def _completeness(self, data):
accessor = JT(data) # can go direct if elements are always present
#organ = accessor.query('meta', 'organ')
try:
organ = adops.get(data, ['meta', 'organ'])
except:
organ = None
if isinstance(organ, list) or isinstance(organ, tuple):
if len(organ) == 1:
organ, = organ
organ = OntTerm(organ)
else:
organ = [OntTerm(o) for o in organ]
elif organ == 'othertargets':
pass
elif organ:
organ = OntTerm(organ)
return (
accessor.status.submission_index,
accessor.status.curation_index,
accessor.status.error_index,
#accessor.submission_completeness_index,
#dataset.name, # from filename (do we not have that in meta!?)
accessor.query('meta', 'folder_name'),
accessor.id, #if 'id' in dowe else None,
accessor.query('meta', 'award_number'),
organ,
)
def award_number(self, value):
_, s = self.c.award_number(value)
yield s, a, owl.NamedIndividual
yield s, a, TEMP.FundedResearchProject
return
o = self.integrator.organ(value)
if o:
if o != 'othertargets':
o = OntId(o)
if o.prefix == 'FMA':
ot = OntTerm(o)
o = next(OntTerm.query(label=ot.label, prefix='UBERON')).OntTerm
yield s, isAbout, o.u
def added(self):
data = super().added
# FIXME conditional lifts ...
if 'award_number' not in data['meta']:
am = self.lifters.award_manual
if am:
data['meta']['award_number'] = am
if 'modality' not in data['meta']:
m = self.lifters.modality
if m:
data['meta']['modality'] = m
if 'organ' not in data['meta']:
if 'award_number' in data['meta']:
an = data['meta']['award_number']
o = self.lifters.organ(an)
if o:
if o != 'othertargets':
o = OntId(o)
if o.prefix == 'FMA':
ot = OntTerm(o)
o = next(
OntTerm.query(label=ot.label,
prefix='UBERON')).OntTerm
data['meta']['organ'] = o
if 'organ' not in data['meta'] or data['meta'][
'organ'] == 'othertargets':
o = self.lifters.organ_term
if o:
if isinstance(o, str):
o = o,
out = tuple()
for _o in o:
_o = OntId(_o)
if _o.prefix == 'FMA':
ot = OntTerm(_o)
_o = next(
OntTerm.query(label=ot.label,
prefix='UBERON')).OntTerm
out += (_o, )
data['meta']['organ'] = out
return data
def triples_protcur(self, protocol_subject):
# XXX deprecated and extremely slow
# triples from the protcur pipeline are exported only once as part of
# protcur.ttl
ps = list(self._protcur(str(protocol_subject)))
anatomy = [(p,
OntId('UBERON:' +
str(p).split('UBERON:', 1)[-1].split(' ', 1)[0]))
for p in ps if p.astType == 'protc:input'
and '(protc:input (term UBERON' in str(p)]
#breakpoint()
dataset_subject = rdflib.URIRef(self.uri_api)
yield protocol_subject, TEMP.hasNumberOfProtcurAnnotations, rdflib.Literal(
len(ps))
done = set()
for anno, term in anatomy:
if term in done:
continue
done.add(term)
yield from OntTerm(term).triples_simple
o = term.u
t = dataset_subject, TEMP.involvesAnatomicalRegion, o
sl = rdflib.URIRef(anno.shareLink)
av = (((ilxtr.annotationValue, rdflib.Literal(anno.value)), )
if anno.value != o else tuple())
notes = [(ilxtr.curatorNote, rdflib.Literal(n))
for n in anno.curatorNotes]
prov = [(ilxtr.hasAnnotation, sl)]
yield t
yield from cmb.annotation(t, *av, *notes, *prov)()
def process_dict(self, dict_):
""" deal with multiple fields """
out = {k:v for k, v in dict_.items() if k not in self.skip}
for h_unit, h_value in zip(self.h_unit, self.h_value):
if h_value not in dict_: # we drop null cells so if one of these was null then we have to skip it here too
continue
dhv = dict_[h_value]
if isinstance(dhv, str):
try:
dhv = ast.literal_eval(dhv)
except ValueError as e:
raise exc.UnhandledTypeError(f'{h_value} {dhv!r} was not parsed!') from e
compose = dhv * pyru.ur.parse_units(dict_[h_unit])
#_, v, rest = parameter_expression(compose)
#out[h_value] = str(UnitsParser(compose).for_text) # FIXME sparc repr
#breakpoint()
out[h_value] = compose #UnitsParser(compose).asPython()
if 'gender' in out and 'species' in out:
if out['species'] != OntTerm('NCBITaxon:9606'):
out['sex'] = out.pop('gender')
return out
def normv(v):
if is_list_or_tuple(v):
return [normv(_) for _ in v]
if isinstance(v, dict):
return {k: normv(v) for k, v in v.items()}
if isinstance(v, str) and v.startswith('http'):
