def normalize_prefixes(graph, curies):
mg = makeGraph('nifall',
makePrefixes('owl', 'skos', 'oboInOwl'),
graph=graph)
mg.del_namespace('')
old_namespaces = list(graph.namespaces())
ng_ = makeGraph('', prefixes=makePrefixes('oboInOwl', 'skos'))
[ng_.g.add(t) for t in mg.g]
[ng_.add_namespace(n, p) for n, p in curies.items() if n != '']
#[mg.add_namespace(n, p) for n, p in old_namespaces if n.startswith('ns') or n.startswith('default')]
#[mg.del_namespace(n) for n in list(mg.namespaces)]
#graph.namespace_manager.reset()
#[mg.add_namespace(n, p) for n, p in wat.items() if n != '']
return mg, ng_
def ilx_json_to_tripples(
j
): # this will be much eaiser if everything can be exported as a relationship or an anotation
g = makeGraph('do not write me',
prefixes=makePrefixes('ILX', 'ilx', 'owl', 'skos', 'NIFRID'))
def pref(inp):
return makePrefixes('ilx')['ilx'] + inp
id_ = pref(j['ilx'])
type_ = {
'term': 'owl:Class',
'relationship': 'owl:ObjectProperty',
'annotation': 'owl:AnnotationProperty'
}[j['type']]
out = [] # TODO need to expand these
out.append((id_, rdflib.RDF.type, type_))
out.append((id_, rdflib.RDFS.label, j['label']))
out.append((id_, 'skos:definition', j['definition']))
for syndict in j['synonyms']:
out.append((id_, 'NIFRID:synonym', syndict['literal']))
for superdict in j[
'superclasses']: # should we be returning the preferred id here not the ilx? or maybe that is a different json output?
out.append((id_, rdflib.RDFS.subClassOf, pref(superdict['ilx'])))
for eid in j['existing_ids']:
out.append((id_, 'ilx:someOtherId', eid['iri'])) # predicate TODO
[g.add_trip(*o) for o in out]
return g.g.serialize(format='nifttl') # other formats can be choosen
def load(file):
filepath = os.path.expanduser(file)
_, ext = os.path.splitext(filepath)
filetype = ext.strip('.')
if filetype == 'ttl':
infmt = 'turtle'
else:
infmt = None
print(filepath)
graph = rdflib.Graph()
try:
graph.parse(filepath, format=infmt)
except rdflib.plugins.parsers.notation3.BadSyntax as e:
print('PARSING FAILED', filepath)
raise e
og = makeGraph('', graph=graph)
# FIXME this should really just be a function :/
curie, *prefs = kludge(filepath)
name = os.path.splitext(os.path.basename(filepath))[0]
if 'slim' in name:
name = name.replace('slim', '')
try:
version = list(graph.subject_objects(owl.versionIRI))[0][1]
except IndexError:
version = list(graph.subjects(rdf.type, owl.Ontology))[0]
ng = createOntology(
f'{name}-dead', f'NIF {curie} deprecated',
makePrefixes('replacedBy', 'NIFRID', curie, *prefs), f'{name}dead',
f'Classes from {curie} with owl:deprecated true that we want rdfs:subClassOf NIFRID:birnlexRetiredClass, or classes hiding in a oboInOwl:hasAlternativeId annotation. This file was generated by pyontutils/necromancy from {version}.'
)
extract(og, ng, curie)
def uri_switch(filenames, get_values):
replacement_graph = createOntology(
'NIF-NIFSTD-mapping', 'NIF* to NIFSTD equivalents',
makePrefixes('BIRNANN', 'BIRNOBI', 'BIRNOBO', 'NIFANN', 'NIFCELL',
'NIFCHEM', 'NIFDYS', 'NIFFUN', 'NIFGA', 'NIFGG', 'NIFINV',
'NIFMOL', 'NIFMOLINF', 'NIFMOLROLE', 'NIFNCBISLIM',
'NIFNEURBR', 'NIFNEURBR2', 'NIFNEURCIR', 'NIFNEURMC',
'NIFNEURMOR', 'NIFNEURNT', 'NIFORG', 'NIFQUAL', 'NIFRES',
'NIFRET', 'NIFSCID', 'NIFSUB', 'NIFUNCL', 'OBOANN',
'SAOCORE'))
fragment_prefixes, ureps = get_values(replacement_graph)
print('Start writing')
trips_lists = Parallel(n_jobs=9)(
delayed(do_file)(f, swapUriSwitch, ureps, fragment_prefixes)
for f in filenames)
print('Done writing')
[replacement_graph.g.add(t) for trips in trips_lists for t in trips]
replacement_graph.write()
def main():
records = get_records()
graph = createOntology(
'scicrunch-registry',
'scicrunch registry exported ontology',
makePrefixes(
'owl',
'rdf',
'rdfs',
'skos',
'definition',
'SCR', # generate base from this directly?
