diff options
author | Xiangrui Meng <meng@databricks.com> | 2015-05-21 17:59:03 -0700 |
---|---|---|
committer | Xiangrui Meng <meng@databricks.com> | 2015-05-21 17:59:03 -0700 |
commit | f5db4b416c922db7a8f1b0c098b4f08647106231 (patch) | |
tree | 97ddc7467b77b169324bba8b0ec93c5d76005ca9 /python/pyspark/ml | |
parent | 17791a58159b3e4619d0367f54a4c5332342658b (diff) | |
download | spark-f5db4b416c922db7a8f1b0c098b4f08647106231.tar.gz spark-f5db4b416c922db7a8f1b0c098b4f08647106231.tar.bz2 spark-f5db4b416c922db7a8f1b0c098b4f08647106231.zip |
[SPARK-7794] [MLLIB] update RegexTokenizer default settings
The previous default is `{gaps: false, pattern: "\\p{L}+|[^\\p{L}\\s]+"}`. The default pattern is hard to understand. This PR changes the default to `{gaps: true, pattern: "\\s+"}`. jkbradley
Author: Xiangrui Meng <meng@databricks.com>
Closes #6330 from mengxr/SPARK-7794 and squashes the following commits:
5ee7cde [Xiangrui Meng] update RegexTokenizer default settings
Diffstat (limited to 'python/pyspark/ml')
-rw-r--r-- | python/pyspark/ml/feature.py | 40 |
1 files changed, 19 insertions, 21 deletions
diff --git a/python/pyspark/ml/feature.py b/python/pyspark/ml/feature.py index 5511dceb70..b0479d9b07 100644 --- a/python/pyspark/ml/feature.py +++ b/python/pyspark/ml/feature.py @@ -446,23 +446,25 @@ class PolynomialExpansion(JavaTransformer, HasInputCol, HasOutputCol): @ignore_unicode_prefix class RegexTokenizer(JavaTransformer, HasInputCol, HasOutputCol): """ - A regex based tokenizer that extracts tokens either by repeatedly matching the regex(default) - or using it to split the text (set matching to false). Optional parameters also allow filtering - tokens using a minimal length. + A regex based tokenizer that extracts tokens either by using the + provided regex pattern (in Java dialect) to split the text + (default) or repeatedly matching the regex (if gaps is true). + Optional parameters also allow filtering tokens using a minimal + length. It returns an array of strings that can be empty. - >>> df = sqlContext.createDataFrame([("a b c",)], ["text"]) + >>> df = sqlContext.createDataFrame([("a b c",)], ["text"]) >>> reTokenizer = RegexTokenizer(inputCol="text", outputCol="words") >>> reTokenizer.transform(df).head() - Row(text=u'a b c', words=[u'a', u'b', u'c']) + Row(text=u'a b c', words=[u'a', u'b', u'c']) >>> # Change a parameter. >>> reTokenizer.setParams(outputCol="tokens").transform(df).head() - Row(text=u'a b c', tokens=[u'a', u'b', u'c']) + Row(text=u'a b c', tokens=[u'a', u'b', u'c']) >>> # Temporarily modify a parameter. >>> reTokenizer.transform(df, {reTokenizer.outputCol: "words"}).head() - Row(text=u'a b c', words=[u'a', u'b', u'c']) + Row(text=u'a b c', words=[u'a', u'b', u'c']) >>> reTokenizer.transform(df).head() - Row(text=u'a b c', tokens=[u'a', u'b', u'c']) + Row(text=u'a b c', tokens=[u'a', u'b', u'c']) >>> # Must use keyword arguments to specify params. >>> reTokenizer.setParams("text") Traceback (most recent call last): @@ -472,31 +474,27 @@ class RegexTokenizer(JavaTransformer, HasInputCol, HasOutputCol): # a placeholder to make it appear in the generated doc minTokenLength = Param(Params._dummy(), "minTokenLength", "minimum token length (>= 0)") - gaps = Param(Params._dummy(), "gaps", "Set regex to match gaps or tokens") - pattern = Param(Params._dummy(), "pattern", "regex pattern used for tokenizing") + gaps = Param(Params._dummy(), "gaps", "whether regex splits on gaps (True) or matches tokens") + pattern = Param(Params._dummy(), "pattern", "regex pattern (Java dialect) used for tokenizing") @keyword_only - def __init__(self, minTokenLength=1, gaps=False, pattern="\\p{L}+|[^\\p{L}\\s]+", - inputCol=None, outputCol=None): + def __init__(self, minTokenLength=1, gaps=True, pattern="\\s+", inputCol=None, outputCol=None): """ - __init__(self, minTokenLength=1, gaps=False, pattern="\\p{L}+|[^\\p{L}\\s]+", \ - inputCol=None, outputCol=None) + __init__(self, minTokenLength=1, gaps=True, pattern="\\s+", inputCol=None, outputCol=None) """ super(RegexTokenizer, self).__init__() self._java_obj = self._new_java_obj("org.apache.spark.ml.feature.RegexTokenizer", self.uid) self.minTokenLength = Param(self, "minTokenLength", "minimum token length (>= 0)") - self.gaps = Param(self, "gaps", "Set regex to match gaps or tokens") - self.pattern = Param(self, "pattern", "regex pattern used for tokenizing") - self._setDefault(minTokenLength=1, gaps=False, pattern="\\p{L}+|[^\\p{L}\\s]+") + self.gaps = Param(self, "gaps", "whether regex splits on gaps (True) or matches tokens") + self.pattern = Param(self, "pattern", "regex pattern (Java dialect) used for tokenizing") + self._setDefault(minTokenLength=1, gaps=True, pattern="\\s+") kwargs = self.__init__._input_kwargs self.setParams(**kwargs) @keyword_only - def setParams(self, minTokenLength=1, gaps=False, pattern="\\p{L}+|[^\\p{L}\\s]+", - inputCol=None, outputCol=None): + def setParams(self, minTokenLength=1, gaps=True, pattern="\\s+", inputCol=None, outputCol=None): """ - setParams(self, minTokenLength=1, gaps=False, pattern="\\p{L}+|[^\\p{L}\\s]+", \ - inputCol="input", outputCol="output") + setParams(self, minTokenLength=1, gaps=True, pattern="\\s+", inputCol=None, outputCol=None) Sets params for this RegexTokenizer. """ kwargs = self.setParams._input_kwargs |