def test_merge_error_not_raised():
m1 = Metadata.from_dict({
"name":
"merge_test",
"columns": [
{
"name": "c1",
"type": "string"
},
{
"name": "c2",
"type": "int64"
},
],
"partitions": ["c2"],
})
m2 = Metadata.from_dict({
"name":
"merge_test",
"columns": [
{
"name": "c2",
"type": "int64"
},
{
"name": "c3",
"type": "string"
},
],
"partitions": ["c2"],
})
expected = Metadata.from_dict({
"name":
"merge_test",
"columns": [
{
"name": "c1",
"type": "string"
},
{
"name": "c2",
"type": "int64"
},
{
"name": "c3",
"type": "string"
},
],
"partitions": ["c2"],
})
merged = Metadata.merge(m1, m2)
assert (sorted(x.items() for x in merged.columns) == sorted(
x.items() for x in expected.columns) and merged.name == expected.name
and sorted(merged.partitions) == sorted(expected.partitions))
def test_preservation_of_underlying_metadata():
# Test if additional data is preserved
test_dict = {
"name": "test",
"description": "test",
"file_format": "test",
"sensitive": False,
"columns": [{
"name": "test",
"type": "null"
}],
"primary_key": ["test"],
"partitions": ["test"],
"additional-attr": "test",
}
meta = Metadata.from_dict(test_dict)
out_dict = meta.to_dict()
for k, v in test_dict.items():
assert out_dict[k] == v
# make sure data is copied and not just a pointer
assert id(test_dict) != id(meta._data)
test_dict["columns"] = [{"name": "new_test", "type": "bool_"}]
assert test_dict != meta.columns
# Assert Metadata instances are different
m1 = Metadata()
m2 = Metadata()
assert m1.columns == m2.columns
m1.columns.append({"name": "new_test", "type": "bool_"})
assert m1.columns != m2.columns
def test_generate_from_meta(spec_name, serde_name, expected_file_name):
md = Metadata.from_dict({
"name":
"test_table",
"file_format":
spec_name,
"columns": [
{
"name": "my_int",
"type": "int64",
"description": "This is an integer",
},
{
"name": "my_double",
"type": "float64"
},
{
"name": "my_date",
"type": "date64"
},
{
"name": "my_decimal",
"type": "decimal128(10,2)"
},
{
"name": "my_timestamp",
"type": "timestamp(s)",
"description": "Partition column",
},
],
"partitions": ["my_timestamp"],
})
gc = GlueConverter()
if spec_name == "csv":
gc.options.set_csv_serde(serde_name)
if spec_name == "json":
gc.options.set_json_serde(serde_name)
opts = GlueConverterOptions(default_db_base_path="s3://bucket/",
default_db_name="test_db")
gc_default_opts = GlueConverter(opts)
table_path = "s3://bucket/test_table"
# DO DICT TEST
spec = gc.generate_from_meta(md,
database_name="test_db",
table_location=table_path)
spec_default_opts = gc_default_opts.generate_from_meta(md, )
assert spec == spec_default_opts
with open(f"tests/data/glue_converter/{expected_file_name}.json") as f:
expected_spec = json.load(f)
assert spec == expected_spec
def metadata(self, meta_dict: dict):
try:
meta_obj = Metadata.from_dict(meta_dict)
meta_obj.set_col_type_category_from_types()
self._metadata = meta_obj.to_dict()
if "file_format" not in self.metadata:
raise ValidationError(
"metadata given must have a file_format property")
except ValidationError as e:
error_msg = ("Pandas validator requires schemas that conform "
"to those found in the mojap_metadata package. "
f"Metadata given failed validation: {str(e)}")
raise ValidationError(error_msg)
def test_generate_from_meta():
md = Metadata.from_dict({
"name":
"test_table",
"file_format":
"test-format",
"columns": [
{
"name": "my_int",
"type": "int64",
"description": "This is an integer",
"nullable": False,
},
{
"name": "my_double",
"type": "float64",
"nullable": True
},
{
"name": "my_date",
"type": "date64"
},
{
"name": "my_decimal",
