def include_array_schemas(columns, schema):
schema['definitions'] = copy.deepcopy(BASE_RECURSIVE_SCHEMAS)
decimal_array_columns = [key for key, value in columns.items() if value.sql_data_type == 'numeric[]']
for col in decimal_array_columns:
scale = post_db.numeric_scale(columns[col])
precision = post_db.numeric_precision(columns[col])
schema_name = schema_name_for_numeric_array(precision, scale)
schema['definitions'][schema_name] = {'type': ['null', 'number', 'array'],
'multipleOf': post_db.numeric_multiple_of(scale),
'exclusiveMaximum': True,
'maximum': post_db.numeric_max(precision, scale),
'exclusiveMinimum': True,
'minimum': post_db.numeric_min(precision, scale),
'items': {'$ref': '#/definitions/{}'.format(schema_name)}}
return schema
def include_array_schemas(columns, schema):
schema["definitions"] = copy.deepcopy(BASE_RECURSIVE_SCHEMAS)
decimal_array_columns = [
key for key, value in columns.items()
if value.sql_data_type == "numeric[]"
]
for c in decimal_array_columns:
scale = post_db.numeric_scale(columns[c])
precision = post_db.numeric_precision(columns[c])
schema_name = schema_name_for_numeric_array(precision, scale)
schema["definitions"][schema_name] = {
"type": ["null", "number", "array"],
"multipleOf": post_db.numeric_multiple_of(scale),
"exclusiveMaximum": True,
"maximum": post_db.numeric_max(precision, scale),
"exclusiveMinimum": True,
"minimum": post_db.numeric_min(precision, scale),
"items": {
"$ref": "#/definitions/{}".format(schema_name)
},
}
return schema
def schema_for_column_datatype(c):
schema = {}
#remove any array notation from type information as we use a separate field for that
data_type = c.sql_data_type.lower().replace('[]', '')
if data_type in INTEGER_TYPES:
schema['type'] = nullable_column('integer', c.is_primary_key)
schema['minimum'] = -1 * (2**(c.numeric_precision - 1))
schema['maximum'] = 2**(c.numeric_precision - 1) - 1
return schema
if data_type == 'money':
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if c.is_enum:
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if data_type == 'bit' and c.character_maximum_length == 1:
schema['type'] = nullable_column('boolean', c.is_primary_key)
return schema
if data_type == 'boolean':
schema['type'] = nullable_column('boolean', c.is_primary_key)
return schema
if data_type == 'uuid':
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if data_type == 'hstore':
schema['type'] = nullable_column('object', c.is_primary_key)
schema['properties'] = {}
return schema
if data_type == 'citext':
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if data_type in JSON_TYPES:
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if data_type == 'numeric':
schema['type'] = nullable_column('number', c.is_primary_key)
scale = post_db.numeric_scale(c)
precision = post_db.numeric_precision(c)
schema['exclusiveMaximum'] = True
schema['maximum'] = post_db.numeric_max(precision, scale)
schema['multipleOf'] = post_db.numeric_multiple_of(scale)
schema['exclusiveMinimum'] = True
schema['minimum'] = post_db.numeric_min(precision, scale)
return schema
if data_type in {'time without time zone', 'time with time zone'}:
#times are treated as ordinary strings as they can not possible match RFC3339
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if data_type in ('date', 'timestamp without time zone',
'timestamp with time zone'):
schema['type'] = nullable_column('string', c.is_primary_key)
schema['format'] = 'date-time'
return schema
if data_type in FLOAT_TYPES:
schema['type'] = nullable_column('number', c.is_primary_key)
return schema
if data_type == 'text':
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
if data_type == 'character varying':
schema['type'] = nullable_column('string', c.is_primary_key)
if c.character_maximum_length:
schema['maxLength'] = c.character_maximum_length
return schema
if data_type == 'character':
schema['type'] = nullable_column('string', c.is_primary_key)
if c.character_maximum_length:
schema['maxLength'] = c.character_maximum_length
return schema
if data_type in {'cidr', 'inet', 'macaddr'}:
schema['type'] = nullable_column('string', c.is_primary_key)
return schema
return schema
def schema_for_column(c):
#NB> from the post postgres docs: The current implementation does not enforce the declared number of dimensions either.
