import os
import shutil
import time
from IPython.display import display, IFrame, clear_output
import tempfile
from ..utils import Message, ClassError
from .jstreecontrol import create_concept_tree, ConceptTree
import tmtk
[docs]def call_boris(to_be_shuffled=None, **kwargs):
"""
This function loads the Arborist if it has been properly installed in your environment.
:param to_be_shuffled: has to be either a tmtk.Study object, a Pandas
column mapping dataframe, or a path to column mapping file.
"""
if not isinstance(to_be_shuffled, (tmtk.Clinical, tmtk.Study)):
Message.error("No path to column mapping file or a valid object given.")
raise ClassError(type(to_be_shuffled, 'tmtk.Clinical or tmtk.Study'))
json_data = create_concept_tree(to_be_shuffled)
json_data = launch_arborist_gui(json_data, **kwargs) # Returns modified json_data
if json_data:
Message.okay('Successfully closed The Arborist. The updated column'
' mapping file has been returned as a dataframe.')
Message.warning("Don't forget to save this dataframe to disk if you want to store it.")
else:
raise Exception('No json file returned from Arborist.')
if isinstance(to_be_shuffled, tmtk.Study):
update_study_from_json(to_be_shuffled, json_data=json_data)
elif isinstance(to_be_shuffled, tmtk.Clinical):
update_clinical_from_json(to_be_shuffled, json_data=json_data)
else:
raise TypeError("Arborist could not find object to update.")
[docs]def valid_arborist_or_exception(**kwargs):
from notebook import __version__ as notebook_version
if notebook_version < '4.2.0':
print("Version of notebook package should be atleast 4.2.0 for Arborist, consider:")
print(" $ pip3 install --upgrade notebook")
raise RuntimeError("Notebook too old for Arborist.")
from notebook.serverextensions import validate_serverextension
from notebook.nbextensions import check_nbextension
warnings = validate_serverextension('tmtk.arborist')
if warnings or not check_nbextension('transmart-arborist', **kwargs):
print('For the Arborist to work you need to install a jupyter serverextension using something like:')
print(' $ jupyter nbextension install --py tmtk.arborist')
print(' $ jupyter serverextension enable --py tmtk.arborist')
print('Then to verify installation run:')
print(' $ jupyter serverextension list')
raise RuntimeError('Transmart-arborist extension not found.')
[docs]def launch_arborist_gui(json_data: str, height=650):
"""
:param json_data:
:param height:
:return:
"""
new_temp_dir = tempfile.mkdtemp()
tmp_json = os.path.join(new_temp_dir, 'tmp_json')
with open(tmp_json, 'w') as f:
f.write(json_data)
# Signal created by Javascript to continue work here.
done_signal = os.path.join(os.path.dirname(tmp_json), 'DONE')
base_url = os.environ.get("ARBORIST_BASE_URL", "/")
running_on = '{}transmart-arborist?treefile={}'.format(base_url, os.path.abspath(tmp_json))
display(IFrame(src=running_on, width='100%', height=height))
try:
# Wait for the done signal file to be created before breaking the GIL.
while not os.path.exists(done_signal):
time.sleep(0.1)
except KeyboardInterrupt:
# This stops the interpreter without showing a stacktrace.
pass
else:
with open(tmp_json, 'r') as f:
json_data = f.read()
return json_data
finally:
shutil.rmtree(new_temp_dir)
# Clear output from Jupyter Notebook cell
clear_output()
print('Cleaning up before closing...')
[docs]def update_clinical_from_json(clinical, json_data):
"""
:param clinical:
:param json_data:
:return:
"""
concept_tree = ConceptTree(json_data)
clinical.ColumnMapping.df = concept_tree.column_mapping_file
clinical.WordMapping.df = concept_tree.word_mapping
[docs]def update_study_from_json(study, json_data):
"""
:param study:
:param json_data:
:return:
"""
concept_tree = ConceptTree(json_data)
study.Clinical.ColumnMapping.df = concept_tree.column_mapping_file
study.Clinical.WordMapping.df = concept_tree.word_mapping
# Some checks for whether to create Tags in the study.
ct_tags = concept_tree.tags_file
if hasattr(study, 'Tags') or ct_tags.shape[0]:
study.ensure_metadata()
study.Tags.df = concept_tree.tags_file
high_dim_paths = concept_tree.high_dim_paths
if high_dim_paths:
study.HighDim.update_high_dim_paths(high_dim_paths)