def get_agent(cls, cl_agent):
"""Get an agent from the kqml cl-json representation (KQMLList)."""
agent_json = cls.converter.cl_to_json(cl_agent)
if isinstance(agent_json, list):
return [ensure_agent_type(Agent._from_json(agj))
for agj in agent_json]
else:
return ensure_agent_type(Agent._from_json(agent_json))
def map_agent(self, agent, do_rename):
"""Return the given Agent with its grounding mapped.
This function grounds a single agent. It returns the new Agent object
(which might be a different object if we load a new agent state
from json) or the same object otherwise.
Parameters
----------
agent : :py:class:`indra.statements.Agent`
The Agent to map.
do_rename: bool
If True, the Agent name is updated based on the mapped grounding.
If do_rename is True the priority for setting the name is
FamPlex ID, HGNC symbol, then the gene name
from Uniprot.
Returns
-------
grounded_agent : :py:class:`indra.statements.Agent`
The grounded Agent.
maps_to_none : bool
True if the Agent is in the grounding map and maps to None.
"""
# We always standardize DB refs as a functionality in the
# GroundingMapper. If a new module is implemented which is
# responsible for standardizing grounding, this can be removed.
agent.db_refs = self.standardize_db_refs(agent.db_refs)
# If there is no TEXT available, we can return immediately since we
# can't do mapping
agent_text = agent.db_refs.get('TEXT')
if not agent_text:
# We still do the name standardization here
if do_rename:
self.standardize_agent_name(agent, standardize_refs=False)
return agent, False
mapped_to_agent_json = self.agent_map.get(agent_text)
if mapped_to_agent_json:
mapped_to_agent = \
Agent._from_json(mapped_to_agent_json['agent'])
return mapped_to_agent, False
# Look this string up in the grounding map
# If not in the map, leave agent alone and continue
if agent_text in self.gm:
map_db_refs = self.gm[agent_text]
# If it's in the map but it maps to None, then filter out
# this statement by skipping it
if map_db_refs is None:
logger.debug("Skipping %s" % agent_text)
return None, True
# If it has a value that's not None, map it and add it
else:
self.update_agent_db_refs(agent, agent_text, do_rename)
# This happens whene there is an Agent text but it is not in the
# grounding map. We still do the name standardization here.
if do_rename:
self.standardize_agent_name(agent, standardize_refs=False)
# Otherwise just return
return agent, False
def map_agent(self, agent, do_rename):
"""Return the given Agent with its grounding mapped.
This function grounds a single agent. It returns the new Agent object
(which might be a different object if we load a new agent state
from json) or the same object otherwise.
Parameters
----------
agent : :py:class:`indra.statements.Agent`
The Agent to map.
do_rename: bool
If True, the Agent name is updated based on the mapped grounding.
If do_rename is True the priority for setting the name is
FamPlex ID, HGNC symbol, then the gene name
from Uniprot.
Returns
-------
grounded_agent : :py:class:`indra.statements.Agent`
The grounded Agent.
"""
# We always standardize DB refs as a functionality in the
# GroundingMapper. If a new module is implemented which is
# responsible for standardizing grounding, this can be removed.
agent.db_refs = self.standardize_db_refs(agent.db_refs)
# If there is no TEXT available, we can return immediately since we
# can't do mapping
agent_text = agent.db_refs.get('TEXT')
if not agent_text:
# We still do the name standardization here
if do_rename:
self.standardize_agent_name(agent, standardize_refs=False)
return agent
# 1. Check if there is a full agent mapping and apply if there is
if agent_text in self.agent_map:
mapped_to_agent = \
Agent._from_json(self.agent_map[agent_text]['agent'])
return mapped_to_agent
# 2. Look agent text up in the grounding map
if agent_text in self.grounding_map:
self.update_agent_db_refs(agent, self.grounding_map[agent_text],
do_rename)
# 3. Look agent text up in the misgrounding map
if agent_text in self.misgrounding_map:
self.remove_agent_db_refs(agent, self.misgrounding_map[agent_text])
# This happens when there is an Agent text but it is not in the
# grounding map. We still do the name standardization here.
if do_rename:
self.standardize_agent_name(agent, standardize_refs=False)
# Otherwise just return
return agent
def map_agent(self, agent, do_rename):
"""Return the given Agent with its grounding mapped.
