def test_ds_extraction(path=None):
skip_if_no_module('libxmp')
ds = Dataset(path).create()
copy(testpath, path)
ds.save()
assert_repo_status(ds.path)
res = extract_metadata(
types=['xmp'],
dataset=ds,
# artificially disable extraction from any file in the dataset
files=[])
assert_result_count(
res, 1,
type='dataset', status='ok', action='metadata', path=path, refds=ds.path)
assert_in('xmp', res[0]['metadata'])
# now the more useful case: getting everything for xmp from a dataset
res = extract_metadata(
types=['xmp'],
dataset=ds)
assert_result_count(res, 2)
assert_result_count(
res, 1,
type='dataset', status='ok', action='metadata', path=path, refds=ds.path)
assert_result_count(
res, 1,
type='file', status='ok', action='metadata', path=opj(path, 'xmp.pdf'),
parentds=ds.path)
for r in res:
assert_in('xmp', r['metadata'])
def test_completion(out_fn=None):
skip_if_no_module('argcomplete')
from datalad.cmd import WitlessRunner
runner = WitlessRunner()
def get_completions(s: str, expected) -> list:
"""Run 'datalad' external command and collect completions
Parameters
----------
s: str
what to append to 'datalad ' invocation
expected: iterable of str
What entries to expect - would raise AssertionError if any is
not present in output
exit_code: int, optional
If incomplete/malformed we seems to get 2, most frequently used
so default
Returns
-------
list of str
Entries output
"""
if os.path.exists(out_fn): # reuse but ensure it is gone
os.unlink(out_fn)
comp_line = f'datalad {s}'
runner.run(
comp_line.split(' '),
env=dict(
os.environ,
_ARGCOMPLETE='1',
_ARGCOMPLETE_STDOUT_FILENAME=out_fn,
COMP_LINE=comp_line,
# without -1 seems to get "finished completion", someone can investigate more
COMP_POINT=str(len(comp_line) - 1), # always at the end ATM
))
with open(out_fn, 'rb') as f:
entries = f.read().split(b'\x0b')
entries = [e.decode() for e in entries]
diff = set(expected).difference(entries)
if diff:
raise AssertionError(
f"Entries {sorted(diff)} were expected but not found in the completion output: {entries}"
)
return entries # for extra analyzes if so desired
all_commands = get_all_commands()
get_completions('i', {'install'})
get_completions(' ', ['--dbg', '-c'] + all_commands)
# if command already matches -- we get only that hit ATM, not others which begin with it
get_completions('create', ['create '])
get_completions('create -', ['--dataset'])
# but for incomplete one we do get all create* commands
get_completions('creat',
[c for c in all_commands if c.startswith('create')])
def test_file_extraction(path=None):
skip_if_no_module('libxmp')
# go into virgin dir to avoid detection of any dataset
with chpwd(path):
res = extract_metadata(
types=['xmp'],
files=[testpath])
assert_result_count(res, 1, type='file', status='ok', action='metadata', path=testpath)
assert_in('xmp', res[0]['metadata'])
def test_check_dates_invalid_date():
skip_if_no_module("dateutil")
with swallow_outputs() as cmo:
assert_raises(IncompleteResultsError,
check_dates, [],
reference_date="not a valid date",
return_type="list")
out = cmo.out
# The error makes it through the standard renderer.
assert_in('"status": "error"', out)
def test_check_dates(path=None):
skip_if_no_module("dateutil")
ref_ts = 1218182889 # Fri, 08 Aug 2008 04:08:09 -0400
refdate = "@{}".format(ref_ts)
repo = os.path.join(path, "repo")
with set_date(ref_ts + 5000):
ar = AnnexRepo(repo)
ar.add(".")
ar.commit()
# The standard renderer outputs json.
with swallow_outputs() as cmo:
# Set level to WARNING to avoid the progress bar when
# DATALAD_TESTS_UI_BACKEND=console.
with swallow_logs(new_level=logging.WARNING):
check_dates([repo], reference_date=refdate, return_type="list")
assert_in("report", json.loads(cmo.out).keys())
# We find the newer objects.
newer = call([path], reference_date=refdate)
eq_(len(newer), 1)
ok_(newer[0]["report"]["objects"])
# There are no older objects to find.
older = call([repo], reference_date=refdate, older=True)
assert_false(older[0]["report"]["objects"])
# We can pass the date in RFC 2822 format.
assert_dict_equal(
newer[0],
call([path], reference_date="08 Aug 2008 04:08:09 -0400")[0])
# paths=None defaults to the current directory.
with chpwd(path):
assert_dict_equal(
newer[0]["report"],
call(paths=None, reference_date=refdate)[0]["report"])
