ClearML/clearml
1
0
Fork 0
mirror of https://github.com/clearml/clearml synced 2025-01-31 09:07:00 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

External files: Improve performance (#962)

Browse Source 
* External files: Improve performance
Add option for faster .add_external_files() option
Add option for parallel external file download

* Fix bug in add_external_files

* Remove max_workers_external_files parameter
Integrate PR feedback

* Integrate feedback

* Fix PEP suggestion
This commit is contained in:
john-zielke-snkeos 2023-04-24 15:21:08 +02:00 committed by GitHub
parent 71e1608ad8
commit a07f396f17
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 33 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
clearml/datasets/dataset.py
View File

@ -1,4 +1,5 @@
import calendar
import itertools
import json
import os
import shutil
@ -416,7 +417,7 @@ class Dataset(object):
self,
source_url, # type: Union[str, Sequence[str]]
wildcard=None, # type: Optional[Union[str, Sequence[str]]]
dataset_path=None, # type: Optional[str]
dataset_path=None, # type: Optional[Union[str,Sequence[str]]]
recursive=True, # type: bool
verbose=False, # type: bool
max_workers=None # type: Optional[int]
@ -458,14 +459,21 @@ class Dataset(object):
source_url_list = source_url if not isinstance(source_url, str) else [source_url]
max_workers = max_workers or psutil.cpu_count()
futures_ = []
if isinstance(dataset_path, str) or dataset_path is None:
dataset_paths = itertools.repeat(dataset_path)
else:
if len(dataset_path) != len(source_url):
raise ValueError("dataset_path must be a string or a list of strings with the same length as source_url"
f" (received {len(dataset_path)} paths for {len(source_url)} source urls))")
dataset_paths = dataset_path
with ThreadPoolExecutor(max_workers=max_workers) as tp:
for source_url_ in source_url_list:
for source_url_, dataset_path_ in zip(source_url_list, dataset_paths):
futures_.append(
tp.submit(
self._add_external_files,
source_url_,
wildcard=wildcard,
dataset_path=dataset_path,
dataset_path=dataset_path_,
recursive=recursive,
verbose=verbose,
)
@ -2194,12 +2202,12 @@ class Dataset(object):
max_workers=max_workers
)
self._download_external_files(
target_folder=target_folder, lock_target_folder=lock_target_folder
target_folder=target_folder, lock_target_folder=lock_target_folder, max_workers=max_workers
)
return local_folder
def _download_external_files(
self, target_folder=None, lock_target_folder=False
self, target_folder=None, lock_target_folder=False, max_workers=None
):
# (Union(Path, str), bool) -> None
"""
@ -2211,6 +2219,7 @@ class Dataset(object):
:param target_folder: If provided use the specified target folder, default, auto generate from Dataset ID.
:param lock_target_folder: If True, local the target folder so the next cleanup will not delete
Notice you should unlock it manually, or wait for the process to finish for auto unlocking.
:param max_workers: Number of threads to be spawned when getting dataset files. Defaults to no multi-threading.
"""
target_folder = (
Path(target_folder)
@ -2227,15 +2236,14 @@ class Dataset(object):
ds = Dataset.get(dependency)
links.update(ds._dataset_link_entries)
links.update(self._dataset_link_entries)
for relative_path, link in links.items():
target_path = os.path.join(target_folder, relative_path)
def _download_link(link,target_path):
if os.path.exists(target_path):
LoggerRoot.get_base_logger().info(
"{} already exists. Skipping downloading {}".format(
target_path, link
LoggerRoot.get_base_logger().info(
"{} already exists. Skipping downloading {}".format(
target_path, link
)
)
)
continue
return
ok = False
error = None
try:
@ -2257,6 +2265,19 @@ class Dataset(object):
LoggerRoot.get_base_logger().info(log_string)
else:
link.size = Path(target_path).stat().st_size
if not max_workers:
for relative_path, link in links.items():
target_path = os.path.join(target_folder, relative_path)
_download_link(link,target_path)
else:
with ThreadPoolExecutor(max_workers=max_workers) as pool:
for relative_path, link in links.items():
target_path = os.path.join(target_folder, relative_path)
pool.submit(_download_link,link,target_path)
def _extract_dataset_archive(
self,