# needed for loading from json that has been serialized
# rather than from our internal representation
# probably better to centralized the reload ...
v = OntTerm(v)
return v.asCell()
if isinstance(
v,
rdflib.URIRef): # FIXME why is this getting converted early?
ot = OntTerm(v)
return ot.asCell()
if isinstance(v, ProtcurExpression):
return str(v) # FIXME for xml?
if isinstance(v, Quantity):
return str(v)
elif isinstance(v, pathlib.Path):
return str(v)
elif isinstance(v, idlib.Stream):
return v.asCell()
#elif isinstance(v, list) or isinstance(v, str):
#return v
elif isinstance(v, BaseException):
return repr(v)
else:
#loge.debug(repr(v))
return v
def _update_technique(self, cell):
# NOTE some rows won't update if the dataset no longer exists
value = cell.value
if value:
try:
term = next(OntTerm.query(label=value))
cell.value = term.asCellHyperlink()
except StopIteration:
log.info(f'no term for technique {value}')
def map(self, anno):
row = self._annotation_row(anno)
mapping_ok = row.mapping_ok().value == 'TRUE' # FIXME
not_input = row.not_input_().value
bad_for_mapping = row.bad_for_mapping_().value
manual_mapping = row.manual_mapping().value
if mapping_ok and not not_input:
pass
if manual_mapping and ' ' in manual_mapping:
log.error(
f'Why does a manual mapping have a space in it {manual_mapping!r}'
)
elif manual_mapping:
return OntTerm(manual_mapping)
elif mapping_ok: # FIXME anno.astValue can drift from auto_mapping
# this is so hilariously inefficient, we parse the same stuff
# 3 times or something
return OntTerm(anno.asPython().asPython().black_box.curie)
def normv(v):
if isinstance(v, str) and v.startswith('http'):
# needed for loading from json that has been serialized
# rather than from our internal representation
# probably better to centralized the reload ...
oid = OntId(v)
if oid.prefix in want_prefixes:
return OntTerm(v).tabular()
else:
return oid.iri
if isinstance(v, OntId):
if not isinstance(v, OntTerm):
v = OntTerm(v)
v = v.tabular()
if isinstance(v, list) or isinstance(v, tuple):
v = ','.join(
json.dumps(_, cls=JEncode) if isinstance(_, dict
) else normv(_)
for _ in v)
v = v.replace('\n', ' ').replace('\t', ' ')
elif any(isinstance(v, c) for c in (int, float, str)):
v = str(v)
v = v.replace('\n',
' ').replace('\t',
' ') # FIXME tests to catch this
elif isinstance(v, dict):
v = json.dumps(v, cls=JEncode)
return v
def yield_from_id(s, matid, predicate=predicate):
mat = rm[matid]
if 'external' in mat:
mat_s = OntTerm(mat['external'][0])
yield s, predicate, mat_s.u
yield mat_s.u, a, owl.Class
yield mat_s.u, rdfs.label, rdflib.Literal(mat_s.label)
if 'materials' in mat:
for submat_id in mat['materials']:
yield from yield_from_id(mat_s, submat_id, TEMP.hasConstituent)
else:
log.warning(f'no external id for {mat}')
def query(value, prefix):
for query_type in ('term', 'search'):
terms = [q.OntTerm for q in OntTerm.query(prefix=prefix, **{query_type:value})]
if terms:
#print('matching', terms[0], value)
#print('extra terms for', value, terms[1:])
return terms[0]
else:
continue
else:
log.warning(f'No ontology id found for {value}')
return value
def normv(v):
if is_list_or_tuple(v):
return [normv(_) for _ in v]
elif isinstance(v, dict):
return {k:normv(v) for k, v in v.items()}
elif isinstance(v, str) and v.startswith('http'):