#'obo':'http://purl.obolibrary.org/obo/',
#'FIXME':'http://fixme.org/',
'NIFSTD', # for old ids??
'NIFRID',
'oboInOwl'
), # these aren't really from OBO files but they will be friendly known identifiers to people in the community
'screxp',
'This file is automatically generated from the SciCrunch resource registry on a weekly basis via https://github.com/tgbugs/pyontutils/blob/master/scr_sync.py.',
path='')
make_file(graph, records)
def convert(f, prefixes):
if f in exclude:
print('skipping', f)
return f
pi = {v: k for k, v in prefixes.items()}
graph = rdflib.Graph()
graph.parse(f, format='turtle')
namespaces = [str(n) for p, n in graph.namespaces()]
prefs = ['']
asdf = {} #{v:k for k, v in ps.items()}
asdf.update(pi)
# determine which prefixes we need
for uri in list(graph.subjects()) + list(graph.predicates()) + list(
graph.objects()):
if uri.endswith('.owl') or uri.endswith('.ttl'):
continue # don't prefix imports
for rn, rp in sorted(
asdf.items(),
key=lambda a: -len(a[0])): # make sure we get longest first
lrn = len(rn)
if type(uri) == rdflib.BNode:
continue
elif uri.startswith(
rn) and '#' not in uri[lrn:] and '/' not in uri[
lrn:]: # prevent prefixing when there is another sep
prefs.append(rp)
break
if prefs:
ps = makePrefixes(*prefs)
else:
ps = makePrefixes('rdfs')
if 'parcellation/' in f:
nsl = {p: n for p, n in graph.namespaces()}
if '' in nsl:
ps[''] = nsl['']
elif 'hbaslim' in f:
ps['HBA'] = nsl['HBA']
elif 'mbaslim' in f:
ps['MBA'] = nsl['MBA']
elif 'cocomac' in f:
ps['cocomac'] = nsl['cocomac']
# special cases for NIFORG, NIFINV, NIFRET where there identifiers in
# annotation properties that share the prefix, so we warn incase
# at some point in the future for some reason want them again...
if f == 'NIF-Organism.ttl':
print('WARNING: special case for NIFORG')
ps.pop('NIFORG')
elif f == 'NIF-Investigation.ttl':
print('WARNING: special case for NIFINV')
ps.pop('NIFINV')
elif f == 'unused/NIF-Retired.ttl':
print('WARNING: special case for NIFRET')
ps.pop('NIFGA')
ng = makeGraph('', prefixes=ps)
ng.filename = f
[ng.g.add(t) for t in graph]
#[ng.add_trip(*n) for n in graph.triples([None]*3)]
#print(f, len(ng.g))
ng.write()
return f
def convert(f):
if f in exclude:
print('skipping', f)
return f
ps = {
'PROTEGE': 'http://protege.stanford.edu/plugins/owl/protege#',
}
PREFIXES.update(ps)
pi = {v: k for k, v in PREFIXES.items()}
pi.pop(None)
graph = rdflib.Graph()
graph.parse(f, format='turtle')
namespaces = [str(n) for p, n in graph.namespaces()]
prefs = []
if f == 'NIF-Dysfunction.ttl':
prefs.append('OBO')
elif f == 'NIF-Eagle-I-Bridge.ttl':
prefs.append('IAO')
elif f == 'resources.ttl':
prefs.append('IAO')
elif f == 'NIF-Investigation.ttl':
prefs.append('IAO')
asdf = {v: k for k, v in ps.items()}
asdf.update(pi)
# determine which prefixes we need
for rn, rp in asdf.items():
for uri in list(graph.subjects()) + list(graph.predicates()) + list(
graph.objects()):
if type(uri) == rdflib.BNode:
continue
elif uri.startswith(rn):
if rp == 'OBO' or rp == 'IAO' or rp == 'NIFTTL':
