def test_read_write_insecure(self):
secure_type = TypedDfBuilder("a").secure().build()
bad_type = TypedDfBuilder("a").recommended_only().build()
with pytest.raises(UnsupportedOperationError):
# noinspection HttpUrlsUsage
secure_type.read_url("http://google.com") # nosec
secure = secure_type.new_df()
bad = bad_type.new_df()
for fmt in FileFormat:
for suffix in fmt.suffixes:
try:
with tmpfile(suffix) as path:
# should always complain about insecurity FIRST
if not fmt.is_secure:
with pytest.raises(FormatInsecureError):
secure_type.read_file(path)
with pytest.raises(FormatInsecureError):
secure.write_file(path)
path.unlink(missing_ok=True)
if not fmt.is_recommended:
with pytest.raises(FormatDiscouragedError):
bad_type.read_file(path)
with pytest.raises(FormatDiscouragedError):
bad.write_file(path)
except Exception:
logger.error(f"Failed on suffix {suffix}")
raise
def test_dir_hash(self):
t = TypedDfBuilder("a").reserve("x", "y").build()
df = t.convert(pd.DataFrame([pd.Series(dict(x="cat", y="kitten"))]))
with tmpfile(".csv") as path:
hash_dir = Checksums().get_dirsum_of_file(path)
hash_dir.unlink(missing_ok=True)
df.write_file(path, dir_hash=True)
assert hash_dir.exists()
got = Checksums().load_dirsum_exact(hash_dir)
assert list(got.keys()) == [path]
hit = got[path]
assert len(hit) == 64
t.read_file(path, dir_hash=True)
t.read_file(path, hex_hash=hit)
def test_file_hash(self):
t = TypedDfBuilder("a").reserve("x", "y").build()
df = t.convert(pd.DataFrame([pd.Series(dict(x="cat", y="dog"))]))
# unfortunately, the file that gets output is os-dependent
# \n vs \r\n is an issue, so we can't check the exact hash
with tmpfile(".csv") as path:
df.write_file(path, file_hash=True)
hash_file = Checksums().get_filesum_of_file(path)
assert hash_file.exists()
got = Checksums().load_filesum_of_file(path)
assert got.file_path == path
hit = got.hash_value
assert len(hit) == 64
t.read_file(path, file_hash=True)
t.read_file(path, hex_hash=hit)
def test_attrs_hard(self):
meta = None
try:
t = TypedDfBuilder("a").reserve("x", "y").build()
df = t.convert(pd.DataFrame([pd.Series(dict(x="cat",
y="kitten"))]))
df.attrs["matrix"] = np.zeros((2, 2))
with tmpfile(".csv") as path:
df.write_file(path, attrs=True)
meta = Path(str(path) + ".attrs.json")
assert meta.exists()
df = t.read_file(path, attrs=True)
assert df.attrs == {"matrix": [["0.0", "0.0"], ["0.0", "0.0"]]}
finally:
if meta is not None:
meta.unlink(missing_ok=True)
def test_attrs(self):
meta = None
try:
t = TypedDfBuilder("a").reserve("x", "y").build()
df = t.convert(pd.DataFrame([pd.Series(dict(x="cat",
y="kitten"))]))
df.attrs["fruit"] = "apple"
with tmpfile(".csv") as path:
df.write_file(path, attrs=True)
meta = Path(str(path) + ".attrs.json")
assert meta.exists()
data = meta.read_text(encoding="utf-8").replace("\n",
"").replace(
" ", "")
assert data == '{"fruit": "apple"}'
df = t.read_file(path, attrs=True)
assert df.attrs == {"fruit": "apple"}
finally:
if meta is not None:
meta.unlink(missing_ok=True)