#these means we can say nothing about an array column. its items may be more arrays or primitive types like integers
#and this can vary on a row by row basis
column_schema = {'type': ["null", "array"]}
if not c.is_array:
return schema_for_column_datatype(c)
if c.sql_data_type == 'integer[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_integer_array'
}
elif c.sql_data_type == 'bit[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_boolean_array'
}
elif c.sql_data_type == 'boolean[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_boolean_array'
}
elif c.sql_data_type == 'character varying[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'cidr[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'citext[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'date[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_timestamp_array'
}
elif c.sql_data_type == 'numeric[]':
scale = post_db.numeric_scale(c)
precision = post_db.numeric_precision(c)
schema_name = schema_name_for_numeric_array(precision, scale)
column_schema['items'] = {
'$ref': '#/definitions/{}'.format(schema_name)
}
elif c.sql_data_type == 'double precision[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_number_array'
}
elif c.sql_data_type == 'hstore[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_object_array'
}
elif c.sql_data_type == 'inet[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'json[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'jsonb[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'mac[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'money[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'real[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_number_array'
}
elif c.sql_data_type == 'smallint[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_integer_array'
}
elif c.sql_data_type == 'text[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'timestamp without time zone[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_timestamp_array'
}
elif c.sql_data_type == 'timestamp with time zone[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_timestamp_array'
}
elif c.sql_data_type == 'time[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
elif c.sql_data_type == 'uuid[]':
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
else:
#custom datatypes like enums
column_schema['items'] = {
'$ref': '#/definitions/sdc_recursive_string_array'
}
return column_schema
def schema_for_column_datatype(c):
schema = {}
# remove any array notation from type information as we use a separate field for that
data_type = c.sql_data_type.lower().replace("[]", "")
if data_type in INTEGER_TYPES:
schema["type"] = nullable_column("integer", c.is_primary_key)
schema["minimum"] = -1 * (2**(c.numeric_precision - 1))
schema["maximum"] = 2**(c.numeric_precision - 1) - 1
return schema
if data_type == "money":
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if c.is_enum:
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type == "bit" and c.character_maximum_length == 1:
schema["type"] = nullable_column("boolean", c.is_primary_key)
return schema
if data_type == "boolean":
schema["type"] = nullable_column("boolean", c.is_primary_key)
return schema
if data_type == "uuid":
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type == "hstore":
schema["type"] = nullable_column("object", c.is_primary_key)
schema["properties"] = {}
return schema
if data_type == "citext":
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type in JSON_TYPES:
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type == "numeric":
schema["type"] = nullable_column("number", c.is_primary_key)
scale = post_db.numeric_scale(c)
precision = post_db.numeric_precision(c)
schema["exclusiveMaximum"] = True
schema["maximum"] = post_db.numeric_max(precision, scale)
schema["multipleOf"] = post_db.numeric_multiple_of(scale)
schema["exclusiveMinimum"] = True
schema["minimum"] = post_db.numeric_min(precision, scale)
return schema
if data_type in {"time without time zone", "time with time zone"}:
# times are treated as ordinary strings as they can not possible match RFC3339
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type in ("date", "timestamp without time zone",
"timestamp with time zone"):
schema["type"] = nullable_column("string", c.is_primary_key)
schema["format"] = "date-time"
return schema
if data_type in FLOAT_TYPES:
schema["type"] = nullable_column("number", c.is_primary_key)
return schema
if data_type == "text":
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type == "character varying":
schema["type"] = nullable_column("string", c.is_primary_key)
if c.character_maximum_length:
schema["maxLength"] = c.character_maximum_length
return schema
if data_type == "character":
schema["type"] = nullable_column("string", c.is_primary_key)
if c.character_maximum_length:
schema["maxLength"] = c.character_maximum_length
return schema
if data_type in {"cidr", "inet", "macaddr"}:
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
if data_type == "geometry":
schema["type"] = nullable_column("string", c.is_primary_key)
return schema
def schema_for_column(c):
# NB> from the post postgres docs: The current implementation does not enforce the declared
# number of dimensions either. This means we can say nothing about an array column. Its items
# may be more arrays or primitive types like integers and this can vary on a row by row basis
column_schema = {"type": ["null", "array"]}
if not c.is_array:
return schema_for_column_datatype(c)
if c.sql_data_type == "integer[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_integer_array"
}
elif c.sql_data_type == "bigint[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_integer_array"
}
elif c.sql_data_type == "bit[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_boolean_array"
}
elif c.sql_data_type == "boolean[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_boolean_array"
}
elif c.sql_data_type == "character varying[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "cidr[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "citext[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "date[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_timestamp_array"
}
elif c.sql_data_type == "numeric[]":
scale = post_db.numeric_scale(c)
precision = post_db.numeric_precision(c)
schema_name = schema_name_for_numeric_array(precision, scale)
column_schema["items"] = {
"$ref": "#/definitions/{}".format(schema_name)
}
elif c.sql_data_type == "double precision[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_number_array"
}
elif c.sql_data_type == "hstore[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_object_array"
}
elif c.sql_data_type == "inet[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "json[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "jsonb[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "mac[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "money[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "real[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_number_array"
}
elif c.sql_data_type == "smallint[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_integer_array"
}
elif c.sql_data_type == "text[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "timestamp without time zone[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_timestamp_array"
}
elif c.sql_data_type == "timestamp with time zone[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_timestamp_array"
}
elif c.sql_data_type == "time[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "uuid[]":
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
elif c.sql_data_type == "bytea[]":
# bytea[] are unsupported
del column_schema["type"]
else:
# custom datatypes like enums
column_schema["items"] = {
"$ref": "#/definitions/sdc_recursive_string_array"
}
return column_schema