if rp == 'IAO' and 'IAO_0000412' in uri: # for sequence_slim
pass
else:
continue
prefs.append(rp)
break
if prefs:
ps = makePrefixes(*prefs)
else:
ps = makePrefixes('rdfs')
if 'parcellation/' in f:
nsl = {p: n for p, n in graph.namespaces()}
if '' in nsl:
ps[''] = nsl['']
elif 'hbaslim' in f:
ps['HBA'] = nsl['HBA']
elif 'mbaslim' in f:
ps['MBA'] = nsl['MBA']
elif 'cocomac' in f:
ps['cocomac'] = nsl['cocomac']
ng = makeGraph(os.path.splitext(f)[0],
prefixes=ps,
writeloc=os.path.expanduser('~/git/NIF-Ontology/ttl/'))
[ng.add_node(*n) for n in graph.triples([None] * 3)]
#print(f, len(ng.g))
ng.write()
return f
from IPython import embed
source = 'https://raw.githubusercontent.com/BlueBrain/nat/master/nat/modelingDictionary.csv'
delimiter = ';'
resp = requests.get(source)
rows = [
r for r in csv.reader(resp.text.split('\n'), delimiter=delimiter)
if r and r[0][0] != '#'
]
header = [
'Record_ID', 'parent_category', 'name', 'description', 'required_tags'
]
PREFIXES = makePrefixes('owl', 'skos', 'ILX', 'ILXREPLACE', 'definition')
graph = makeGraph('measures', prefixes=PREFIXES)
class nat(rowParse):
def Record_ID(self, value):
print(value)
self.old_id = value
self._id = ILXREPLACE(value)
def parent_category(self, value):
self.super_old_id = value
self.super_id = ILXREPLACE(value)
def name(self, value):
self.hidden = value
fmt = 'turtle' if ext == '.ttl' else 'xml'
if noneMembers(o, 'go.owl', 'uberon.owl', 'pr.owl', 'doid.owl',
'taxslim.owl') or dobig:
graph.parse(o, format=fmt)
for i in range(4):
repeat(True)
with open(
os.path.expanduser(
'~/git/NIF-Ontology/scigraph/nifstd_curie_map.yaml'), 'rt') as f:
wat = yaml.load(f)
vals = set(wat.values())
mg = makeGraph('nifall', makePrefixes('owl', 'skos', 'oboInOwl'), graph=graph)
mg.del_namespace('')
old_namespaces = list(graph.namespaces())
ng_ = makeGraph('',
prefixes=makePrefixes('oboInOwl', 'skos'),
graph=rdflib.Graph())
[ng_.g.add(t) for t in mg.g]
[ng_.add_namespace(n, p) for n, p in wat.items() if n != '']
#[mg.add_namespace(n, p) for n, p in old_namespaces if n.startswith('ns') or n.startswith('default')]
#[mg.del_namespace(n) for n in list(mg.namespaces)]
#graph.namespace_manager.reset()
#[mg.add_namespace(n, p) for n, p in wat.items() if n != '']
def for_burak(ng):
def pref(inp):
return makePrefixes('ilx')['ilx'] + inp
def createRecordsFromGraph(graph, existing, target_graph=None):
mg = graph
graph = mg.g
s = rdflib.URIRef(makePrefixes('NIFRID')['NIFRID'] + 'synonym')
if target_graph is None:
target_ontology_iri = mg.ontid
else:
target_ontology_iri = target_graph.ontid
if 'ILXREPLACE' in mg.namespaces:
namespace = str(mg.namespaces['ILXREPLACE'])
else:
print('Nothing needs to be replaced in', mg.filename)
return
query = ("SELECT DISTINCT ?v "
"WHERE { {?v ?p ?o} UNION {?s ?v ?o} UNION {?s ?p ?v} . "
"FILTER("
"strstarts(str(?v), '%s') )}") % namespace
vals = graph.query(query)
#existing = {} # TODO this needs to populate from an existing source ie the ontology, and a tempid -> realid map?
