Skip to content

[ENH] Improve performance of TimeSeriesDataSet.__getitem__ #806

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

Open
denix56 wants to merge 7 commits into
base: main
Choose a base branch
from
Open

[ENH] Improve performance of TimeSeriesDataSet.__getitem__ #806

Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
39a28ba
Replace DataFrame of indices with np.recarray
denix56 Dec 21, 2021
46646c0
Enable index field (to avoid changes in other files)
denix56 Dec 22, 2021
08cb636
Fix conversion to numpy, when we have numpy already
denix56 Dec 26, 2021
eb706f9
Remove whitespaces
denix56 Dec 28, 2021
957b9c3
Merge branch 'main' into pr/806
fkiraly Jun 8, 2025
8a8f57b
manual merge
fkiraly Jun 8, 2025
c0788fa
Update _timeseries.py
fkiraly Jun 8, 2025
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
18 changes: 11 additions & 7 deletions pytorch_forecasting/data/timeseries/_timeseries.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1860,7 +1860,7 @@ def _construct_index(self, data: pd.DataFrame, predict_mode: bool) -> pd.DataFra
)
assert len(df_index) > 0, msg

return df_index
return df_index.to_records(index=True)

def filter(self, filter_func: Callable, copy: bool = True) -> TimeSeriesDataType:
"""Filter subsequences in dataset.
Expand Down Expand Up @@ -1909,8 +1909,8 @@ def decoded_index(self) -> pd.DataFrame:
pd.DataFrame: index that can be understood in terms of original data
"""
# get dataframe to filter
index_start = self.index["index_start"].to_numpy()
index_last = self.index["index_end"].to_numpy()
index_start = self.index["index_start"]
index_last = self.index["index_end"]
index = (
# get group ids in order of index
pd.DataFrame(
Expand Down Expand Up @@ -2091,16 +2091,20 @@ def __getitem__(self, idx: int) -> tuple[dict[str, torch.Tensor], torch.Tensor]:
Returns:
tuple[dict[str, torch.Tensor], torch.Tensor]: x and y for model
"""
index = self.index.iloc[idx]
index = self.index[idx]

# get index data
index_start = index.index_start
index_end = index.index_end
index_sequence_length = index.sequence_length
# slice data based on index
idx_slice = slice(index.index_start, index.index_end + 1)
idx_slice = slice(index_start, index_end + 1)

data_cont = self.data["reals"][idx_slice].clone()
data_cat = self.data["categoricals"][idx_slice].clone()
time = self.data["time"][idx_slice].clone()
target = [d[idx_slice].clone() for d in self.data["target"]]
groups = self.data["groups"][index.index_start].clone()
groups = self.data["groups"][index_start].clone()
if self.data["weight"] is None:
weight = None
else:
Expand All @@ -2112,7 +2116,7 @@ def __getitem__(self, idx: int) -> tuple[dict[str, torch.Tensor], torch.Tensor]:

# fill in missing values (if not all time indices are specified)
sequence_length = len(time)
if sequence_length < index.sequence_length:
if sequence_length < index_sequence_length:
assert (
self.allow_missing_timesteps
), "allow_missing_timesteps should be True if sequences have gaps"
Expand Down
Loading