def create_dpkg(top_level_dict, ev_ob_dict, directory, joint_id):
"""Create the datapackage representation.
Keyword arguments:
top_level_dict -- the dictionary with the TOP_LEVEL_INFO
ev_ob_dict -- the dictionary containing events and objects
directory -- the directory
joint_id -- the joint_identifier
"""
myDP = dp.DataPackage()
for k, v in top_level_dict.items():
myDP.descriptor[k] = v
myDP.descriptor['resources'] = []
# the events block #
key = 'events'
events_table = ev_ob_dict.get(key)
path = key + '.csv'
with io.open(directory + os.sep + key + '.csv') as stream:
headers = stream.readline().rstrip('\n').split(',')
values = csv.reader(stream)
schema = infer(headers, values, row_limit=50, primary_key=joint_id)
referenced_resource = key + 'Table'
myDP.descriptor['resources'].append({
"name": key + 'Table',
"path": path,
"schema": schema,
})
# the objects block #
key = 'objects'
objects_table = ev_ob_dict.get(key)
path = key + '.csv'
with io.open(directory + os.sep + key + '.csv') as stream:
headers = stream.readline().rstrip('\n').split(',')
values = csv.reader(stream)
schema = infer(headers, values, row_limit=50)
schema['foreignKeys'] = [{
"fields": joint_id,
"reference": {
"datapackage": "",
"resource": referenced_resource,
"fields": joint_id
}
}]
myDP.descriptor['resources'].append({
"name": key + 'Table',
"path": path,
"schema": schema,
})
return myDP
def create_datapackage(ds):
# Create datapackage based on dataset.json
dp = datapackage.DataPackage()
basepath = '{0}/{1}/{2}'.format(DIR,private_or_public(ds),ds['name'])
dp.metadata['name'] = ds['name']
dp.metadata['title'] = ds['label']
dp.metadata['description'] = ds['description']
if ds['territories']:
dp.metadata['countryCode'] = ds['territories']
dp.metadata['profiles'] = {'fiscal': '*','tabular': '*'}
dp.metadata['resources'] = [{}]
resource = dp.resources[0]
resource.metadata['name'] = 'dataset'
resource.metadata['path'] = 'dataset.csv'
# Infer schema of dataset.csv file
with io.open(basepath + '/dataset.csv') as stream:
headers = stream.readline().rstrip('\n').split(',')
values = csv.reader(stream)
schema = infer(headers, values, row_limit=1000)
resource.metadata['schema'] = schema
# Translate mapping
dp.metadata['mapping'] = transform_dataset(ds)
return dp
def pre_run(self, data_table):
if (self.schema is None) and self.infer_schema:
sample_values = data_table.get_sample(300)
self.schema = self.schema_model(jsontableschema.infer(data_table.headers, sample_values))
return True, data_table
def infer(data, row_limit, encoding, to_file):
"""Infer a schema from data.
* data must be a local filepath
* data must be CSV
* the file encoding is assumed to be UTF-8 unless an encoding is passed
with --encoding
* the first line of data must be headers
* these constraints are just for the CLI
"""
if not row_limit:
row_limit = None
with io.open(data, mode='r+t', encoding=encoding) as stream:
try:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
except UnicodeDecodeError:
response = "Could not decode the data file as {0}. " \
"Please specify an encoding to use with the " \
"--encoding argument.".format(encoding)
else:
response = jsontableschema.infer(headers, values,
row_limit=row_limit)
if to_file:
with io.open(to_file, mode='w+t', encoding='utf-8') as dest:
dest.write(json.dumps(response, ensure_ascii=False, indent=2))
click.echo(response)
def infer(data, row_limit, to_file):
"""Infer a schema from data.
* data must be a local filepath
* data must be CSV
* data must be UTF-8 encoded
* the first line of data must be headers
* these constraints are just for the CLI
"""
if not row_limit:
row_limit = None
with io.open(data, mode='r+t', encoding='utf-8') as stream:
headers = stream.readline().rstrip('\n').split(',')
values = csv.reader(stream)
response = jsontableschema.infer(headers, values, row_limit=row_limit)
if to_file:
with io.open(to_file, mode='w+t', encoding='utf-8') as dest:
dest.write(json.dumps(response, ensure_ascii=False, indent=2))
click.echo(response)
def infer_csv(csv_file, outfile, row_limit=0):
with io.open(outfile, 'w') as fp:
with io.open(csv_file) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, row_limit=row_limit)
fp.write(six.u(json.dumps(schema, indent=2, ensure_ascii=False)))
def infer(data, row_limit, encoding, to_file):
"""Infer a schema from data.
