From c887278706b2e276135c8980a45d2fa6a0d0953d Mon Sep 17 00:00:00 2001
From: Abdul Fatir <Abdulfatirs@gmail.com>
Date: Fri, 29 Nov 2024 10:33:41 +0100
Subject: [PATCH] Clean up evaluation script (#218)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

*Issue #, if available:*

*Description of changes:* This PR cleans up eval script by using
`DateTimeIndex.to_period()` instead of the ugly frequency mapping. Not
sure what I was doing before. 🫠


By submitting this pull request, I confirm that you can use, modify,
copy, and redistribute this contribution, under the terms of your
choice.

Co-authored-by: Abdul Fatir Ansari <ansarnd@amazon.de>
---
 scripts/evaluation/evaluate.py | 160 +--------------------------------
 1 file changed, 3 insertions(+), 157 deletions(-)

diff --git a/scripts/evaluation/evaluate.py b/scripts/evaluation/evaluate.py
index 8995e31..9c9acff 100644
--- a/scripts/evaluation/evaluate.py
+++ b/scripts/evaluation/evaluate.py
@@ -24,161 +24,6 @@ from chronos import (
 
 app = typer.Typer(pretty_exceptions_enable=False)
 
-# Taken from pandas._libs.tslibs.dtypes.OFFSET_TO_PERIOD_FREQSTR
-offset_alias_to_period_alias = {
-    "WEEKDAY": "D",
-    "EOM": "M",
-    "BME": "M",
-    "SME": "M",
-    "BQS": "Q",
-    "QS": "Q",
-    "BQE": "Q",
-    "BQE-DEC": "Q",
-    "BQE-JAN": "Q",
-    "BQE-FEB": "Q",
-    "BQE-MAR": "Q",
-    "BQE-APR": "Q",
-    "BQE-MAY": "Q",
-    "BQE-JUN": "Q",
-    "BQE-JUL": "Q",
-    "BQE-AUG": "Q",
-    "BQE-SEP": "Q",
-    "BQE-OCT": "Q",
-    "BQE-NOV": "Q",
-    "MS": "M",
-    "D": "D",
-    "B": "B",
-    "min": "min",
-    "s": "s",
-    "ms": "ms",
-    "us": "us",
-    "ns": "ns",
-    "h": "h",
-    "QE": "Q",
-    "QE-DEC": "Q-DEC",
-    "QE-JAN": "Q-JAN",
-    "QE-FEB": "Q-FEB",
-    "QE-MAR": "Q-MAR",
-    "QE-APR": "Q-APR",
-    "QE-MAY": "Q-MAY",
-    "QE-JUN": "Q-JUN",
-    "QE-JUL": "Q-JUL",
-    "QE-AUG": "Q-AUG",
-    "QE-SEP": "Q-SEP",
-    "QE-OCT": "Q-OCT",
-    "QE-NOV": "Q-NOV",
-    "YE": "Y",
-    "YE-DEC": "Y-DEC",
-    "YE-JAN": "Y-JAN",
-    "YE-FEB": "Y-FEB",
-    "YE-MAR": "Y-MAR",
-    "YE-APR": "Y-APR",
-    "YE-MAY": "Y-MAY",
-    "YE-JUN": "Y-JUN",
-    "YE-JUL": "Y-JUL",
-    "YE-AUG": "Y-AUG",
-    "YE-SEP": "Y-SEP",
-    "YE-OCT": "Y-OCT",
-    "YE-NOV": "Y-NOV",
-    "W": "W",
-    "ME": "M",
-    "Y": "Y",
-    "BYE": "Y",
-    "BYE-DEC": "Y",
-    "BYE-JAN": "Y",
-    "BYE-FEB": "Y",
-    "BYE-MAR": "Y",
-    "BYE-APR": "Y",
-    "BYE-MAY": "Y",
-    "BYE-JUN": "Y",