# needed for loading from json that has been serialized
# rather than from our internal representation
# probably better to centralized the reload ...
# XXX NOTE these days this will only happen if someone
# supplies us with a uri in a field where we aren't
# expecting one, in which case we should just return it
try:
v = OntTerm(v)
return v.asCell()
except Exception as e:
loge.error(f'something went wrong with {v}')
loge.exception(e)
return v
#raise e
elif isinstance(v, rdflib.URIRef): # FIXME why is this getting converted early?
ot = OntTerm(v)
return ot.asCell()
elif isinstance(v, ProtcurExpression):
return str(v) # FIXME for xml?
elif isinstance(v, Quantity):
return str(v)
elif isinstance(v, AsJson): # XXX returns value not tested, may be extremely strange
return str(v)
elif isinstance(v, pathlib.Path):
return str(v)
elif isinstance(v, idlib.Stream):
return v.asCell()
#elif isinstance(v, list) or isinstance(v, str):
#return v
elif isinstance(v, BaseException):
return repr(v)
else:
#loge.debug(repr(v))
return v
def normv(v):
if isinstance(v, str) and v.startswith('http'):
# needed for loading from json that has been serialized
# rather than from our internal representation
# probably better to centralized the reload ...
v = OntTerm(v)
return v.tabular()
if isinstance(v, rdflib.URIRef): # FIXME why is this getting converted early?
ot = OntTerm(v)
return ot.tabular()
if isinstance(v, Expr):
return str(v) # FIXME for xml?
if isinstance(v, Quantity):
return str(v)
else:
#log.debug(repr(v))
return v
def _term(self):
if not hasattr(self, '_c_term'):
self._c_term = OntTerm(self.id)
return self._c_term
def disco(self):
#dsh = sorted(MetaOutSchema.schema['allOf'][0]['properties'])
dsh = [
'acknowledgements',
'additional_links',
'award_number',
'completeness_of_data_set',
'contributor_count',
'description',
'dirs',
'errors',
'examples',
'files',
'funding',
'keywords',
'links',
'modality',
'name', # -> title
'organ',
'originating_article_doi',
'principal_investigator',
'prior_batch_number',
'protocol_url_or_doi',
'sample_count',
'size',
'species',
'subject_count',
'title_for_complete_data_set',
'uri_api',
'uri_human',
'error_index', # (sum *_index)
'dataset_completeness_index', # dead
'is_about',
'involves_anatomical_region',
'title',
'folder_name',
]
chs = [
'contributor_affiliation',
'contributor_orcid_id',
'contributor_role',
'is_contact_person',
'name',
'first_name',
'last_name',
'middle_name',
'id',
'blackfynn_user_id',
]
datasets = [['id', 'submission_index', 'curation_index'] + dsh]
contributors = [['id'] + chs]
subjects = [['id', 'blob']]
errors = [['id', 'blob']]
resources = [['id', 'blob']]
#cje = JEncode()
def normv(v):
if isinstance(v, str) and v.startswith('http'):
