def test_dont_convert_non_df_to_pandas():
"""
Check that if we just give a number or a string or something else,
we get an ApiException
"""
with pytest.raises(ApiException) as exception_int:
x_conv = convert_to_pandas(345)
with pytest.raises(ApiException) as exception_str:
x_conv = convert_to_pandas("abc")
def test_execute_pd_concat():
"""
given an input dict of value assignments and a code snippet,
substitute the values in, and evaluate.
"""
input_code = "z = pd.concat([x,y],join='outer', ignore_index=True, sort=True)"
input_vals = {
"x": [{
"a": 1,
"b": 2
}, {
"a": 2,
"b": 3
}],
"y": [{
"b": 4,
"c": 2
}, {
"b": 5,
"c": 7
}]
}
output_hash = "somehash"
file_contents = {}
return_targets = find_assignments(input_code)["targets"]
result_dict = execute_code(file_contents, input_code, input_vals,
return_targets, output_hash)
result = result_dict["results"]
assert (len(result) == 1) # only one output of function
assert (isinstance(result['z'], bytes))
result_df = convert_to_pandas(result['z'])
assert (result_df.size == 12) ## 4 rows * 3 columns
def test_json_to_pandas_to_json():
"""
convert a json dataframe into a pandas one,
and back again, and check we get the same one
back.
"""
d_orig = [{"Col1": 123, "Col2": "Abc"}, {"Col1": 456, "Col2": "Def"}]
pd_df = convert_to_pandas(d_orig)
assert (isinstance(pd_df, pd.DataFrame))
d_new = json.loads(convert_from_pandas(pd_df, max_size_json=1024))
assert (d_orig == d_new)
def test_convert_null_to_nan():
"""
check we get NaN in the Pandas DF when we have null in the JSON, in a column that has other numbers
"""
json_string = '[{"a": 1, "b": 33},{"a": 2, "b": null}]'
json_obj = json.loads(json_string)
## should be converted to None
assert (json_obj[1]["b"] == None)
df = convert_to_pandas(json_obj)
## will now be NaN
assert (np.isnan(df["b"][1]))
## but when we convert it back into json, we want it to be None
new_json = json.loads(convert_from_pandas(df, max_size_json=1024))
assert (new_json[1]["b"] == None)
def test_execute_simple_func():
"""
import numpy, and define a trivial function in the code snippet, which is
then used when filling a dataframe
"""
input_code = 'import numpy\ndef squareroot(x):\n return numpy.sqrt(x)\n\ndf= pd.DataFrame({\"a\":[numpy.sqrt(9),squareroot(16),13],\"b\":[14,15,16]})'
input_vals = {}
file_contents = {}
return_targets = find_assignments(input_code)["targets"]
output_hash = "somehash"
result_dict = execute_code(file_contents, input_code, input_vals,
return_targets, output_hash)
result = result_dict["results"]
assert (result)
assert (isinstance(result, dict))
assert ("df" in result.keys())
pddf = convert_to_pandas(result["df"])
assert (isinstance(pddf, pd.DataFrame))
assert (pddf["a"][0] == 3)
assert (pddf["a"][1] == 4)
