class RawUniqueLinks(Base, IndraDBTable):
__tablename__ = 'raw_unique_links'
_always_disp = ['raw_stmt_id', 'pa_stmt_mk_hash']
_indices = [
BtreeIndex('raw_unique_links_raw_stmt_id_idx', 'raw_stmt_id'),
BtreeIndex('raw_unique_links_pa_stmt_mk_hash_idx',
'pa_stmt_mk_hash')
]
id = Column(Integer, primary_key=True)
raw_stmt_id = Column(Integer,
ForeignKey('raw_statements.id'),
nullable=False)
pa_stmt_mk_hash = Column(BigInteger,
ForeignKey('pa_statements.mk_hash'),
nullable=False)
__table_args = (UniqueConstraint('raw_stmt_id',
'pa_stmt_mk_hash',
name='stmt-link-uniqueness'), )
class MtiRefAnnotationsTest(Base, IndraDBTable):
__tablename__ = 'mti_ref_annotations_test'
_always_disp = ['pmid_num', 'mesh_num', 'qual_num']
_indices = [
BtreeIndex('mti_ref_annotations_test_pmid_idx', 'pmid_num'),
BtreeIndex('mti_ref_annotations_test_mesh_id_idx', 'mesh_num'),
BtreeIndex('mti_ref_annotations_test_qual_id_idx', 'qual_num')
]
id = Column(Integer, primary_key=True)
pmid_num = Column(Integer, nullable=False)
mesh_num = Column(Integer, nullable=False)
qual_num = Column(Integer)
major_topic = Column(Boolean, default=False)
is_concept = Column(Boolean, default=False)
__table_args__ = (UniqueConstraint('pmid_num',
'mesh_num',
'qual_num',
'is_concept',
name='mesh-uniqueness'), )
class TextRef(Base, IndraDBTable):
__tablename__ = 'text_ref'
_ref_cols = ['pmid', 'pmcid', 'doi', 'pii', 'url', 'manuscript_id']
_always_disp = ['id', 'pmid', 'pmcid']
_indices = [
StringIndex('text_ref_pmid_idx', 'pmid'),
StringIndex('text_ref_pmcid_idx', 'pmcid'),
BtreeIndex('text_ref_pmid_num_idx', 'pmid_num'),
BtreeIndex('text_ref_pmcid_num_idx', 'pmcid_num'),
BtreeIndex('text_ref_doi_ns_idx', 'doi_ns'),
BtreeIndex('text_ref_doi_id_idx', 'doi_id'),
StringIndex('text_ref_doi_idx', 'doi')
]
id = Column(Integer, primary_key=True)
pmid = Column(String(20))
pmid_num = Column(Integer)
pmcid = Column(String(20))
pmcid_num = Column(Integer)
pmcid_version = Column(Integer)
doi = Column(String(100))
doi_ns = Column(Integer)
doi_id = Column(String)
pii = Column(String(250))
url = Column(String, unique=True)
manuscript_id = Column(String(100), unique=True)
create_date = Column(DateTime, default=func.now())
last_updated = Column(DateTime, onupdate=func.now())
__table_args__ = (UniqueConstraint('pmid', 'doi', name='pmid-doi'),
UniqueConstraint('pmid', 'pmcid', name='pmid-pmcid'),
UniqueConstraint('pmcid', 'doi', name='pmcid-doi'))
def __repr__(self):
terms = [f'id={self.id}']
for col in ['pmid', 'pmcid', 'doi', 'pii', 'url', 'manuscript_id']:
if getattr(self, col) is not None:
terms.append(f'{col}={getattr(self, col)}')
if len(terms) > 2:
break
return f'{self.__class__.__name__}({", ".join(terms)})'
@classmethod
def new(cls,
pmid=None,
pmcid=None,
doi=None,
pii=None,
url=None,
manuscript_id=None):
pmid, pmid_num = cls.process_pmid(pmid)
pmcid, pmcid_num, pmcid_version = cls.process_pmcid(pmcid)
doi, doi_ns, doi_id = cls.process_doi(doi)
return cls(pmid=pmid,
pmid_num=pmid_num,
pmcid=pmcid,
pmcid_num=pmcid_num,
pmcid_version=pmcid_version,
doi=doi,
doi_ns=doi_ns,
doi_id=doi_id,
pii=pii,
url=url,
manuscript_id=manuscript_id)
def update(self, **updates):
for id_type, id_val in updates.items():
if id_type == 'pmid':
self.pmid, self.pmid_num = self.process_pmid(id_val)
elif id_type == 'pmcid':
self.pmcid, self.pmcid_num, self.pmcid_version = \
self.process_pmcid(id_val)
elif id_type == 'doi':
self.doi, self.doi_ns, self.doi_id = \
self.process_doi(id_val)
else:
setattr(self, id_type, id_val)
return
@staticmethod
def process_pmid(pmid):
if not pmid:
return None, None
if not pmid.isdigit():
return pmid, None
return pmid, int(pmid)
@staticmethod
def process_pmcid(pmcid):
if not pmcid:
return None, None, None
if not pmcid.startswith('PMC'):
return pmcid, None, None
if '.' in pmcid:
pmcid, version_number_str = pmcid.split('.')