# needed for loading from json that has been serialized
# rather than from our internal representation
# probably better to centralized the reload ...
oid = OntId(v)
if oid.prefix in want_prefixes:
return OntTerm(v).tabular()
else:
return oid.iri
if isinstance(v, OntId):
if not isinstance(v, OntTerm):
v = OntTerm(v)
v = v.tabular()
if isinstance(v, list) or isinstance(v, tuple):
v = ','.join(
json.dumps(_, cls=JEncode) if isinstance(_, dict
) else normv(_)
for _ in v)
v = v.replace('\n', ' ').replace('\t', ' ')
elif any(isinstance(v, c) for c in (int, float, str)):
v = str(v)
v = v.replace('\n',
' ').replace('\t',
' ') # FIXME tests to catch this
elif isinstance(v, dict):
v = json.dumps(v, cls=JEncode)
return v
for dataset_blob in self:
id = dataset_blob['id']
dowe = dataset_blob
graph = rdflib.Graph()
TriplesExportDataset(dataset_blob).populate(graph)
is_about = [
OntTerm(o) for s, o in graph[:isAbout:]
if isinstance(o, rdflib.URIRef)
]
involves = [
OntTerm(o) for s, o in graph[:TEMP.involvesAnatomicalRegion:]
]
inv = ','.join(i.tabular() for i in involves)
ia = ','.join(a.tabular() for a in is_about)
#row = [id, dowe['error_index'], dowe['submission_completeness_index']] # FIXME this doubles up on the row
row = [
id, dowe['status']['submission_index'],
dowe['status']['curation_index']
] # FIXME this doubles up on the row
if 'meta' in dowe:
meta = dowe['meta']
for k in dsh:
if k in meta:
v = meta[k]
v = normv(v)
elif k == 'is_about':
v = ia
elif k == 'involves_anatomical_region':
v = inv
else:
v = None
row.append(v)
else:
row += [None for k in sc.MetaOutSchema.schema['properties']]
datasets.append(row)
# contribs
if 'contributors' in dowe:
cs = dowe['contributors']
for c in cs:
row = [id]
for k in chs:
if k in c:
v = c[k]
v = normv(v)
row.append(v)
else:
row.append(None)
contributors.append(row)
if 'subjects' in dowe:
for subject in dowe['subjects']:
row = [id]
row.append(json.dumps(subject, cls=JEncode))
subjects.append(row)
# moved to resources if exists already
#if 'software' in sbs:
#for software in sbs['software']:
#row = [id]
#row.append(json.dumps(software, cls=JEncode))
#resources.append(row)
if 'resources' in dowe:
for res in dowe['resources']:
row = [id]
row.append(json.dumps(res, cls=JEncode))
resources.append(row)
if 'errors' in dowe:
ers = get_all_errors(dowe)
for er in ers:
row = [id]
row.append(json.dumps(er, cls=JEncode))
errors.append(row)
# TODO samples resources
return (('datasets', datasets), ('contributors', contributors),
('subjects', subjects), ('resources', resources), ('errors',
errors))
def map(self, anno):
row = self._annotation_row(anno)
oid = row.ontology_id().value
label = row.ontology_label().value
ilx_curie = row.interlex_id().value
return OntTerm(oid, label=label), OntTerm(interlex_curie)
def triples_gen(self):
rm = self._source
# FIXME there doesn't seem to be a section that tells me the name
# of top level model so I have to know its name beforhand
# the id is in the model, having the id in the resource map
# prevents issues if these things get sent decoupled
id = rm['id']
mid = id.replace(' ', '-')
links = rm[id]['links']
#linknodes = [n for n in rm[id]['nodes'] if n['class'] == 'Link'] # visible confusion
st = []
from_to = []
ot = None
