Update pipelines to handle integer timestamp #241
Conversation
| @@ -173,6 +176,7 @@ def _forecast(self, ts: TSDataset, return_components: bool) -> TSDataset: | |||
|
|
|||
| return prediction_ts | |||
|
|
|||
| # TODO: зачем это надо? Мб удалить? | |||
There was a problem hiding this comment.
We should discuss this.
There was a problem hiding this comment.
It seems that everything okey here, what do you want to discuss?
There was a problem hiding this comment.
I can't wee why do we need this thing. It seems to me it does nothing.
|
🚀 Deployed on https://deploy-preview-241--etna-docs.netlify.app |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## unaligned-data #241 +/- ##
=================================================
Coverage ? 90.00%
=================================================
Files ? 198
Lines ? 13330
Branches ? 0
=================================================
Hits ? 11997
Misses ? 1333
Partials ? 0 ☔ View full report in Codecov by Sentry. |
| @pytest.fixture | ||
| def voting_ensemble_pipeline_int_timestamp( |
There was a problem hiding this comment.
Why we use "int_timestamp" part in naming? It seems, that this pipeline can work with normal timestamps too.
There was a problem hiding this comment.
Yes, it can. But I created it exactly to work in integer timestamp. It replaces prophet with sarimax in voting_ensemble_pipeline.
| if ts.freq is None: | ||
| expected_index = pd.Index(np.arange(start_timestamp, end_timestamp + 1), name="timestamp") | ||
| else: | ||
| expected_index = pd.date_range( | ||
| start=start_timestamp, periods=num_points, freq=predict_ts.freq, name="timestamp" | ||
| ) |
There was a problem hiding this comment.
Perhaps there should be a separate function for the creation of the index? We use this logic in many places.
There was a problem hiding this comment.
Yes, I'll add this logic in a separate PR.
Before submitting (must do checklist)
Proposed Changes
Look #221.
Closing issues
Closes #221.