if version_number_str.isdigit():
version_number = int(version_number_str)
else:
version_number = None
else:
version_number = None
if not pmcid[3:].isdigit():
return pmcid, None, version_number
return pmcid, int(pmcid[3:]), version_number
@staticmethod
def process_doi(doi):
# Check for invalid DOIs
if not doi:
return None, None, None
# Regularize case.
doi = doi.upper()
if not doi.startswith('10.'):
return doi, None, None
# Split up the parts of the DOI
parts = doi[3:].split('/')
if len(parts) < 2:
return doi, None, None
# Check the namespace number, make it an integer.
namespace_str = parts[0]
if not namespace_str.isdigit():
return doi, None, None
namespace = int(namespace_str)
# Join the res of the parts together.
group_id = '/'.join(parts[1:])
return doi, namespace, group_id
@classmethod
def pmid_in(cls, pmid_list, filter_ids=False):
"""Get sqlalchemy clauses for a list of pmids."""
pmid_num_set = set()
for pmid in pmid_list:
_, pmid_num = cls.process_pmid(pmid)
if pmid_num is None:
if filter_ids:
logger.warning('"%s" is not a valid pmid. Skipping.' %
pmid)
continue
else:
ValueError('"%s" is not a valid pmid.' % pmid)
pmid_num_set.add(pmid_num)
return cls.pmid_num.in_(pmid_num_set)
@classmethod
def pmcid_in(cls, pmcid_list, filter_ids=False):
"""Get the sqlalchemy clauses for a list of pmcids."""
pmcid_num_set = set()
for pmcid in pmcid_list:
_, pmcid_num, _ = cls.process_pmcid(pmcid)
if not pmcid_num:
if filter_ids:
logger.warning('"%s" does not look like a valid '
'pmcid. Skipping.' % pmcid)
continue
else:
raise ValueError('"%s" is not a valid pmcid.' % pmcid)
else:
pmcid_num_set.add(pmcid_num)
return cls.pmcid_num.in_(pmcid_num_set)
@classmethod
def doi_in(cls, doi_list, filter_ids=False):
"""Get clause for looking up a list of dois."""
doi_tuple_set = set()
for doi in doi_list:
doi, doi_ns, doi_id = cls.process_doi(doi)
if not doi_ns:
if filter_ids:
logger.warning('"%s" does not look like a normal doi. '
'Skipping.' % doi)
continue
else:
raise ValueError('"%s" is not a valid doi.' % doi)
else:
doi_tuple_set.add((doi_ns, doi_id))
return tuple_(cls.doi_ns, cls.doi_id).in_(doi_tuple_set)
def get_ref_dict(self):
ref_dict = {}
for ref in self._ref_cols:
val = getattr(self, ref, None)
if val:
ref_dict[ref.upper()] = val
ref_dict['TRID'] = self.id
return ref_dict