yield from self.apinatbase()
for link in links:
if 'conveyingType' in link:
if link['conveyingType'] == 'ADVECTIVE':
p_is = TEMP.isAdvectivelyConnectedTo
p_from = TEMP.advectivelyConnectsFrom
p_to = TEMP.advectivelyConnectsTo
p_cmat = TEMP.advectivelyConnectsMaterial
diffusive = False
elif link['conveyingType'] == 'DIFFUSIVE':
p_is = TEMP.isDiffusivelyConnectedTo
p_from = TEMP.diffusivelyConnectsFrom
p_to = TEMP.diffusivelyConnectsTo
p_cmat = TEMP.diffusivelyConnectsMaterial
diffusive = True
else:
log.critical(f'unhandled conveying type {link}')
continue
source = link['source']
target = link['target']
ok = True
if len(from_to) == 2: # otherwise
st = []
from_to = []
for i, e in enumerate((source, target)):
ed = rm[e]
if 'external' not in ed:
if not i and from_to:
# TODO make sure the intermediate ids match
pass
else:
ok = False
break
else:
st.append(e)
from_to.append(OntId(ed['external'][0]))
conveying = link['conveyingLyph']
cd = rm[conveying]
if 'external' in cd:
old_ot = ot
ot = OntTerm(cd['external'][0])
yield ot.u, rdf.type, owl.Class
yield ot.u, TEMP.internalId, rdflib.Literal(conveying)
yield ot.u, rdfs.label, rdflib.Literal(ot.label)
yield from self.materialTriples(
ot.u, link, p_cmat) # FIXME locate this correctly
if ok:
u, d = from_to
if st[0] == source:
yield u, rdfs.label, rdflib.Literal(
OntTerm(u).label)
yield u, rdf.type, owl.Class
yield from cmb.restriction.serialize(
ot.u, p_from, u)
if st[1] == target:
yield d, rdfs.label, rdflib.Literal(
OntTerm(d).label)
yield d, rdf.type, owl.Class
yield from cmb.restriction.serialize(ot.u, p_to, d)
if old_ot is not None and old_ot != ot:
yield from cmb.restriction.serialize(
ot.u, p_from, old_ot.u)
if diffusive:
# we can try to hack this using named individuals
# but it is not going to do exactly what is desired
s_link = TEMP[f'ApiNATOMY/{mid}/{link["id"]}']
s_cd = TEMP[f'ApiNATOMY/{mid}/{cd["id"]}']
yield s_link, rdf.type, owl.NamedIndividual
yield s_link, rdf.type, TEMP.diffusiveLink # FIXME I'm not sure these go in the model ...
yield s_cd, rdf.type, owl.NamedIndividual
if 'external' in cd and cd['external']:
oid = OntId(cd['external'][0])
yield s_cd, rdf.type, oid.u
ot = oid.asTerm()
if ot.label:
yield oid.u, rdfs.label, ot.label
else:
yield s_cd, rdf.type, TEMP.conveyingLyph
for icd in cd['inCoalescences']:
dcd = rm[icd]
log.info(lj(dcd))
s_icd = TEMP[f'ApiNATOMY/{mid}/{dcd["id"]}']
yield s_cd, TEMP.partOfCoalescence, s_icd
yield s_icd, rdf.type, owl.NamedIndividual
yield s_icd, rdf.type, TEMP[
'ApiNATOMY/Coalescence']
if 'external' in dcd and dcd['external']:
oid = OntId(dcd['external'][0])
yield s_icd, rdf.type, oid.u
ot = oid.asTerm()
if ot.label:
yield oid.u, rdfs.label, ot.label
for lyphid in dcd['lyphs']:
ild = rm[lyphid]
log.info(lj(ild))
if 'external' in ild and ild['external']:
yield s_icd, TEMP.hasLyphWithMaterial, OntId(
ild['external'][0])
if not ok:
logd.info(f'{source} {target} issue')
continue
for inid, e in zip(st, from_to):
yield e.u, rdf.type, owl.Class
yield e.u, rdfs.label, rdflib.Literal(OntTerm(e).label)
yield e.u, TEMP.internalId, rdflib.Literal(inid)
f, t = from_to
yield from cmb.restriction.serialize(f.u, p_is, t.u)
def added(self):
data = super().added
if data['meta'] == {'techniques': []}:
breakpoint()