"type": "decimal128(10,2)"
},
{
"name": "my_timestamp",
"type": "timestamp(s)",
"description": "Partition column",
},
],
"partitions": ["my_timestamp"],
})
ac = ArrowConverter()
assert isinstance(ac.options, BaseConverterOptions)
schema1 = ac.generate_from_meta(md)
schema2 = ac.generate_from_meta(md, False)
assert isinstance(schema1, pa.Schema)
assert isinstance(schema2, pa.Schema)
schema_str1 = ("my_int: int64 not null\nmy_double: double\n"
"my_date: date64[ms]\nmy_decimal: decimal(10, 2)")
schema_str2 = schema_str1 + "\nmy_timestamp: timestamp[s]"
assert schema1.to_string() == schema_str1
assert schema2.to_string() == schema_str2
def test_infer_file_format_from_meta(file_format, expected):
meta = {
"name": "test",
"columns": [{
"name": "a",
"type": "string"
}],
"file_format": file_format,
}
actual = infer_file_format_from_meta(meta)
assert actual == expected
metadata = Metadata.from_dict(meta)
actual = infer_file_format_from_meta(metadata)
assert actual == expected
def test_from_dict():
test_dict = {
"name": "test",
"description": "test",
"file_format": "test",
"sensitive": False,
"columns": [{
"name": "test",
"type": "null"
}],
"primary_key": ["test"],
"partitions": ["test"],
}
meta = Metadata.from_dict(test_dict)
for k, v in test_dict.items():
assert getattr(meta, k) == v
meta.name = "bob"
assert meta.name == meta._data["name"]
def test_file_reader_works_with_both_meta_types():
csv_meta = {
"columns": [
{
"name": "test",
"type_category": "string"
},
{
"name": "a_column",
"type_category": "string"
},
]
}
df_csv1 = reader.csv.read("tests/data/example_data.csv", csv_meta)
csv_meta = Metadata.from_dict(csv_meta)
df_csv2 = reader.csv.read("tests/data/example_data.csv", csv_meta)
assert_frame_equal(df_csv1, df_csv2)
with pytest.raises(ValidationError):
m = {"columns": [{"name": "broken", "type": "not-a-type"}]}
reader.csv.read("tests/data/example_data.jsonl", metadata=m)
def get_meta(ff: str, additional_params: Optional[dict] = None):
additional_params = {} if not additional_params else additional_params
md = Metadata.from_dict({
"name":
"test_table",
"file_format":
ff,
"columns": [
{
"name": "my_int",
"type": "int64",
"description": "This is an integer",
},
{
"name": "my_double",
"type": "float64"
},
{
"name": "my_date",
"type": "date64"
},
{
"name": "my_decimal",
"type": "decimal128(10,2)"
},
{
"name": "my_complex",
"type": "large_list<int64>"
},
{
"name": "my_timestamp",
"type": "timestamp(s)",
"description": "Partition column",
},
],
"partitions": ["my_timestamp"],
**additional_params,
})
return md
def test_to_from_json_yaml(tmpdir, writer):
path_file = tmpdir.mkdir("test_outputs").join("meta.{writer}")
test_dict = {
"name": "test",
"description": "test",
"file_format": "test",
"sensitive": False,
"columns": [{
"name": "test",
"type": "null"
}],
"primary_key": ["test"],
"partitions": ["test"],
}
meta = Metadata.from_dict(test_dict)
# test in/out reader
getattr(meta, f"to_{writer}")(str(path_file))
read_meta = getattr(Metadata, f"from_{writer}")(str(path_file))
out_dict = read_meta.to_dict()
for k, v in test_dict.items():
assert out_dict[k] == v
def cast_pandas_table_to_schema(
df: pd.DataFrame,
metadata: Union[Metadata, dict],
ignore_columns: List = None,
drop_columns: List = None,
pd_integer: bool = True,
pd_string: bool = True,
pd_boolean: bool = True,
pd_date_type: str = "datetime_object",
pd_timestamp_type: str = "datetime_object",
bool_map: Union[Callable, dict] = None,
num_error_map: dict = None,
):
"""
Casts the columns in dataframe provided to the meta data dictionary provided.