# Only commit type is present when annex='none'.
newer_noannex = call([path], reference_date=refdate, annex="none")
for entry in newer_noannex[0]["report"]["objects"].values():
ok_(entry["type"] == "commit")
def testish2():
skip_if_no_module("datalad")
return "magic"
def testish():
skip_if_no_module("nonexistingforsuremodule")
raise ValueError
def test_wtf(topdir=None):
path = opj(topdir, OBSCURE_FILENAME)
# smoke test for now
with swallow_outputs() as cmo:
wtf(dataset=path, on_failure="ignore")
assert_not_in('## dataset', cmo.out)
assert_in('## configuration', cmo.out)
# Those sections get sensored out by default now
assert_not_in('user.name: ', cmo.out)
with chpwd(path):
with swallow_outputs() as cmo:
wtf()
assert_not_in('## dataset', cmo.out)
assert_in('## configuration', cmo.out)
# now with a dataset
ds = create(path)
with swallow_outputs() as cmo:
wtf(dataset=ds.path)
assert_in('## configuration', cmo.out)
assert_in('## dataset', cmo.out)
assert_in(u'path: {}'.format(ds.path), ensure_unicode(cmo.out))
assert_in('branches', cmo.out)
assert_in(DEFAULT_BRANCH + '@', cmo.out)
assert_in('git-annex@', cmo.out)
# and if we run with all sensitive
for sensitive in ('some', True):
with swallow_outputs() as cmo:
wtf(dataset=ds.path, sensitive=sensitive)
# we fake those for tests anyways, but we do show cfg in this mode
# and explicitly not showing them
assert_in('user.name: %s' % _HIDDEN, cmo.out)
with swallow_outputs() as cmo:
wtf(dataset=ds.path, sensitive='all')
assert_not_in(_HIDDEN, cmo.out) # all is shown
assert_in('user.name: ', cmo.out)
# Sections selection
#
# If we ask for no sections and there is no dataset
with chpwd(path):
with swallow_outputs() as cmo:
wtf(sections=[])
assert_not_in('## dataset', cmo.out)
for s in SECTION_CALLABLES:
assert_not_in('## %s' % s.lower(), cmo.out.lower())
# ask for a selected set
secs = ['git-annex', 'configuration']
with chpwd(path):
with swallow_outputs() as cmo:
wtf(sections=secs)
for s in SECTION_CALLABLES:
(assert_in if s in secs else assert_not_in)('## %s' %
s.lower(),
cmo.out.lower())
# order should match our desired one, not alphabetical
# but because of https://github.com/datalad/datalad/issues/3915
# alphanum is now desired
assert cmo.out.index('## git-annex') > cmo.out.index(
'## configuration')
# not achievable from cmdline is to pass an empty list of sections.
with chpwd(path):
with swallow_outputs() as cmo:
wtf(sections=[])
eq_(cmo.out.rstrip(), '# WTF')
# and we could decorate it nicely for embedding e.g. into github issues
with swallow_outputs() as cmo:
wtf(sections=['dependencies'], decor='html_details')
ok_startswith(cmo.out,
'<details><summary>DataLad %s WTF' % __version__)
assert_in('## dependencies', cmo.out)
# short flavor
with swallow_outputs() as cmo:
wtf(flavor='short')
assert_in("- datalad: version=%s" % __version__, cmo.out)
assert_in("- dependencies: ", cmo.out)
eq_(len(cmo.out.splitlines()),
4) # #WTF, datalad, dependencies, trailing new line
with swallow_outputs() as cmo:
wtf(flavor='short', sections='*')
assert_greater(len(cmo.out.splitlines()), 10) # many more
# check that wtf of an unavailable section yields impossible result (#6712)
res = wtf(sections=['murkie'], on_failure='ignore')
eq_(res[0]["status"], "impossible")
# should result only in '# WTF'
skip_if_no_module('pyperclip')
# verify that it works correctly in the env/platform
import pyperclip
with swallow_outputs() as cmo:
try:
pyperclip.copy("xxx")
pyperclip_works = pyperclip.paste().strip() == "xxx"
wtf(dataset=ds.path, clipboard=True)
except (AttributeError, pyperclip.PyperclipException) as exc:
# AttributeError could come from pyperclip if no DISPLAY
raise SkipTest(str(exc))
assert_in("WTF information of length", cmo.out)
assert_not_in('user.name', cmo.out)
if not pyperclip_works:
# Some times does not throw but just fails to work
raise SkipTest(
"Pyperclip seems to be not functioning here correctly")
assert_not_in('user.name', pyperclip.paste())
assert_in(_HIDDEN, pyperclip.paste()) # by default no sensitive info
assert_in("cmd:annex:", pyperclip.paste()) # but the content is there