for val, in vals:
qn = graph.namespace_manager.qname(val)
try:
labs = list(graph.objects(val, rdflib.RDFS.label))
label = str(labs[0])
except IndexError:
label = None # not defined here but we need to collect info here anyway
definition = list(graph.objects(val, rdflib.namespace.SKOS.definition))
if definition: definition = definition[0]
synonyms = list(graph.objects(val, s))
superclass = [
_ for _ in graph.objects(val, rdflib.RDFS.subClassOf)
if type(_) == rdflib.URIRef
]
superclass = superclass[0] if superclass else None
try:
superclass = graph.namespace_manager.qname(superclass)
except (TypeError, ValueError) as e:
print('ERROR: superclass of', qn, 'not a proper uri', superclass)
superclass = None
rec = makeIlxRec(label,
definition,
type='term',
comment=qn,
synonyms=synonyms,
existing_ids=[],
superclass=superclass,
ontologies=[target_ontology_iri])
if qn in existing:
_tmp = set(existing[qn]['files'])
_tmp.add(mg.filename) # prevent duplicate files from multiple runs
existing[qn]['files'] = list(_tmp)
newrec = {}
for k, v in existing[qn]['rec'].items():
if v:
newrec[k] = v
else:
newrec[k] = rec[k]
existing[qn]['rec'] = newrec
else:
existing[qn] = {
'id': None,
'sc': superclass, # make sorting easier
'files': [mg.filename],
#'done':False, # we probably don't need this
'rec': rec
}
superToLabel(existing)
def backend_refactor_values():
uri_reps_lit = {
# from https://github.com/information-artifact-ontology/IAO/blob/master/docs/BFO%201.1%20to%202.0%20conversion/mapping.txt
'http://www.ifomis.org/bfo/1.1#Entity': 'BFO:0000001',
'BFO1SNAP:Continuant': 'BFO:0000002',
'BFO1SNAP:Disposition': 'BFO:0000016',
'BFO1SNAP:Function': 'BFO:0000034',
'BFO1SNAP:GenericallyDependentContinuant': 'BFO:0000031',
'BFO1SNAP:IndependentContinuant': 'BFO:0000004',
'BFO1SNAP:MaterialEntity': 'BFO:0000040',
'BFO1SNAP:Quality': 'BFO:0000019',
'BFO1SNAP:RealizableEntity': 'BFO:0000017',
'BFO1SNAP:Role': 'BFO:0000023',
'BFO1SNAP:Site': 'BFO:0000029',
'BFO1SNAP:SpecificallyDependentContinuant': 'BFO:0000020',
'BFO1SPAN:Occurrent': 'BFO:0000003',
'BFO1SPAN:ProcessualEntity': 'BFO:0000015',
'BFO1SPAN:Process': 'BFO:0000015',
'BFO1SNAP:ZeroDimensionalRegion': 'BFO:0000018',
'BFO1SNAP:OneDimensionalRegion': 'BFO:0000026',
'BFO1SNAP:TwoDimensionalRegion': 'BFO:0000009',
'BFO1SNAP:ThreeDimensionalRegion': 'BFO:0000028',
'http://purl.org/obo/owl/OBO_REL#bearer_of': 'RO:0000053',
'http://purl.org/obo/owl/OBO_REL#inheres_in': 'RO:0000052',
'ro:has_part': 'BFO:0000051',
'ro:part_of': 'BFO:0000050',
'ro:has_participant': 'RO:0000057',
'ro:participates_in': 'RO:0000056',
'http://purl.obolibrary.org/obo/OBI_0000294': 'RO:0000059',
'http://purl.obolibrary.org/obo/OBI_0000297': 'RO:0000058',
'http://purl.obolibrary.org/obo/OBI_0000300': 'BFO:0000054',
'http://purl.obolibrary.org/obo/OBI_0000308': 'BFO:0000055',
# more bfo
'BFO1SNAP:SpatialRegion': 'BFO:0000006',
'BFO1SNAP:FiatObjectPart': 'BFO:0000024',
'BFO1SNAP:ObjectAggregate': 'BFO:0000027',
'BFO1SNAP:Object': 'BFO:0000030',
#'BFO1SNAP:ObjectBoundary' # no direct replacement, only occurs in unused
#'BFO1SPAN:ProcessAggregate' # was not replaced, could simply be a process itself??