Safest casting occurs when all coltypes of input dataframe are strings.
df: Pandas dataframe
meta: Metadata or dict representation of metadata
ignore_columns: a list of column names to not cast to the meta data dictionary.
These columns are remained unchanged.
drop_columns: Removes these columns from the dataframe
bool_map (Callable, dict, optional): A custom mapping function that is applied
to str cols to be converted to booleans before conversion to boolean type.
e.g. {"Yes": True, "No": False}. If not set bool values are inferred by the
_default_str_bool_mapper.
"""
default_num_errors = "raise"
if ignore_columns is None:
ignore_columns = []
if drop_columns is None:
drop_columns = []
if isinstance(metadata, Metadata):
meta = metadata.to_dict()
elif isinstance(metadata, dict):
if "columns" not in metadata:
raise KeyError('metadata missing a "columns" key')
_ = Metadata.from_dict(metadata) # Check metadata is valid
meta = deepcopy(metadata)
else:
error_msg = ("Input metadata must be of type Metadata "
f"or dict got {type(metadata)}")
raise ValueError(error_msg)
df = df.copy()
all_exclude_cols = ignore_columns + drop_columns
meta_cols_to_convert = [
c for c in meta["columns"] if c["name"] not in all_exclude_cols
or c["name"] not in meta["partitions"]
]
for c in meta_cols_to_convert:
# Null first if applicable
if c["name"] not in df.columns:
raise ValueError(f"Column '{c['name']}' not in df")
else:
# must get num_errors from either meta or num_error_map. Meta has precedence
if c.get("num_errors"):
num_errors = c.get("num_errors")
elif isinstance(num_error_map, dict):
num_errors = num_error_map.get(c["name"])
else:
num_errors = default_num_errors
df[c["name"]] = cast_pandas_column_to_schema(
df[c["name"]],
metacol=c,
pd_integer=pd_integer,
pd_string=pd_string,
pd_boolean=pd_boolean,
pd_date_type=pd_date_type,
pd_timestamp_type=pd_timestamp_type,
num_errors=num_errors,
bool_map=bool_map,
)
final_cols = [
c["name"] for c in meta["columns"]
if c["name"] not in drop_columns or c["name"] not in meta["partitions"]
]
df = df[final_cols]
return df
def test_generate_from_meta():
md = Metadata.from_dict({
"name":
"test_table",
"file_format":
"test-format",
"columns": [
{
"name": "my_int",
"type": "int64",
"description": "This is an integer",
"nullable": False,
},
{
"name": "my_double",
"type": "float64",
"nullable": True
},
{
"name": "my_date",
"type": "date64"
},
{
"name": "my_decimal",
"type": "decimal128(10,2)"
},
{
"name": "my_timestamp",
"type": "timestamp(s)",
"description": "Partition column",
},
],
"partitions": ["my_timestamp"],
})
ac = ArrowConverter()
assert isinstance(ac.options, BaseConverterOptions)
schema1 = ac.generate_from_meta(md)
schema2 = ac.generate_from_meta(md, False)
assert isinstance(schema1, pa.Schema)
assert isinstance(schema2, pa.Schema)
expected_names = ["my_int", "my_double", "my_date", "my_decimal"]
expected_types = [
pa.int64(),
pa.float64(),
pa.date64(),
pa.decimal128(10, 2)
]
assert schema1.names == expected_names
checks1 = [a.equals(e) for a, e in zip(schema1.types, expected_types)]
assert all(checks1)
# Do schema2 assertions
expected_names.append("my_timestamp")
expected_types.append(pa.timestamp("s"))
assert schema2.names == expected_names
checks2 = [a.equals(e) for a, e in zip(schema2.types, expected_types)]
assert all(checks2)
# Also check specific type properties
assert schema2.field("my_decimal").type.precision == 10
assert schema2.field("my_decimal").type.scale == 2
assert schema2.field("my_timestamp").type.unit == "s"
def test_set_col_types_from_type_category():