* data must be a local filepath
* data must be CSV
* the file encoding is assumed to be UTF-8 unless an encoding is passed
with --encoding
* the first line of data must be headers
* these constraints are just for the CLI
"""
if not row_limit:
row_limit = None
with io.open(data, mode='r+t', encoding=encoding) as stream:
try:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
except UnicodeDecodeError:
response = "Could not decode the data file as {0}. " \
"Please specify an encoding to use with the " \
"--encoding argument.".format(encoding)
else:
response = jsontableschema.infer(headers,
values,
row_limit=row_limit)
if to_file:
with io.open(to_file, mode='w+t', encoding='utf-8') as dest:
dest.write(json.dumps(response, ensure_ascii=False, indent=2))
click.echo(response)
def infer_from_df(df, **kwargs):
# df.iterrows does not preserve types
h = df.head()
fields = list(df)
iterrows = ([str(h[_].values[i]) for _ in fields]
for i in range(h.shape[0]))
return infer(fields, iterrows, **kwargs)
def test_infer_explicit_true(self):
filepath = os.path.join(self.data_dir, 'data_infer.csv')
with io.open(filepath) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, explicit=True)
self.assertTrue(schema['fields'][0].get('constraints'))
def test_infer_explicit_true(self):
filepath = os.path.join(self.data_dir, 'data_infer.csv')
with io.open(filepath) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, explicit=True)
self.assertTrue(schema['fields'][0].get('constraints'))
def pre_run(self, data_table):
sample_values = data_table.get_sample(300)
if (self.schema is None) and self.infer_schema:
self.schema = self.schema_model(jsontableschema.infer(data_table.headers, sample_values))
if self.schema and self.process_extra_fields:
self.extra_fields = (set(data_table.headers)).difference(set(self.schema.headers))
infered_schema = jsontableschema.infer(data_table.headers, sample_values)
complete_schema_dict = self.schema._to_python()
for field in infered_schema['fields']:
if field['name'] in self.extra_fields:
complete_schema_dict['fields'].append(copy.deepcopy(field))
self.schema = self.schema_model(complete_schema_dict)
return True, data_table
def test_infer_schema_primary_key_list(self):
primary_key = ['id', 'age']
filepath = os.path.join(self.data_dir, 'data_infer.csv')
with io.open(filepath) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, primary_key=primary_key)
schema_model = jsontableschema.models.SchemaModel(schema)
self.assertTrue(schema_model.primaryKey, primary_key)
def test_infer_schema_primary_key_list(self):
primary_key = ['id', 'age']
filepath = os.path.join(self.data_dir, 'data_infer.csv')
with io.open(filepath) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, primary_key=primary_key)
schema_model = jsontableschema.model.SchemaModel(schema)
self.assertTrue(schema_model.primaryKey, primary_key)
def test_infer_schema_row_limit(self):
filepath = os.path.join(self.data_dir, 'data_infer_row_limit.csv')
with io.open(filepath) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, row_limit=4)
schema_model = jsontableschema.models.SchemaModel(schema)
self.assertEqual(schema_model.get_field('id')['type'], 'integer')
self.assertEqual(schema_model.get_field('age')['type'], 'integer')
self.assertEqual(schema_model.get_field('name')['type'], 'string')
def test_infer_schema_row_limit(self):
filepath = os.path.join(self.data_dir, 'data_infer_row_limit.csv')
with io.open(filepath) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = jsontableschema.compat.csv_reader(stream)
schema = jsontableschema.infer(headers, values, row_limit=4)
schema_model = jsontableschema.model.SchemaModel(schema)
self.assertEqual(schema_model.get_field('id')['type'], 'integer')
self.assertEqual(schema_model.get_field('age')['type'], 'integer')
self.assertEqual(schema_model.get_field('name')['type'], 'string')
def pre_run(self, data_table):
sample_values = data_table.get_sample(300)
if (self.schema is None) and self.infer_schema:
self.schema = self.schema_model(
jsontableschema.infer(data_table.headers, sample_values))
if self.schema and self.process_extra_fields:
self.extra_fields = (set(data_table.headers)).difference(
set(self.schema.headers))
infered_schema = jsontableschema.infer(data_table.headers,
sample_values)
complete_schema_dict = self.schema._to_python()
for field in infered_schema['fields']:
if field['name'] in self.extra_fields:
complete_schema_dict['fields'].append(copy.deepcopy(field))