This function grounds a single agent. It returns the new Agent object
(which might be a different object if we load a new agent state
from json) or the same object otherwise.
Parameters
----------
agent : :py:class:`indra.statements.Agent`
The Agent to map.
do_rename: bool
If True, the Agent name is updated based on the mapped grounding.
If do_rename is True the priority for setting the name is
FamPlex ID, HGNC symbol, then the gene name
from Uniprot.
Returns
-------
grounded_agent : :py:class:`indra.statements.Agent`
The grounded Agent.
maps_to_none : bool
True if the Agent is in the grounding map and maps to None.
"""
agent_text = agent.db_refs.get('TEXT')
mapped_to_agent_json = self.agent_map.get(agent_text)
if mapped_to_agent_json:
mapped_to_agent = \
Agent._from_json(mapped_to_agent_json['agent'])
return mapped_to_agent, False
# Look this string up in the grounding map
# If not in the map, leave agent alone and continue
if agent_text in self.gm.keys():
map_db_refs = self.gm[agent_text]
else:
return agent, False
# If it's in the map but it maps to None, then filter out
# this statement by skipping it
if map_db_refs is None:
# Increase counter if this statement has not already
# been skipped via another agent
logger.debug("Skipping %s" % agent_text)
return None, True
# If it has a value that's not None, map it and add it
else:
# Otherwise, update the agent's db_refs field
self.update_agent_db_refs(agent, agent_text, do_rename)
return agent, False
def map_agent(self, agent, do_rename):
"""Return the given Agent with its grounding mapped.
This function grounds a single agent. It returns the new Agent object
(which might be a different object if we load a new agent state
from json) or the same object otherwise.
Parameters
----------
agent : :py:class:`indra.statements.Agent`
The Agent to map.
do_rename: bool
If True, the Agent name is updated based on the mapped grounding.
If do_rename is True the priority for setting the name is
FamPlex ID, HGNC symbol, then the gene name
from Uniprot.
Returns
-------
grounded_agent : :py:class:`indra.statements.Agent`
The grounded Agent.
maps_to_none : bool
True if the Agent is in the grounding map and maps to None.
"""
agent_text = agent.db_refs.get('TEXT')
mapped_to_agent_json = self.agent_map.get(agent_text)
if mapped_to_agent_json:
mapped_to_agent = \
Agent._from_json(mapped_to_agent_json['agent'])
return mapped_to_agent, False
# Look this string up in the grounding map
# If not in the map, leave agent alone and continue
if agent_text in self.gm.keys():
map_db_refs = self.gm[agent_text]
else:
return agent, False
# If it's in the map but it maps to None, then filter out
# this statement by skipping it
if map_db_refs is None:
# Increase counter if this statement has not already
# been skipped via another agent
logger.debug("Skipping %s" % agent_text)
return None, True
# If it has a value that's not None, map it and add it
else:
# Otherwise, update the agent's db_refs field
self.update_agent_db_refs(agent, agent_text, do_rename)
return agent, False
def map_agent(self, agent, do_rename):
"""Grounds an agent; returns the new agent object (which might be
a different object if we load a new agent state from javascript).