test_dict = {
"name":
"test",
"description":
"test",
"file_format":
"test",
"sensitive":
False,
"columns": [
{
"name": "test_null",
"type_category": "null"
},
{
"name": "test_integer",
"type_category": "integer"
},
{
"name": "test_float",
"type_category": "float"
},
{
"name": "test_string",
"type_category": "string"
},
{
"name": "test_timestamp",
"type_category": "timestamp"
},
{
"name": "test_binary",
"type_category": "binary"
},
{
"name": "test_boolean",
"type_category": "boolean"
},
{
"name": "test_list",
"type_category": "list"
},
{
"name": "test_struct",
"type_category": "struct"
},
],
}
meta = Metadata.from_dict(test_dict)
with pytest.warns(UserWarning):
meta.set_col_types_from_type_category()
for c in meta.columns:
default_type_cat = c["name"].replace("test_", "")
expected_type = meta.default_type_category_lookup.get(default_type_cat)
assert c["type"] == expected_type
new_dict = {
"null": "null",
"integer": "uint8",
"float": "decimal128(2,5)",
"string": "large_string",
"timestamp": "timestamp(us)",
"binary": "large_binary",
"boolean": "bool_",
"list": "large_list<null>",
"struct": "map_<null>",
}
meta2 = Metadata.from_dict(test_dict)
meta2.set_col_types_from_type_category(
lambda x: new_dict.get(x["type_category"]))
for c in meta2.columns:
default_type_cat = c["name"].replace("test_", "")
assert c["type"] == new_dict.get(default_type_cat)
def test_generate_from_meta():
md = Metadata.from_dict(
{
"name": "test_table",
"file_format": "csv",
"description": "A test table",
"columns": [
{
"name": "my_int",
"type": "int64",
"description": "This is an integer",
"nullable": False,
"minimum": 10,
},
{"name": "my_double", "type": "float64", "nullable": True},
{"name": "my_date", "type": "date64"},
{"name": "my_decimal", "type": "decimal128(10,2)"},
{"name": "my_string", "type": "string", "enum": ["cat", "dog"]},
{
"name": "my_timestamp",
"type": "timestamp(s)",
"description": "Partition column",
},
],
"partitions": ["my_timestamp"],
}
)
expected1 = {
"$schema": "https://moj-analytical-services.github.io/metadata_schema/table/v1.4.0.json", # noqa: 401
"name": "test_table",
"data_format": "csv",
"location": "test_table/",
"description": "A test table",
"columns": [
{
"name": "my_int",
"type": "long",
"description": "This is an integer",
"nullable": False,
"minimum": 10,
},
{
"name": "my_double",
"type": "double",
"nullable": True,
"description": "",
},
{"name": "my_date", "type": "date", "description": ""},
{"name": "my_decimal", "type": "decimal(10,2)", "description": ""},
{
"name": "my_string",
"type": "character",
"enum": ["cat", "dog"],
"description": "",
},
{
"name": "my_timestamp",
"type": "datetime",
"description": "Partition column",
},
],
"partitions": ["my_timestamp"],
}
emc = EtlManagerConverter()
assert isinstance(emc.options, BaseConverterOptions)
etl1 = emc.generate_from_meta(md)
assert isinstance(etl1, TableMeta)
etl1 = etl1.to_dict()
assert etl1 == expected1
expected2 = copy.deepcopy(expected1)
# Remove additional cols not native to etl_manager
del expected2["columns"][0]["minimum"]
etl2 = emc.generate_from_meta(md, include_extra_column_params=False).to_dict()
assert etl2 == expected2
# Check table_location works
expected3 = copy.deepcopy(expected1)
expected3["location"] = "some/new/tablepath/"
etl3 = emc.generate_from_meta(md, table_location="some/new/tablepath/").to_dict()
assert etl3 == expected3
# Check file_format_mapper
mapper = {"csv": "csv_quoted_nodate"}.get
expected4 = copy.deepcopy(expected1)
expected4["data_format"] = "csv_quoted_nodate"