self.schema = self.schema_model(complete_schema_dict)
return True, data_table
def package_yearly_data():
csv_file_path = CSV_DATE_DIR + YEARLY_DATA_FILE_CSV
pkg_file_path = PKG_DATE_DIR + YEARLY_DATA_FILE_PKG
dp = datapackage.DataPackage()
dp.descriptor['name'] = 'yearly-gas-price'
dp.descriptor['title'] = 'Yearly Avg Gas Price'
with io.open(csv_file_path) as stream:
headers = stream.readline().rstrip('\n').split(',')
values = csv.reader(stream)
schema = infer(headers, values)
dp.descriptor['resources'] = [
{
'name': 'data',
'path': csv_file_path,
'schema': schema
}
]
with open(pkg_file_path, 'w') as f:
f.write(dp.to_json())
def extra_header(errors, columns, sample, infer_fields=False):
for column in copy(columns):
if 'field' not in column:
# Infer field
if infer_fields:
column_sample = []
for row in sample:
value = None
if len(row) > column['number']:
value = row[column['number']]
column_sample.append(value)
descriptor = infer([column['header']], column_sample)
column['field'] = Schema(descriptor).fields[0]
# Add error/remove column
else:
message = spec['errors']['extra-header']['message']
message = message.format(column_number=column['number'])
errors.append({
'code': 'extra-header',
'message': message,
'row-number': None,
'column-number': column['number'],
})
columns.remove(column)
def __inspect_table(self, table):
# Start timer
start = datetime.datetime.now()
# Prepare vars
errors = []
headers = None
row_number = 0
fatal_error = False
checks = copy(self.__checks)
source = table['source']
stream = table['stream']
schema = table['schema']
extra = table['extra']
# Prepare table
try:
stream.open()
sample = stream.sample
headers = stream.headers
if self.__filter_checks(checks, type='schema'):
if schema is None and self.__infer_schema:
schema = Schema(infer(headers, sample))
if schema is None:
checks = self.__filter_checks(checks,
type='schema',
inverse=True)
except Exception as exception:
fatal_error = True
message = str(exception)
if isinstance(exception, tabulator.exceptions.SourceError):
code = 'source-error'
elif isinstance(exception, tabulator.exceptions.SchemeError):
code = 'scheme-error'
elif isinstance(exception, tabulator.exceptions.FormatError):
code = 'format-error'
elif isinstance(exception, tabulator.exceptions.EncodingError):
code = 'encoding-error'
elif isinstance(exception, tabulator.exceptions.IOError):
code = 'io-error'
elif isinstance(exception, tabulator.exceptions.HTTPError):
code = 'http-error'
else:
raise
errors.append({
'row': None,
'code': code,
'message': message,
'row-number': None,
'column-number': None,
})
# Prepare columns
if not fatal_error:
columns = []
fields = [None] * len(headers)
if schema is not None:
fields = schema.fields
iterator = zip_longest(headers, fields, fillvalue=_FILLVALUE)
for number, (header, field) in enumerate(iterator, start=1):
column = {'number': number}
if header is not _FILLVALUE:
column['header'] = header
if field is not _FILLVALUE:
column['field'] = field
columns.append(column)
# Head checks
if not fatal_error:
head_checks = self.__filter_checks(checks, context='head')
for check in head_checks:
if not columns:
break
check['func'](errors, columns, sample)
for error in errors:
error['row'] = None
# Body checks
if not fatal_error:
states = {}
colmap = {column['number']: column for column in columns}
body_checks = self.__filter_checks(checks, context='body')
with stream:
for row_number, headers, row in stream.iter(extended=True):
columns = []
iterator = zip_longest(headers, row, fillvalue=_FILLVALUE)
for number, (header, value) in enumerate(iterator,
start=1):
colref = colmap.get(number, {})
column = {'number': number}
if header is not _FILLVALUE:
column['header'] = colref.get('header', header)
if 'field' in colref:
column['field'] = colref['field']
if value is not _FILLVALUE:
column['value'] = value
columns.append(column)
for check in body_checks:
if not columns:
break
state = states.setdefault(check['code'], {})
check['func'](errors, columns, row_number, state)
for error in reversed(errors):
if 'row' in error:
break
error['row'] = row
if row_number >= self.__row_limit:
break
if len(errors) >= self.__error_limit:
break
# Stop timer
stop = datetime.datetime.now()
# Compose report
errors = errors[:self.__error_limit]
report = copy(extra)
report.update({
'time': round((stop - start).total_seconds(), 3),
'valid': not bool(errors),
'error-count': len(errors),
'row-count': row_number,
'headers': headers,
'source': source,
'errors': errors,
})
return report
