[BUG] Holt Winters Anomaly Strategy fails due to NoClassDefFoundError

[komashk]

Holt Winters Anomaly Strategy fails due to NoClassDefFoundError

To Reproduce
Steps to reproduce the behavior:

1. Replicating Glue tutorial for the anomaly detection using Holt Winters Anomaly Strategy.
2. Attempted with the following combinations of versions:

• AWS Glue version 4, Spark 3.3, pydeequ 1.4.0, deequ jar com.amazon.deequ:deequ:2.0.7-spark-3.3.
• AWS Glue version 3, Spark 3.1, pydeeqy 1.4.0, deequ jar com.amazon.deequ:deequ:2.0.7-spark-3.1

4. I ran the code below after generating the metrics for years 2013 and 2014:

for month in range(1,9):
    date = "\'2015" +'/'+str(month)+"\'"
    df = df_2015.filter("review_date =" + date)
    key_tags = {'tag':  date}
    result_key_2015 = ResultKey(session, ResultKey.current_milli_time(), key_tags)

    jewelry_result = VerificationSuite(session).onData(df)\
        .useRepository(metricsRepository) \
        .saveOrAppendResult(result_key_2015) \
        .addAnomalyCheck(HoltWinters(session, MetricInterval.Monthly, SeriesSeasonality.Yearly), Sum('total_votes'))\
        .run()

5. I get the following error:

Py4JJavaError: An error occurred while calling o560.run.
: java.lang.NoClassDefFoundError: breeze/stats/package$
	at com.amazon.deequ.anomalydetection.seasonal.HoltWinters.detect(HoltWinters.scala:243)
	at com.amazon.deequ.anomalydetection.AnomalyDetector.detectAnomaliesInHistory(AnomalyDetector.scala:98)
.....

Expected behavior
The expected behavior is to obtain a message that anomaly has been detected.

What can be done to fix this error? Thank you.