Parameters
----------
agent: indra.statements.Agent
The agent to map
do_rename: bool
Whether to rename the agent text
Returns
-------
grounded_agent: indra.statements.Agent
The grounded agent
maps_to_none: bool
Whether the agent is in the grounding map and maps to None
"""
agent_text = agent.db_refs.get('TEXT')
mapped_to_agent_json = self.agent_map.get(agent_text)
if mapped_to_agent_json:
mapped_to_agent = \
Agent._from_json(mapped_to_agent_json['agent'])
return mapped_to_agent, False
# Look this string up in the grounding map
# If not in the map, leave agent alone and continue
if agent_text in self.gm.keys():
map_db_refs = self.gm[agent_text]
else:
return agent, False
# If it's in the map but it maps to None, then filter out
# this statement by skipping it
if map_db_refs is None:
# Increase counter if this statement has not already
# been skipped via another agent
logger.debug("Skipping %s" % agent_text)
return None, True
# If it has a value that's not None, map it and add it
else:
# Otherwise, update the agent's db_refs field
self.update_agent_db_refs(agent, agent_text, do_rename)
return agent, False
def map_agents(self, stmts, do_rename=True):
# Make a copy of the stmts
mapped_stmts = []
num_skipped = 0
# Iterate over the statements
for stmt in stmts:
mapped_stmt = deepcopy(stmt)
# Iterate over the agents
skip_stmt = False
mapped_agent_list = mapped_stmt.agent_list()
for idx, agent in enumerate(mapped_agent_list):
if agent is None or agent.db_refs.get('TEXT') is None:
continue
agent_text = agent.db_refs.get('TEXT')
mapped_to_agent_json = self.agent_map.get(agent_text)
if mapped_to_agent_json:
mapped_to_agent = \
Agent._from_json(mapped_to_agent_json['agent'])
mapped_agent_list[idx] = mapped_to_agent
mapped_stmt.set_agent_list(mapped_agent_list)
# Look this string up in the grounding map
# If not in the map, leave agent alone and continue
try:
map_db_refs = self.gm[agent_text]
except KeyError:
continue
# If it's in the map but it maps to None, then filter out
# this statement by skipping it
if map_db_refs is None:
# Increase counter if this statement has not already
# been skipped via another agent
if not skip_stmt:
num_skipped += 1
logger.debug("Skipping %s" % agent_text)
skip_stmt = True
# If it has a value that's not None, map it and add it
else:
# Otherwise, update the agent's db_refs field
gene_name = None
map_db_refs = deepcopy(self.gm.get(agent_text))
up_id = map_db_refs.get('UP')
hgnc_sym = map_db_refs.get('HGNC')
if up_id and not hgnc_sym:
gene_name = uniprot_client.get_gene_name(up_id, False)
if gene_name:
hgnc_id = hgnc_client.get_hgnc_id(gene_name)
if hgnc_id:
map_db_refs['HGNC'] = hgnc_id
elif hgnc_sym and not up_id:
# Override the HGNC symbol entry from the grounding
# map with an HGNC ID
hgnc_id = hgnc_client.get_hgnc_id(hgnc_sym)
if hgnc_id:
map_db_refs['HGNC'] = hgnc_id
# Now get the Uniprot ID for the gene
up_id = hgnc_client.get_uniprot_id(hgnc_id)
if up_id:
map_db_refs['UP'] = up_id
# If there's no HGNC ID for this symbol, raise an
# Exception
else:
raise ValueError('No HGNC ID corresponding to gene '
'symbol %s in grounding map.' %
hgnc_sym)
# If we have both, check the gene symbol ID against the
# mapping from Uniprot
elif up_id and hgnc_sym:
# Get HGNC Symbol from Uniprot
gene_name = uniprot_client.get_gene_name(up_id)
if not gene_name:
raise ValueError('No gene name found for Uniprot '
'ID %s (expected %s)' %
(up_id, hgnc_sym))
# We got gene name, compare it to the HGNC name
else:
if gene_name != hgnc_sym:
raise ValueError('Gene name %s for Uniprot ID '
'%s does not match HGNC '
'symbol %s given in grounding '
'map.' %
(gene_name, up_id, hgnc_sym))
else:
hgnc_id = hgnc_client.get_hgnc_id(hgnc_sym)
if not hgnc_id:
raise ValueError('No HGNC ID '
'corresponding to gene '
'symbol %s in grounding '
'map.' % hgnc_sym)
# Assign the DB refs from the grounding map to the agent
agent.db_refs = map_db_refs
# Are we renaming right now?
if do_rename:
# If there's a Bioentities ID, prefer that for the name
if agent.db_refs.get('BE'):
agent.name = agent.db_refs.get('BE')
# Get the HGNC symbol or gene name (retrieved above)
elif hgnc_sym is not None:
agent.name = hgnc_sym
elif gene_name is not None:
agent.name = gene_name
# Check if we should skip the statement
if not skip_stmt:
mapped_stmts.append(mapped_stmt)
logger.info('%s statements filtered out' % num_skipped)
return mapped_stmts
def test_all_target_list():
d = DTDA()
assert d.all_targets
assert all('HGNC' in Agent._from_json(e).db_refs.keys()
for e in d.all_targets)
def test_all_drug_list():
d = DTDA()
assert d.all_drugs
assert all('HMS-LINCS' in Agent._from_json(e).db_refs.keys()
for e in d.all_drugs)