ClearML/clearml
1
0
Fork 0
mirror of https://github.com/clearml/clearml synced 2025-06-26 18:16:07 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Dataset lineage view is broken with multiple dataset dependencies

Browse Source
This commit is contained in:
allegroai 2022-12-22 21:53:23 +02:00
parent 3dee5854bd
commit 6529c0dc11
1 changed files with 29 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
clearml/datasets/dataset.py
View File

@ -2747,22 +2747,48 @@ class Dataset(object):
current_index = 0 current_index = 0
dataset_struct = {} dataset_struct = {}
indices = {} indices = {}
dependency_graph_ex = deepcopy(self._dependency_graph) dependency_graph_ex_copy = deepcopy(self._dependency_graph)
# Make sure that id we reference a node as a parent, they exist on the DAG itself
for parents in self._dependency_graph.values(): for parents in self._dependency_graph.values():
for parent in parents: for parent in parents:
if parent not in self._dependency_graph: if parent not in self._dependency_graph:
dependency_graph_ex[parent] = [] dependency_graph_ex_copy[parent] = []
for id_, parents in dependency_graph_ex.items(): # get data from the parent versions
dependency_graph_ex = {}
while dependency_graph_ex_copy:
id_, parents = dependency_graph_ex_copy.popitem()
dependency_graph_ex[id_] = parents
task = Task.get_task(task_id=id_) task = Task.get_task(task_id=id_)
dataset_struct_entry = {"job_id": id_, "status": task.status} dataset_struct_entry = {"job_id": id_, "status": task.status}
# noinspection PyProtectedMember # noinspection PyProtectedMember
last_update = task._get_last_update() last_update = task._get_last_update()
if last_update: if last_update:
last_update = calendar.timegm(last_update.timetuple()) last_update = calendar.timegm(last_update.timetuple())
# fetch the parents of this version (task) based on what we have on the Task itself.
# noinspection PyBroadException
try:
dataset_version_node = task.get_configuration_object_as_dict("Dataset Struct")
# fine the one that is us
for node in dataset_version_node.values():
if node["job_id"] != id_:
continue
for parent in node.get("parents", []):
parent_id = dataset_version_node[parent]["job_id"]
if parent_id not in dependency_graph_ex_copy and parent_id not in dependency_graph_ex:
# add p to dependency_graph_ex
dependency_graph_ex_copy[parent_id] = []
if parent_id not in parents:
parents.append(parent_id)
break
except Exception:
pass
dataset_struct_entry["last_update"] = last_update dataset_struct_entry["last_update"] = last_update
dataset_struct_entry["parents"] = parents dataset_struct_entry["parents"] = parents
# noinspection PyProtectedMember
dataset_struct_entry["job_size"] = task._get_runtime_properties().get("ds_total_size") dataset_struct_entry["job_size"] = task._get_runtime_properties().get("ds_total_size")
dataset_struct_entry["name"] = task.name dataset_struct_entry["name"] = task.name
# noinspection PyProtectedMember
dataset_struct_entry["version"] = task._get_runtime_properties().get("version") dataset_struct_entry["version"] = task._get_runtime_properties().get("version")
dataset_struct[str(current_index)] = dataset_struct_entry dataset_struct[str(current_index)] = dataset_struct_entry
indices[id_] = str(current_index) indices[id_] = str(current_index)