def test_stderr_bad_encoding(crx_bytes):
"""The errors and warnings can include cited garbage from invalid files
and we should handle that gracefully."""
with pytest.raises(HatanakaException) as excinfo:
crx2rnx(crx_bytes[:-1] + b'\xFF\0')
msg = excinfo.value.args[0]
assert msg.startswith('The file seems to be truncated')
assert msg.endswith('\\xff<end')
def test_non_ascii_is_tolerated(rnx_bytes, crx_bytes):
def add_non_ascii(txt):
return txt.replace(
b'VERSION / TYPE',
b'VERSION / TYPE' + 'õäü'.encode().ljust(60) + b'COMMENT\n')
converted = crx2rnx(add_non_ascii(crx_bytes))
assert clean(converted) == clean(add_non_ascii(rnx_bytes))
converted = rnx2crx(add_non_ascii(rnx_bytes))
assert clean(converted) == clean(add_non_ascii(crx_bytes))
def test_invalid():
with pytest.raises(HatanakaException) as excinfo:
crx2rnx('blah')
msg = excinfo.value.args[0]
assert msg.startswith('The file seems to be truncated in the middle.')
def test_crx2rnx_extra_args_good(rnx_str, crx_str):
converted = crx2rnx(crx_str, skip_strange_epochs=True)
assert clean(converted) == clean(rnx_str)
def test_crx2rnx_bytes_stream(crx_bytes_stream, rnx_bytes):
assert clean(crx2rnx(crx_bytes_stream)) == clean(rnx_bytes)
def test_crx2rnx_str_stream(crx_str_stream, rnx_str):
assert crx2rnx(crx_str_stream) == rnx_str
def test_crx2rnx_bytes(crx_bytes, rnx_bytes):
assert clean(crx2rnx(crx_bytes)) == clean(rnx_bytes)
def test_crx2rnx_str(crx_str, rnx_str):
assert crx2rnx(crx_str) == rnx_str
def opener(fn: T.TextIO | Path, header: bool = False) -> T.Iterator[T.TextIO]:
"""provides file handle for regular ASCII or gzip files transparently"""
if isinstance(fn, str):
fn = Path(fn).expanduser()
if isinstance(fn, io.StringIO):
fn.seek(0)
yield fn
elif isinstance(fn, Path):
# need to have this check for Windows
if not fn.is_file():
raise FileNotFoundError(fn)
finf = fn.stat()
if finf.st_size > 100e6:
logging.info(f"opening {finf.st_size/1e6} MByte {fn.name}")
suffix = fn.suffix.lower()
if suffix == ".gz":
with gzip.open(fn, "rt") as f:
_, is_crinex = rinex_version(first_nonblank_line(f))
f.seek(0)
if is_crinex and not header:
# Conversion to string is necessary because of a quirk where gzip.open()
# even with 'rt' doesn't decompress until read.
f = io.StringIO(crx2rnx(f.read()))
yield f
elif suffix == ".bz2":
# this is for plain bzip2 files, NOT tar.bz2, which requires f.seek(512)
with bz2.open(fn, "rt") as f:
_, is_crinex = rinex_version(first_nonblank_line(f))
f.seek(0)
if is_crinex and not header:
f = io.StringIO(crx2rnx(f.read()))
yield f
elif suffix == ".zip":
with zipfile.ZipFile(fn, "r") as z:
flist = z.namelist()
for rinexfn in flist:
with z.open(rinexfn, "r") as bf:
f = io.StringIO(
io.TextIOWrapper(
bf, encoding="ascii",
errors="ignore").read() # type: ignore
)
yield f
elif suffix == ".z":
if unlzw is None:
raise ImportError("pip install unlzw3")
with fn.open("rb") as zu:
with io.StringIO(unlzw(zu.read()).decode("ascii")) as f:
yield f
else: # assume not compressed (or Hatanaka)
with fn.open("r", encoding="ascii", errors="ignore") as f:
_, is_crinex = rinex_version(first_nonblank_line(f))
f.seek(0)
if is_crinex and not header:
f = io.StringIO(crx2rnx(f))
yield f
else:
raise OSError(f"Unsure what to do with input of type: {type(fn)}")