#'BFO1SNAP:DependentContinuant' # was not replaced
# other
#'ro:participates_in' # above
#'ro:has_participant' # above
#'ro:has_part', # above
#'ro:part_of', # above
#'ro:precedes' # unused and only in inferred
#'ro:preceded_by' # unused and only in inferred
#'ro:transformation_of' # unused and only in inferred
#'ro:transformed_into' # unused and only in inferred
'http://purl.org/obo/owl/obo#inheres_in': 'RO:0000052',
'http://purl.obolibrary.org/obo/obo#towards': 'RO:0002503',
'http://purl.org/obo/owl/pato#towards': 'RO:0002503',
'http://purl.obolibrary.org/obo/pato#inheres_in': 'RO:0000052',
'BIRNLEX:17': 'RO:0000053', # is_bearer_of
'http://purl.obolibrary.org/obo/pato#towards': 'RO:0002503',
'ro:adjacent_to': 'RO:0002220',
'ro:derives_from': 'RO:0001000',
'ro:derives_into': 'RO:0001001',
'ro:agent_in': 'RO:0002217',
'ro:has_agent': 'RO:0002218',
'ro:contained_in': 'RO:0001018',
'ro:contains': 'RO:0001019',
'ro:located_in': 'RO:0001025',
'ro:location_of': 'RO:0001015',
'ro:has_proper_part': 'NIFRID:has_proper_part',
'ro:proper_part_of':
'NIFRID:proper_part_of', # part of where things are not part of themsevles need to review
}
ug = makeGraph('',
prefixes=makePrefixes('ro', 'RO', 'BIRNLEX', 'NIFRID',
'BFO', 'BFO1SNAP', 'BFO1SPAN'))
ureps = {
ug.check_thing(k): ug.check_thing(v)
for k, v in uri_reps_lit.items()
}
return ureps
def attachPrefixes(*prefixes, graph=None):
return makeGraph('', prefixes=makePrefixes(*prefixes), graph=graph)
def main():
# load in our existing graph
# note: while it would be nice to allow specification of phenotypes to be decoupled
# from insertion into the graph... maybe we could enable this, but it definitely seems
# to break a number of nice features... and we would need the phenotype graph anyway
EXISTING_GRAPH = rdflib.Graph()
sources = ('/tmp/NIF-Neuron-Phenotype.ttl', '/tmp/NIF-Neuron-Defined.ttl',
'/tmp/NIF-Neuron.ttl', '/tmp/NIF-Phenotype-Core.ttl',
'/tmp/NIF-Phenotypes.ttl', '/tmp/hbp-special.ttl')
for file in sources:
EXISTING_GRAPH.parse(file, format='turtle')
EXISTING_GRAPH.namespace_manager.bind(
'ILXREPLACE',
makePrefixes('ILXREPLACE')['ILXREPLACE'])
#EXISTING_GRAPH.namespace_manager.bind('PR', makePrefixes('PR')['PR'])
PREFIXES = makePrefixes(
'owl',
'skos',
'PR',
'UBERON',
'NCBITaxon',
'ILXREPLACE',
'ilx',
'ILX',
'NIFCELL',
'NIFMOL',
)
graphBase.core_graph = EXISTING_GRAPH
graphBase.out_graph = rdflib.Graph()
graphBase._predicates = getPhenotypePredicates(EXISTING_GRAPH)
g = makeGraph('merged',
prefixes={k: str(v)
for k, v in EXISTING_GRAPH.namespaces()},
graph=EXISTING_GRAPH)
reg_neurons = list(
g.g.subjects(rdflib.RDFS.subClassOf, g.expand(NIFCELL_NEURON)))
tc_neurons = [
_ for (_, ) in
g.g.query('SELECT DISTINCT ?match WHERE {?match rdfs:subClassOf+ %s}' %
NIFCELL_NEURON)
]
def_neurons = g.get_equiv_inter(NIFCELL_NEURON)
nodef = sorted(set(tc_neurons) - set(def_neurons))
MeasuredNeuron.out_graph = rdflib.Graph()
Neuron.out_graph = rdflib.Graph()
mns = [MeasuredNeuron(id_=n) for n in nodef]
dns = [Neuron(id_=n) for n in sorted(def_neurons)]
#dns += [Neuron(*m.pes) if m.pes else m.id_ for m in mns]
dns += [Neuron(*m.pes) for m in mns if m.pes]
# reset everything for export
Neuron.out_graph = rdflib.Graph()
ng = makeGraph('output', prefixes=PREFIXES, graph=Neuron.out_graph)
Neuron.existing_pes = {
} # reset this as well because the old Class references have vanished
dns = [Neuron(*d.pes) for d in set(dns)
] # TODO remove the set and use this to test existing bags?
from neuron_lang import WRITEPYTHON
WRITEPYTHON(sorted(dns))
ng.add_ont(ILXREPLACE('defined-neurons'), 'Defined Neurons', 'NIFDEFNEU',
'VERY EXPERIMENTAL', '0.0.0.1a')
ng.add_node(ILXREPLACE('defined-neurons'), 'owl:imports',
'http://ontology.neuinfo.org/NIF/ttl/NIF-Phenotype-Core.ttl')
ng.add_node(ILXREPLACE('defined-neurons'), 'owl:imports',
'http://ontology.neuinfo.org/NIF/ttl/NIF-Phenotypes.ttl')
ng.write()
bads = [
n for n in ng.g.subjects(rdflib.RDF.type, rdflib.OWL.Class)
if len(list(ng.g.predicate_objects(n))) == 1
]
embed()