def setup_config(data_type="databases"):
"""
Reads YAML configuration file and converts it into a Python dictionary.
:param data_type: configuration type ('databases', 'ontologies', 'experiments' or 'builder').
:return: Dictionary.
.. note:: This function should be used to obtain the configuration for databases_controller.py, \
ontologies_controller.py, experiments_controller.py and builder.py.
"""
try:
dirname = os.path.abspath(os.path.dirname(__file__))
if data_type == 'databases':
config = ckg_utils.get_configuration(
os.path.join(dirname, ckg_config.databases_config_file))
elif data_type == 'ontologies':
config = ckg_utils.get_configuration(
os.path.join(dirname, ckg_config.ontologies_config_file))
elif data_type == "experiments":
config = ckg_utils.get_configuration(
os.path.join(dirname, ckg_config.experiments_config_file))
elif data_type == 'builder':
config = ckg_utils.get_configuration(
os.path.join(dirname, ckg_config.builder_config_file))
elif data_type == 'users':
config = ckg_utils.get_configuration(
os.path.join(dirname, ckg_config.users_config_file))
except Exception as err:
raise Exception(
"builder_utils - Reading configuration > {}.".format(err))
return config
def set_configuration_from_file(self, configuration_file):
try:
cwd = os.path.abspath(os.path.dirname(__file__))
config_path = os.path.join("config/", configuration_file)
self.configuration = ckg_utils.get_configuration(os.path.join(cwd, config_path))
except Exception as err:
logger.error("Error: {} reading configuration file: {}.".format(err, config_path))
def get_config(config_name, data_type='databases'):
"""
Reads YAML configuration file and converts it into a Python dictionary.
:param str config_name: name of the configuration YAML file.
:param str data_type: configuration type ('databases' or 'ontologies').
:return: Dictionary.
.. note:: Use this function to obtain configuration for individual database/ontology parsers.
"""
cwd = os.path.abspath(os.path.dirname(__file__))
config = ckg_utils.get_configuration(
os.path.join(cwd, '{}/config/{}'.format(data_type, config_name)))
return config
def get_full_path_directories():
"""
Reads Builder YAML configuration file and returns the full path of all directories.
:return: Dictionary.
"""
directories = {}
try:
dirname = os.path.abspath(os.path.dirname(__file__))
config = ckg_utils.get_configuration(
os.path.join(dirname, ckg_config.builder_config_file))
if 'directories' in config:
for directory in config['directories']:
directories[directory] = os.path.join(
dirname, config['directories'][directory])
except Exception as err:
raise Exception(
"Error {}: builder_utils - Reading directories from configuration > {}."
.format(err, ckg_config.builder_config_file))
return directories
import sys
import os
import json
import neo4j
import pandas as pd
import ckg_utils
from config import ckg_config
from graphdb_builder import builder_utils
log_config = ckg_config.graphdb_connector_log
logger = builder_utils.setup_logging(log_config, key="connector")
try:
cwd = os.path.abspath(os.path.dirname(__file__))
path = os.path.join(cwd, ckg_config.connector_config_file)
config = ckg_utils.get_configuration(path)
except Exception as err:
logger.error("Reading configuration > {}.".format(err))
def getGraphDatabaseConnectionConfiguration(configuration=None, database=None):
if configuration is None:
configuration = config
host = configuration['db_url']
port = configuration['db_port']
user = configuration['db_user']
password = configuration['db_password']
if database is not None:
host = host + '/' + database
def build_project(self, force=False):
if self.check_report_exists() and not force:
self.load_project_report()
elif force:
self.report = {}
self.datasets = {}
if len(self.report) == 0 or len(self.datasets) == 0:
project_info = self.query_data()
if len(project_info) > 0:
self.set_attributes(project_info)
self.get_similar_projects(project_info)
self.get_projects_overlap(project_info)
for data_type in self.data_types:
dataset = None
configuration = None
if data_type == "proteomics":
if "proteomics" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["proteomics"])
dataset = ProteomicsDataset(
self.identifier,
data={},
configuration=configuration,
analyses={},
analysis_queries={},
report=None)
elif data_type == "clinical":
if "clinical" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["clinical"])
dataset = ClinicalDataset(self.identifier,
data={},
configuration=configuration,
analyses={},
analysis_queries={},
report=None)
elif data_type == "wes" or data_type == "wgs":
if "wes" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["wes"])
elif "wgs" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["wgs"])
dataset = DNAseqDataset(self.identifier,
dataset_type=data_type,
data={},
configuration=configuration,
analyses={},
analysis_queries={},
report=None)
elif data_type == "interactomics":
if "interactomics" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["interactomics"])
dataset = InteractomicsDataset(
self.identifier,
data={},
configuration=configuration,
analyses={},
analysis_queries={},
report=None)
elif data_type == "phosphoproteomics":
if "phosphoproteomics" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["phosphoproteomics"])
dataset = PhosphoproteomicsDataset(
self.identifier,
data={},
configuration=configuration,
analyses={},
analysis_queries={},
report=None)
elif data_type == "longitudinal_proteomics":
if "longitudinal_proteomics" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.
configuration_files["longitudinal_proteomics"])
dataset = LongitudinalProteomicsDataset(
self.identifier,
data={},
configuration=configuration,
analyses={},
analysis_queries={},
report=None)
if dataset is not None:
dataset.generate_dataset()
self.update_dataset({data_type: dataset})
if len(self.datasets) > 1:
if "multiomics" in self.configuration_files:
configuration = ckg_utils.get_configuration(
self.configuration_files["multiomics"])
dataset = MultiOmicsDataset(self.identifier,
data=self.datasets,
configuration=configuration,
analyses={},
report=None)
self.update_dataset({'multiomics': dataset})
self.append_data_type('multiomics')
else:
logger.error(
"Project {} could not be built. Error retrieving information for this project or no information associated to this project"
.format(self.identifier))
print(
"Project {} could not be built. Error retrieving information for this project or no information associated to this project"
.format(self.identifier))
