args.inpath)
# What is the type of output file?
outtype = guess_type(args.outpath)
# If outtype is ARFF then we need to guess field-types.
# Thus we overwrite the args.guess_types to True.
if outtype == arff.MIMETYPE:
args.guess_types = True
if is_url_path(args.inpath):
instream = urllib2.urlopen(args.inpath)
else:
instream = open(args.inpath)
# tsv_types = ['tsv', 'text/tsv', 'text/tab-separated-values']
if intype in ['text/csv', 'csv']:
records, metadata = dcsv.parse(instream, guess_types=args.guess_types)
elif intype in [
'application/vnd.ms-excel',
'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',
'xls'
]:
import dataconverters.xls
excel_type = 'xls' if intype == 'application/vnd.ms-excel' else 'xlsx'
records, metadata = dataconverters.xls.parse(
instream,
excel_type=excel_type,
sheet=args.sheet,
guess_types=args.guess_types,
encoding=args.encoding)
else:
raise ValueError(
'No support for reading file type %s - support for csv or xls only at present'
% intype)
if args.outpath.startswith('_.'):
outstream = sys.stdout
else:
outstream = open(args.outpath, 'w')
if (args.records):
records = itertools.islice(records, int(args.records))
if outtype == 'text/csv':
dcsv.write(outstream, records, metadata)
elif outtype == 'application/json':
import dataconverters.jsondata as js
js.write(outstream, records, metadata)
elif outtype == arff.MIMETYPE:
arff.write(outstream, records, metadata)
else:
raise ValueError('Only support writing to csv and json at present')
def _dataconvert(args):
# What is the type of input file?
if args.format:
intype = args.format
else:
intype = guess_type(args.inpath)
# What is the type of output file?
outtype = guess_type(args.outpath)
# If outtype is ARFF then we need to guess field-types.
# Thus we overwrite the args.guess_types to True.
if outtype == arff.MIMETYPE:
args.guess_types = True
if is_url_path(args.inpath):
instream = urllib2.urlopen(args.inpath)
else:
instream = open(args.inpath)
tsv_types = ['tsv', 'text/tsv', 'text/tab-separated-values']
if intype in ['text/csv', 'csv'] + tsv_types:
records, metadata = dcsv.parse(instream, guess_types=args.guess_types)
elif intype in ['application/vnd.ms-excel',
'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',
'xls'
]:
import dataconverters.xls
excel_type = 'xls' if intype == 'application/vnd.ms-excel' else 'xlsx'
records, metadata = dataconverters.xls.parse(instream,
excel_type=excel_type,
sheet=args.sheet,
guess_types=args.guess_types,
encoding=args.encoding
)
else:
raise ValueError(
'No support for reading file type %s - support for csv or xls only at present' % intype)
if args.outpath.startswith('_.'):
outstream = sys.stdout
else:
outstream = open(args.outpath, 'w')
if (args.records):
records = itertools.islice(records, int(args.records))
if outtype == 'text/csv':
dcsv.write(outstream, records, metadata)
elif outtype == 'application/json':
import dataconverters.jsondata as js
js.write(outstream, records, metadata)
elif outtype == arff.MIMETYPE:
arff.write(outstream, records, metadata)
else:
raise ValueError('Only support writing to csv and json at present')
def test_1(self):
metadata = {"fields": [{"type": "Integer", "id": u"temperature"}, {"type": "String", "id": u"place"}]}
records = [
{"place": u"Cairo", "temperature": 32},
{"place": u"Alexandria", "temperature": 22},
{"place": u"Aswan", "temperature": 42},
]
desired_results = """@RELATION testdataset\n
@ATTRIBUTE temperature NUMERIC
@ATTRIBUTE place STRING\n
@DATA\n32,'Cairo'
22,'Alexandria'
42,'Aswan'
"""
out = StringIO()
arff.write(out, records, metadata, dataset_name="testdataset")
out.seek(0)
result = out.read()
assert_equal(result, desired_results)
def test_1(self):
metadata = {
'fields': [{
'type': 'Integer',
'id': u'temperature'
}, {
'type': 'String',
'id': u'place'
}]
}
records = [
{
'place': u'Cairo',
'temperature': 32
},
{
'place': u'Alexandria',
'temperature': 22
},
{
'place': u'Aswan',
'temperature': 42
},
]
desired_results = """@RELATION testdataset\n
@ATTRIBUTE temperature NUMERIC
@ATTRIBUTE place STRING\n
@DATA\n32,'Cairo'
22,'Alexandria'
42,'Aswan'
"""
out = StringIO()
arff.write(out, records, metadata, dataset_name='testdataset')
out.seek(0)
result = out.read()
assert_equal(result, desired_results)