# FIXME conditional lifts ...
if 'award_number' not in data['meta']:
am = self.lifters.award_manual
if am:
data['meta']['award_number'] = am
if 'modality' not in data['meta']:
m = self.lifters.modality
if m:
data['meta']['modality'] = m
if False and 'organ' not in data['meta']:
# skip here, now attached directly to award
if 'award_number' in data['meta']:
an = data['meta']['award_number']
o = self.lifters.organ(an)
if o:
if o != 'othertargets':
o = OntId(o)
if o.prefix == 'FMA':
ot = OntTerm(o)
o = next(
OntTerm.query(label=ot.label, prefix='UBERON'))
data['meta']['organ'] = o
if 'organ' not in data['meta'] or data['meta'][
'organ'] == 'othertargets':
o = self.lifters.organ_term
if o:
if isinstance(o, str):
o = o,
out = tuple()
for _o in o:
_o = OntId(_o)
if _o.prefix == 'FMA':
ot = OntTerm(_o)
_o = next(
OntTerm.query(label=ot.label, prefix='UBERON'))
out += (_o, )
data['meta']['organ'] = out
if 'protocol_url_or_doi' not in data['meta']:
if self.lifters.protocol_uris:
data['meta']['protocol_url_or_doi'] = tuple(
self.lifters.protocol_uris)
else:
if not isinstance(data['meta']['protocol_url_or_doi'], tuple):
_test_path = deque(['meta', 'protocol_url_or_doi'])
if not [e for e in data['errors'] if e['path'] == _test_path]:
raise ext.ShouldNotHappenError('urg')
else:
data['meta']['protocol_url_or_doi'] += tuple(
self.lifters.protocol_uris)
data['meta']['protocol_url_or_doi'] = tuple(
sorted(set(data['meta']['protocol_url_or_doi']))) # ick
return data
def added(self):
data = super().added
if data['meta'] == {'techniques': []}:
breakpoint()
# FIXME conditional lifts ...
if 'award_number' not in data['meta']:
am = self.lifters.award_manual
if am:
data['meta']['award_number'] = am
if 'modality' not in data['meta']:
m = self.lifters.modality
if m:
data['meta']['modality'] = m
if False and 'organ' not in data['meta']:
# skip here, now attached directly to award
if 'award_number' in data['meta']:
an = data['meta']['award_number']
o = self.lifters.organ(an)
if o:
if o != 'othertargets':
o = OntId(o)
if o.prefix == 'FMA':
ot = OntTerm(o)
o = next(OntTerm.query(label=ot.label, prefix='UBERON'))
data['meta']['organ'] = o
if 'organ' not in data['meta'] or data['meta']['organ'] == 'othertargets':
o = self.lifters.organ_term
if o:
if isinstance(o, str):
o = o,
out = tuple()
for _o in o:
_o = OntId(_o)
if _o.prefix == 'FMA':
ot = OntTerm(_o)
_o = next(OntTerm.query(label=ot.label, prefix='UBERON'))
out += (_o,)
data['meta']['organ'] = out
if 'protocol_url_or_doi' not in data['meta']:
if self.lifters.protocol_uris:
data['meta']['protocol_url_or_doi'] = tuple(self.lifters.protocol_uris)
else:
if not isinstance(data['meta']['protocol_url_or_doi'], tuple):
_test_path = deque(['meta', 'protocol_url_or_doi'])
if not [e for e in data['errors']
if 'path' in e and e['path'] == _test_path]:
raise ext.ShouldNotHappenError('urg')
else:
data['meta']['protocol_url_or_doi'] += tuple(self.lifters.protocol_uris)
data['meta']['protocol_url_or_doi'] = tuple(sorted(set(data['meta']['protocol_url_or_doi']))) # ick
# FIXME this is a really bad way to do this :/ maybe stick the folder in data['prov'] ?
# and indeed, when we added PipelineStart this shifted and broke everything
local = (self
.previous_pipeline.pipelines[0]
.previous_pipeline.pipelines[0]
.previous_pipeline.pipelines[0]
.path)
remote = local.remote
if 'doi' not in data['meta']:
doi = remote.doi
if doi is not None:
try:
metadata = doi.metadata()
if metadata is not None:
data['meta']['doi'] = doi.identifier
except requests.exceptions.HTTPError:
data['meta']['doi'] = None
pass
else:
data['meta']['doi'] = None
if 'status' not in data:
data['status'] = {}
if 'status_on_platform' not in data['status']:
data['status']['status_on_platform'] = remote.bfobject.status
return data