-    "BYE-JUL": "Y",
-    "BYE-AUG": "Y",
-    "BYE-SEP": "Y",
-    "BYE-OCT": "Y",
-    "BYE-NOV": "Y",
-    "YS": "Y",
-    "BYS": "Y",
-    "QS-JAN": "Q",
-    "QS-FEB": "Q",
-    "QS-MAR": "Q",
-    "QS-APR": "Q",
-    "QS-MAY": "Q",
-    "QS-JUN": "Q",
-    "QS-JUL": "Q",
-    "QS-AUG": "Q",
-    "QS-SEP": "Q",
-    "QS-OCT": "Q",
-    "QS-NOV": "Q",
-    "QS-DEC": "Q",
-    "BQS-JAN": "Q",
-    "BQS-FEB": "Q",
-    "BQS-MAR": "Q",
-    "BQS-APR": "Q",
-    "BQS-MAY": "Q",
-    "BQS-JUN": "Q",
-    "BQS-JUL": "Q",
-    "BQS-AUG": "Q",
-    "BQS-SEP": "Q",
-    "BQS-OCT": "Q",
-    "BQS-NOV": "Q",
-    "BQS-DEC": "Q",
-    "YS-JAN": "Y",
-    "YS-FEB": "Y",
-    "YS-MAR": "Y",
-    "YS-APR": "Y",
-    "YS-MAY": "Y",
-    "YS-JUN": "Y",
-    "YS-JUL": "Y",
-    "YS-AUG": "Y",
-    "YS-SEP": "Y",
-    "YS-OCT": "Y",
-    "YS-NOV": "Y",
-    "YS-DEC": "Y",
-    "BYS-JAN": "Y",
-    "BYS-FEB": "Y",
-    "BYS-MAR": "Y",
-    "BYS-APR": "Y",
-    "BYS-MAY": "Y",
-    "BYS-JUN": "Y",
-    "BYS-JUL": "Y",
-    "BYS-AUG": "Y",
-    "BYS-SEP": "Y",
-    "BYS-OCT": "Y",
-    "BYS-NOV": "Y",
-    "BYS-DEC": "Y",
-    "Y-JAN": "Y-JAN",
-    "Y-FEB": "Y-FEB",
-    "Y-MAR": "Y-MAR",
-    "Y-APR": "Y-APR",
-    "Y-MAY": "Y-MAY",
-    "Y-JUN": "Y-JUN",
-    "Y-JUL": "Y-JUL",
-    "Y-AUG": "Y-AUG",
-    "Y-SEP": "Y-SEP",
-    "Y-OCT": "Y-OCT",
-    "Y-NOV": "Y-NOV",
-    "Y-DEC": "Y-DEC",
-    "Q-JAN": "Q-JAN",
-    "Q-FEB": "Q-FEB",
-    "Q-MAR": "Q-MAR",
-    "Q-APR": "Q-APR",
-    "Q-MAY": "Q-MAY",
-    "Q-JUN": "Q-JUN",
-    "Q-JUL": "Q-JUL",
-    "Q-AUG": "Q-AUG",
-    "Q-SEP": "Q-SEP",
-    "Q-OCT": "Q-OCT",
-    "Q-NOV": "Q-NOV",
-    "Q-DEC": "Q-DEC",
-    "W-MON": "W-MON",
-    "W-TUE": "W-TUE",
-    "W-WED": "W-WED",
-    "W-THU": "W-THU",
-    "W-FRI": "W-FRI",
-    "W-SAT": "W-SAT",
-    "W-SUN": "W-SUN",
-}
-
 
 def to_gluonts_univariate(hf_dataset: datasets.Dataset):
     series_fields = [
@@ -188,8 +33,9 @@ def to_gluonts_univariate(hf_dataset: datasets.Dataset):
     ]
     series_fields.remove("timestamp")
     dataset_length = hf_dataset.info.splits["train"].num_examples * len(series_fields)
-    dataset_freq = pd.infer_freq(hf_dataset[0]["timestamp"])
-    dataset_freq = offset_alias_to_period_alias.get(dataset_freq, dataset_freq)
+
+    # Assumes that all time series in the dataset have the same frequency
+    dataset_freq = pd.DatetimeIndex(hf_dataset[0]["timestamp"]).to_period()[0].freqstr
 
     gts_dataset = []
     for hf_entry in hf_dataset: