def store_dataframe_in_db(data_frame, pk):
"""
Given a dataframe and the primary key of a workflow, it dumps its content on
a table that is rewritten every time.
:param data_frame: Pandas data frame containing the data
:param pk: The unique key for the workflow
:return: Nothing. Side effect in the database
"""
return store_table_in_db(data_frame, pk, create_table_name(pk))
def store_dataframe_in_db(data_frame, pk, reset_keys=True):
"""
Given a dataframe and the primary key of a workflow, it dumps its content on
a table that is rewritten every time.
:param data_frame: Pandas data frame containing the data
:param pk: The unique key for the workflow
:param reset_keys: Reset the field is_key computing its value from scratch
:return: Nothing. Side effect in the database
"""
return store_table_in_db(data_frame,
pk,
create_table_name(pk),
reset_keys=reset_keys)
def create(self, validated_data, **kwargs):
# Initial values
workflow_obj = None
try:
workflow_obj = Workflow(
user=self.context['user'],
name=self.context['name'],
description_text=validated_data['description_text'],
nrows=0,
ncols=0,
attributes=validated_data['attributes'],
query_builder_ops=validated_data.get('query_builder_ops', {}))
workflow_obj.save()
# Create the columns
column_data = ColumnSerializer(data=validated_data.get(
'columns', []),
many=True,
context={'workflow': workflow_obj})
# And save its content
if column_data.is_valid():
column_data.save()
else:
raise Exception('Unable to save column information')
# If there is any column with position = 0, recompute (this is to
# guarantee backward compatibility.
if workflow_obj.columns.filter(position=0).exists():
for idx, c in enumerate(workflow_obj.columns.all()):
c.position = idx + 1
c.save()
# Load the data frame
data_frame = validated_data.get('data_frame', None)
if data_frame is not None:
ops.store_dataframe_in_db(data_frame, workflow_obj.id)
# Reconcile now the information in workflow and columns with the
# one loaded
workflow_obj.data_frame_table_name = \
pandas_db.create_table_name(workflow_obj.pk)
workflow_obj.ncols = validated_data['ncols']
workflow_obj.nrows = validated_data['nrows']
workflow_obj.save()
# Create the actions pointing to the workflow
action_data = ActionSerializer(data=validated_data.get(
'actions', []),
many=True,
context={'workflow': workflow_obj})
if action_data.is_valid():
action_data.save()
else:
raise Exception('Unable to save column information')
# Create the views pointing to the workflow
view_data = ViewSerializer(data=validated_data.get('views', []),
many=True,
context={'workflow': workflow_obj})
if view_data.is_valid():
view_data.save()
else:
raise Exception('Unable to save column information')
except Exception:
# Get rid of the objects created
if workflow_obj:
if workflow_obj.has_data_frame():
pandas_db.delete_table(workflow_obj.id)
if workflow_obj.id:
workflow_obj.delete()
raise
return workflow_obj
def create(self, validated_data, **kwargs):
workflow_obj = Workflow(
user=self.context['user'],
name=self.context['name'],
description_text=validated_data['description_text'],
nrows=0,
ncols=0,
attributes=validated_data['attributes'],
query_builder_ops=validated_data.get('query_builder_ops', {})
)
workflow_obj.save()
# Create the columns
column_data = ColumnSerializer(
data=validated_data.get('columns', []),
many=True,
context={'workflow': workflow_obj})
# And save its content
if column_data.is_valid():
column_data.save()
else:
workflow_obj.delete()
return None
# Create the actions pointing to the workflow
action_data = ActionSerializer(
data=validated_data.get('actions', []),
many=True,
context={'workflow': workflow_obj}
)
if action_data.is_valid():
action_data.save()
else:
workflow_obj.delete()
return None
# Create the views pointing to the workflow
view_data = ViewSerializer(
data=validated_data.get('views', []),
many=True,
context={'workflow': workflow_obj}
)
if view_data.is_valid():
view_data.save()
else:
workflow_obj.delete()
return None
# Load the data frame
data_frame = validated_data.get('data_frame', None)
if data_frame is not None:
ops.store_dataframe_in_db(data_frame, workflow_obj.id)
# Reconcile now the information in workflow and columns with the
# one loaded
workflow_obj.data_frame_table_name = \
pandas_db.create_table_name(workflow_obj.pk)
workflow_obj.ncols = validated_data['ncols']
workflow_obj.nrows = validated_data['nrows']
workflow_obj.save()
return workflow_obj
def get_queryset_by_workflow_id(workflow_id):
return get_table_queryset(create_table_name(workflow_id))
def workflow_id_has_table(workflow_id):
return is_table_in_db(create_table_name(workflow_id))
