About an error regarding spark nlp using scala

Question

I am a beginner to spark-nlp and i am learning it by following examples in the johnsnowlabs. I am using SCALA in data bricks.

When i follow the example as follows,

import com.johnsnowlabs.nlp.base._
import com.johnsnowlabs.nlp.annotator._
import org.apache.spark.ml.Pipeline

val documentAssembler = new DocumentAssembler().
    setInputCol("text").
    setOutputCol("document")

val regexTokenizer = new Tokenizer().
    setInputCols(Array("sentence")).
    setOutputCol("token")
val sentenceDetector = new SentenceDetector().
    setInputCols(Array("document")).
    setOutputCol("sentence")

val finisher = new Finisher()
    .setInputCols("token")
    .setIncludeMetadata(true)


finisher.withColumn("newCol", explode(arrays_zip($"finished_token", $"finished_ner")))

I am getting following error when i run the last line :

command-786892578143744:2: error: value withColumn is not a member of com.johnsnowlabs.nlp.Finisher
finisher.withColumn("newCol", explode(arrays_zip($"finished_token", $"finished_ner")))

what may be the reason for this ?

When i try to do the example , by just omitting this line , i added follwoing additional lines of codes

val pipeline = new Pipeline().
    setStages(Array(
        documentAssembler,
        sentenceDetector,
        regexTokenizer,
        finisher
    ))

val data1 = Seq("hello, this is an example sentence").toDF("text")

pipeline.fit(data1).transform(data1).toDF("text")

I got another error when i run the last line :

java.lang.IllegalArgumentException: requirement failed: The number of columns doesn't match.

Can anyone help me to fix this issue ?

Thank you

Answer 1

Here what your code should look like, first construct the Pipeline:

import com.johnsnowlabs.nlp.base._
import com.johnsnowlabs.nlp.annotator._
import org.apache.spark.ml.Pipeline

val documentAssembler = new DocumentAssembler().
    setInputCol("text").
    setOutputCol("document")

val regexTokenizer = new Tokenizer().
    setInputCols(Array("sentence")).
    setOutputCol("token")
val sentenceDetector = new SentenceDetector().
    setInputCols(Array("document")).
    setOutputCol("sentence")

val finisher = new Finisher()
    .setInputCols("token")
    .setIncludeMetadata(true)

val pipeline = new Pipeline().
    setStages(Array(
        documentAssembler,
        sentenceDetector,
        regexTokenizer,
        finisher
    ))

Create a simple DataFrame for testing:

val data1 = Seq("hello, this is an example sentence").toDF("text")

Now we fit and transform your DataFrame on this Pipeline:

val prediction = pipeline.fit(data1).transform(data1)

The variable prediction is a DataFrame which in that you can explode the token column. Let's have a look inside prediction DataFrame:

scala> prediction.show
+--------------------+--------------------+-----------------------+
|                text|      finished_token|finished_token_metadata|
+--------------------+--------------------+-----------------------+
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|
+--------------------+--------------------+-----------------------+

scala> prediction.withColumn("newCol", explode($"finished_token")).show
+--------------------+--------------------+-----------------------+--------+
|                text|      finished_token|finished_token_metadata|  newCol|
+--------------------+--------------------+-----------------------+--------+
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|   hello|
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|       ,|
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|    this|
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|      is|
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|      an|
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...| example|
|hello, this is an...|[hello, ,, this, ...|   [[sentence, 0], [...|sentence|
+--------------------+--------------------+-----------------------+--------+

• Your first issue as Alberto mentioned, thinking that finisher was a DataFrame. It is an annotator until it is transformed.

• The Second issue was having .toDF() in a place you didn't need it. (after pipeline transformation)

• Your explode function being in a bad place aside, you are zipping a column that doesn't even exist in your pipeline: ner

Please feel free to ask any question and I'll update the answer accordingly.