Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Added use of subdirs for cache items. #11

Draft
wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

Added use of subdirs for cache items. #11

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 21 additions & 2 deletions src/mgds/pipelineModules/DiskCache.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -156,6 +156,21 @@ def __clone_for_cache(self, x: Any):
return x.clone()
return x

"""
Utility func to standardize splitting up items into scalable
subdirs. This allows better scaling than 100,000 items in
a single subdirectory. Currently we divide by 10,000

index "12" returns "0", "0/12.pt"
index "21034" returns "2", "2/21034.pt"
index "56712345" returns "5671", "5671/56712345.pt"
"""
def __item_subpaths(self, indexnum) -> tuple[str,str]:
dirnum=str(indexnum // 10000)
filename=str(indexnum) + '.pt'
return dirnum, dirnum + '/' + filename


def __refresh_cache(self, out_variation: int):
if not self.variations_initialized:
self.__init_variations()
Expand Down Expand Up @@ -196,7 +211,9 @@ def fn(group_index, in_index, in_variation):
for name in self.aggregate_names:
aggregate_item[name] = self.__clone_for_cache(self._get_previous_item(in_variation, name, in_index))

torch.save(split_item, os.path.realpath(os.path.join(cache_dir, str(group_index) + '.pt')))
dirname,itempath = self.__item_subpaths(group_index)
os.makedirs(os.path.realpath(os.path.join(cache_dir, dirname)), exist_ok=True)
torch.save(split_item, os.path.realpath(os.path.join(cache_dir, itempath)))
aggregate_cache[group_index] = aggregate_item

fs = (self._state.executor.submit(
Expand Down Expand Up @@ -250,7 +267,9 @@ def get_item(self, index: int, requested_name: str = None) -> dict:
item[name] = aggregate_item[name]

elif requested_name in self.split_names:
cache_path = os.path.join(self.__get_cache_dir(group_key, in_variation), str(group_index) + '.pt')
_, item_path = self.__item_subpaths(group_index)

cache_path = os.path.join(self.__get_cache_dir(group_key, in_variation), item_path)
split_item = torch.load(os.path.realpath(cache_path), weights_only=False)

for name in self.split_names:
Expand Down