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

Jump to bottom

Feature: Panel splitter classes #1220

Merged

emhbrine merged 7 commits into develop from feature/splitter

Nov 17, 2023

Contributor

emhbrine commented Nov 9, 2023 •

edited

The original PanelTimeSeriesSplit class is refactored into three classes:

ForwardPanelSplit
- Maintains the functionality of PanelTimeSeriesSplit with n_split_method="expanding"
KFoldPanelSplit
- Maintains the functionality of PanelTimeSeriesSplit with n_split_method="rolling"
IntervalPanelSplit
- Maintains the functionality of PanelTimeSeriesSplit with train_intervals

The split() methods now yield the splits instead of a returning a list of all splits.

emhbrine force-pushed the feature/splitter branch 4 times, most recently from 3c90dad to 1ef087b Compare

November 10, 2023 10:16

codecov bot commented Nov 10, 2023 •

edited

Codecov Report

Merging #1220 (4a6755b) into develop (841a957) will increase coverage by 0.37%.
Report is 1 commits behind head on develop.
The diff coverage is 59.88%.

❗ Current head 4a6755b differs from pull request most recent head 1af7db2. Consider uploading reports for the commit 1af7db2 to get more accurate results

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1220      +/-   ##
===========================================
+ Coverage    75.97%   76.35%   +0.37%     
===========================================
  Files           60       60              
  Lines         5345     5349       +4     
===========================================
+ Hits          4061     4084      +23     
+ Misses        1284     1265      -19

Files	Coverage Δ
macrosynergy/learning/__init__.py	`100.00% <100.00%> (ø)`
macrosynergy/learning/metrics.py	`19.48% <100.00%> (ø)`
macrosynergy/learning/prediction_tools.py	`15.92% <50.00%> (ø)`
macrosynergy/learning/cv_tools.py	`13.79% <33.33%> (ø)`
macrosynergy/learning/panel_time_series_split.py	`60.00% <60.00%> (ø)`

emhbrine marked this pull request as ready for review

November 10, 2023 12:58

emhbrine requested review from rushilg99 and a team as code owners

November 10, 2023 12:58

emhbrine force-pushed the feature/splitter branch from da64d8c to 55b57f8 Compare

November 10, 2023 15:03

emhbrine commented

View reviewed changes

macrosynergy/learning/panel_time_series_split.py

+                      self.n_splits = n_splits
+                      self.n_folds = n_splits + 1
+                  def split(

Contributor Author

emhbrine Nov 10, 2023 •

edited

Add return type hints

emhbrine requested a review from rsueppel

November 13, 2023 10:18

Contributor

rushilg99 commented Nov 15, 2023 •

edited

To be reviewed on my end now.

emhbrine added 5 commits

November 17, 2023 14:26


          refactor PanelTimeSeriesSplit

cdecdc8


          split cross validation methods

e021345


          rename splitter classes

2485b40


          add splitter tests

cc173bc


          add type hints

emhbrine force-pushed the feature/splitter branch from cc27bac to 7a10702 Compare

November 17, 2023 15:15


          fix merge conflicts

44ce2b7

emhbrine force-pushed the feature/splitter branch from 7a10702 to 44ce2b7 Compare

November 17, 2023 15:18

rushilg99 previously approved these changes

View reviewed changes

macrosynergy/learning/panel_time_series_split.py

+                          sorted(Xy.index.get_level_values(0).unique())
+                      )
+                      real_dates = Xy.index.get_level_values(1).unique().sort_values()

Contributor

rushilg99 Nov 17, 2023

Add:
freq_est = pd.infer_freq(real_dates)
freq_offset = pd.tseries.frequencies.to_offset(freq_est)

macrosynergy/learning/panel_time_series_split.py Outdated

+                              xranges_train: List[
+                                  Tuple[pd.Timestamp, pd.Timedelta]
+                              ] = self._calculate_xranges(cs_train_dates, real_dates)

Contributor

rushilg99 Nov 17, 2023

Add freq_offset as an argument to calculate_xranges

macrosynergy/learning/panel_time_series_split.py Outdated

+                          )
+                  def _calculate_xranges(
+                      self, cs_dates: pd.DatetimeIndex, real_dates: pd.DatetimeIndex

Contributor

rushilg99 Nov 17, 2023

Add argument
:param <pd.DateOffset> freq_offset: DateOffset object representing the frequency of the dates in the panel.

macrosynergy/learning/panel_time_series_split.py Outdated

+                      # A single contiguous range of dates.
+                      if len(difference) == 0:
+                          xranges.append(
+                              (cs_dates.min(), cs_dates.max() - cs_dates.min() + pd.Timedelta(days=1))

Contributor

rushilg99 Nov 17, 2023

replace pd.Timedelta(days=1) with freq_offset

macrosynergy/learning/panel_time_series_split.py Outdated

+                              difference = difference[(difference >= cs_dates.min())]
+                          xranges.append(
+                              (cs_dates.min(), cs_dates.max() - cs_dates.min() + pd.Timedelta(days=1))

Contributor

rushilg99 Nov 17, 2023

replace pd.Timedelta(days=1) with freq_offset


          change splitter naming

1af7db2

emhbrine dismissed rushilg99’s stale review via

1af7db2

November 17, 2023 16:20

rushilg99 self-requested a review

November 17, 2023 16:29

rushilg99 approved these changes

View reviewed changes

emhbrine merged commit c5c3327 into develop

4 checks passed

emhbrine deleted the feature/splitter branch

November 17, 2023 16:30

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment