def create_entity(graph, fs_subject_id, filepath, hostname):
""" Create a PROV entity for a file in a FreeSurfer directory
"""
# identify FreeSurfer terms based on directory and file names
_, filename = os.path.split(filepath)
relpath = filepath.split(fs_subject_id)[1].lstrip(os.path.sep)
fstypes = relpath.split('/')[:-1]
additional_types = relpath.split('/')[-1].split('.')
file_md5_hash = hash_infile(filepath, crypto=hashlib.md5)
file_sha512_hash = hash_infile(filepath, crypto=hashlib.sha512)
if file_md5_hash is None:
print('Empty file: %s' % filepath)
#url = "file://%s%s" % (hostname, filepath)
url = filepath
url_get = prov.URIRef("http://localhost/file?file=%s" % filepath)
url_get = prov.URIRef("file://%s" % filepath)
obj_attr = [(prov.PROV["label"], filename),
(fs["relative_path"], "%s" % relpath),
(prov.PROV["location"], url_get),
(crypto["md5"], "%s" % file_md5_hash),
(crypto["sha"], "%s" % file_sha512_hash)
]
for key in fstypes:
obj_attr.append((nidm["tag"], key))
for key in additional_types:
obj_attr.append((nidm["tag"], key))
for key, uris in fs_file_map:
if key in filename:
if key.rstrip('.').lstrip('.') not in fstypes + additional_types:
obj_attr.append((nidm["tag"], key.rstrip('.').lstrip('.')))
for uri in uris:
if isinstance(uri, tuple):
obj_attr.append((uri[0], uri[1]))
else:
obj_attr.append((prov.PROV["type"], uri))
id = get_id()
return graph.entity(id, obj_attr)
"""
Retrieve the file into wd
"""
out_T1_files = []
filemap = {}
for idx, info in enumerate(t1_result.bindings):
o = urlparse.urlparse(info['?t1path'])
if o.scheme.startswith('file'):
uri = 'file://' + o.path
else:
uri = info['?t1path']
filename = os.path.join(cwd,
'file_%d_' % idx + os.path.split(o.path)[-1])
urllib.urlretrieve(uri, filename)
if hash_infile(filename, crypto=hashlib.sha512) != str(info['?sha']):
raise IOError("Hash of file doesn't match remote hash")
out_T1_files.append(filename)
filemap[filename] = (info['?sha'], info['?e'])
"""
Run bet and convert to rdf
"""
provgraph, rdfgraph = run_bet(out_T1_files, cwd)
nipype_files = """
PREFIX nipype: <http://nipy.org/nipype/terms/>
select ?e ?value where {
?e a prov:Entity ;
nipype:value ?value .
FILTER(regex(?value, 'file://'))
subject = 'SAD_024'
ingraph = pm.ProvBundle(identifier=get_id())
ingraph.add_namespace(foaf)
ingraph.add_namespace(niiri)
ingraph.add_namespace(nif)
ingraph.add_namespace(crypto)
agent = ingraph.agent(get_id(), {
pm.PROV["type"]: pm.PROV["Person"],
foaf["name"]: subject
})
t1_collection = ingraph.collection(get_id())
ingraph.wasAttributedTo(t1_collection, agent)
for t1path in T1s:
file_md5_hash = hash_infile(t1path, crypto=hashlib.md5)
file_sha512_hash = hash_infile(t1path, crypto=hashlib.sha512)
url = "file://%s%s" % (getfqdn(), t1path)
url_get = pm.URIRef(url)
obj_attr = [(pm.PROV["location"], url_get),
(crypto["md5"], "%s" % file_md5_hash),
(crypto["sha"], "%s" % file_sha512_hash),
(pm.PROV["type"], nif["nlx_inv_20090243"])]
t1entity = ingraph.entity(get_id(), obj_attr)
ingraph.hadMember(t1_collection, t1entity)
rdfingraph = ingraph.rdf()
print rdfingraph.serialize(format='turtle')
"""
The input graph as generated by the above code
"""
t1_result = rdfingraph.query(t1_query)
"""
Retrieve the file into wd
"""
out_T1_files = []
filemap = {}
for idx, info in enumerate(t1_result.bindings):
o = urlparse.urlparse(info['?t1path'])
if o.scheme.startswith('file'):
uri = 'file://' + o.path
else:
uri = info['?t1path']
filename = os.path.join(cwd,
'file_%d_' % idx + os.path.split(o.path)[-1])
urllib.urlretrieve(uri, filename)
if hash_infile(filename, crypto=hashlib.sha512) != str(info['?sha']):
raise IOError("Hash of file doesn't match remote hash")
out_T1_files.append(filename)
filemap[filename] = (info['?sha'], info['?e'])
"""
Run bet and convert to rdf
"""
provgraph, rdfgraph = run_bet(out_T1_files, cwd)
nipype_files = """
PREFIX nipype: <http://nipy.org/nipype/terms/>
select ?e ?value where {
?e a prov:Entity ;
nipype:value ?value .
FILTER(regex(?value, 'file://'))
}
subject = 'SAD_024'
ingraph = pm.ProvBundle(identifier=get_id())
ingraph.add_namespace(foaf)
ingraph.add_namespace(niiri)
ingraph.add_namespace(nif)
ingraph.add_namespace(crypto)
agent = ingraph.agent(get_id(),
{pm.PROV["type"]: pm.PROV["Person"],
foaf["name"]: subject}
)
t1_collection = ingraph.collection(get_id())
ingraph.wasAttributedTo(t1_collection, agent)
for t1path in T1s:
file_md5_hash = hash_infile(t1path, crypto=hashlib.md5)
file_sha512_hash = hash_infile(t1path, crypto=hashlib.sha512)
url = "file://%s%s" % (getfqdn(), t1path)
url_get = pm.URIRef(url)
obj_attr = [(pm.PROV["location"], url_get),
(crypto["md5"], "%s" % file_md5_hash),
(crypto["sha"], "%s" % file_sha512_hash),
(pm.PROV["type"], nif["nlx_inv_20090243"])
]
t1entity = ingraph.entity(get_id(), obj_attr)
ingraph.hadMember(t1_collection, t1entity)
rdfingraph = ingraph.rdf()
print rdfingraph.serialize(format='turtle')
"""