etl4 = emc.generate_from_meta(md, file_format_mapper=mapper).to_dict()
assert etl4 == expected4
# Check glue_specific
mapper = {"csv": "csv_quoted_nodate"}.get
expected5 = copy.deepcopy(expected1)
gs = {
"Parameters": {"skip.header.line.count": "1"},
"StorageDescriptor": {"Parameters": {"skip.header.line.count": "1"}},
}
expected5["glue_specific"] = gs
etl5 = emc.generate_from_meta(md, glue_specific=gs).to_dict()
assert etl5 == expected5
0.0,
[],
(),
],
)
def test_inferred_input_fails(fake_input):
with pytest.raises(TypeError):
Metadata.from_infer(fake_input)
merge_meta_test = Metadata.from_dict({
"name":
"merge_test",
"columns": [{
"name": "c1",
"type": "int64"
}, {
"name": "c2",
"type": "string"
}],
"partitions": ["c1"],
})
merge_meta_diff_col_type = Metadata.from_dict({
"name":
"merge_test",
"columns": [
{
"name": "c3",
"type": "string"
},
{
def _parse_data_to_pandas(filepath: str, table_params: dict, metadata: dict):
"""
Reads in the data from the given filepath and returns
a dataframe
"""
meta_col_names = [
c["name"] for c in metadata["columns"]
if c["name"] not in metadata.get("partitions", [])
]
# For string based file types convert make arrow readers read them in as strings
# validators will still treat these as dates but will run validation against strings
# cols expecting values to match a timestamp format
if "json" in metadata["file_format"] or "csv" in metadata["file_format"]:
md_obj = Metadata.from_dict(metadata)
cols = md_obj.columns
cols_to_force_str_read_in = []
for c in cols:
if c["type"].startswith("time") or c["type"].startswith("date"):
c["type"] = "string"
c["type_category"] = "string"
cols_to_force_str_read_in.append(c["name"])
md_obj.columns = cols
ac = ArrowConverter()
arrow_schema = ac.generate_from_meta(md_obj)
ts_as_str_schema = pa.schema([])
for cname in cols_to_force_str_read_in:
ts_as_str_schema = ts_as_str_schema.append(
arrow_schema.field(cname))
# Set the reader type
if filepath.startswith("s3://"):
reader_fs = fs.S3FileSystem(region="eu-west-1")
fp_for_file_reader = filepath.replace("s3://", "", 1)
else:
reader_fs = fs.LocalFileSystem()
fp_for_file_reader = filepath
with reader_fs.open_input_stream(fp_for_file_reader) as f:
if "csv" in metadata["file_format"]:
# Safer CSV load for newlines_in_values set to True
if table_params.get("expect-header", True):
po = csv.ParseOptions(newlines_in_values=True)
else:
po = csv.ParseOptions(newlines_in_values=True,
column_names=meta_col_names)
if ts_as_str_schema:
co = csv.ConvertOptions(column_types=ts_as_str_schema)
else:
co = None
df = pa_read_csv_to_pandas(
input_file=f,
schema=arrow_schema,
expect_full_schema=False,
parse_options=po,
convert_options=co,
)
# dates/datetimes == string
elif "json" in metadata["file_format"]:
po = json.ParseOptions(
newlines_in_values=True,
explicit_schema=ts_as_str_schema if ts_as_str_schema else None,
)
df = pa_read_json_to_pandas(
input_file=f,
schema=arrow_schema,
expect_full_schema=False,
parse_options=po,
)
# dates/datetimes == string
elif "parquet" in metadata["file_format"]:
df = arrow_to_pandas(pq.read_table(f))
# dates/datetimes == datetime / date
else:
raise ValueError(
f"Unknown file_format in metadata: {metadata['file_format']}.")
if table_params.get("row-limit"):
df = df.sample(table_params.get("row-limit"))
if table_params.get("headers-ignore-case"):
df_cols = [c.lower() for c in df.columns]
df.columns = df_cols
if table_params.get("only-test-cols-in-metadata", False):
keep_cols = [c for c in df.columns if c in meta_col_names]
df = df[keep_cols]
return df
