John Snow Labs Spark-NLP is a natural language processing library built on top of Apache Spark ML. It provides simple, performant & accurate NLP annotations for machine learning pipelines, that scale easily in a distributed environment.
Take a look at our official spark-nlp page: http://nlp.johnsnowlabs.com/ for user documentation and examples
Questions? Feedback? Request access sending an email to nlp@johnsnowlabs.com
Spark-NLP 1.8.2 has been built on top of Apache Spark 2.4.0
Note that Spark is not retrocompatible with Spark 2.3.x, so models and environments might not work.
If you are still stuck on Spark 2.3.x feel free to use this assembly jar instead. Support is limited. For OCR module, this is for spark 2.3.x
Spark NLP | Spark 2.0.0 / Spark 2.3.x | Spark 2.4 |
---|---|---|
1.8.x | Partially | YES |
1.7.3 | YES | N/A |
1.6.3 | YES | N/A |
1.5.0 | YES | N/A |
Find out more about Spark-NLP
versions from our release notes.
This library has been uploaded to the spark-packages repository.
Benefit of spark-packages is that makes it available for both Scala-Java and Python
To use the most recent version just add the --packages JohnSnowLabs:spark-nlp:1.8.2
to you spark command
spark-shell --packages JohnSnowLabs:spark-nlp:1.8.2
pyspark --packages JohnSnowLabs:spark-nlp:1.8.2
spark-submit --packages JohnSnowLabs:spark-nlp:1.8.2
This can also be used to create a SparkSession manually by using the spark.jars.packages
option in both Python and Scala
Either download pre-compiled packages here or build from source using sbt assembly
Spark-NLP FAT-JAR from here (Does NOT include Spark): Spark-NLP 1.8.2 FAT-JAR
Spark-NLP GPU Enhanced Tensorflow FAT-JAR: Spark-NLP 1.8.2-gpu FAT-JAR
Spark-NLP-OCR Module (Requires native Tesseract 4.x+ for image based OCR. Does not require Spark-NLP to work but highly suggested) Spark-NLP-OCR 1.8.2 FAT-JAR
If for some reason you need to use the JAR, you can either download the Fat JARs provided here or download it from Maven Central.
To add JARs to spark programs use the --jars
option:
spark-shell --jars spark-nlp.jar
The preferred way to use the library when running spark programs is using the --packages
option as specified in the spark-packages
section.
Our package is deployed to maven central. In order to add this package as a dependency in your application:
<!-- https://mvnrepository.com/artifact/com.johnsnowlabs.nlp/spark-nlp -->
<dependency>
<groupId>com.johnsnowlabs.nlp</groupId>
<artifactId>spark-nlp_2.11</artifactId>
<version>1.8.2</version>
</dependency>
and
<!-- https://mvnrepository.com/artifact/com.johnsnowlabs.nlp/spark-nlp-ocr -->
<dependency>
<groupId>com.johnsnowlabs.nlp</groupId>
<artifactId>spark-nlp-ocr_2.11</artifactId>
<version>1.8.2</version>
</dependency>
// https://mvnrepository.com/artifact/com.johnsnowlabs.nlp/spark-nlp
libraryDependencies += "com.johnsnowlabs.nlp" %% "spark-nlp" % "1.8.2"
and
// https://mvnrepository.com/artifact/com.johnsnowlabs.nlp/spark-nlp-ocr
libraryDependencies += "com.johnsnowlabs.nlp" %% "spark-nlp-ocr" % "1.8.2"
Maven Central: https://mvnrepository.com/artifact/com.johnsnowlabs.nlp
If you installed pyspark through pip, you can install spark-nlp
through pip as well.
pip install spark-nlp==1.8.2
PyPI spark-nlp package
If you are using Anaconda/Conda for managing Python packages, you can install spark-nlp
as follow:
conda install -c johnsnowlabs spark-nlp
Anaconda spark-nlp package
Then you'll have to create a SparkSession manually, for example:
spark = SparkSession.builder \
.appName("ner")\
.master("local[4]")\
.config("spark.driver.memory","4G")\
.config("spark.driver.maxResultSize", "2G") \
.config("spark.jars.packages", "JohnSnowLabs:spark-nlp:1.8.2")\
.config("spark.kryoserializer.buffer.max", "500m")\
.getOrCreate()
If using local jars, you can use spark.jars
instead for a comma delimited jar files. For cluster setups, of course you'll have to put the jars in a reachable location for all driver and executor nodes
Use either one of the following options
- Add the following Maven Coordinates to the interpreter's library list
com.johnsnowlabs.nlp:spark-nlp_2.11:1.8.2
- Add path to pre-built jar from here in the interpreter's library list making sure the jar is available to driver path
Apart from previous step, install python module through pip
pip install spark-nlp==1.8.2
Or you can install spark-nlp
from inside Zeppelin by using Conda:
%python.conda install -c johnsnowlabs spark-nlp
Configure Zeppelin properly, use cells with %spark.pyspark or any interpreter name you chose.
Finally, in Zeppelin interpreter settings, make sure you set properly zeppelin.python to the python you want to use and installed the pip library with (e.g. python3
).
An alternative option would be to set SPARK_SUBMIT_OPTIONS
(zeppelin-env.sh) and make sure --packages
is there as shown earlier, since it includes both scala and python side installation.
Easiest way to get this done is by making Jupyter Notebook run using pyspark as follows:
export SPARK_HOME=/path/to/your/spark/folder
export PYSPARK_PYTHON=python3
export PYSPARK_DRIVER_PYTHON=jupyter
export PYSPARK_DRIVER_PYTHON_OPTS=notebook
pyspark --packages JohnSnowLabs:spark-nlp:1.8.2
Alternatively, you can mix in using --jars
option for pyspark + pip install spark-nlp
If not using pyspark at all, you'll have to run the instructions pointed here
If your distributed storage is S3 and you don't have a standard hadoop configuration (i.e. fs.defaultFS) You need to specify where in the cluster distributed storage you want to store Spark-NLP's tmp files. First, decide where you want to put your application.conf file
import com.johnsnowlabs.uti.ConfigLoader
ConfigLoader.setConfigPath("/somewhere/to/put/application.conf")
And then we need to put in such application.conf the following content
sparknlp {
settings {
cluster_tmp_dir = "somewhere in s3n:// path to some folder"
}
}
If you have troubles using pretrained() models in your environment, here a list to various models (only valid for latest versions). If there is any older than current version of a model, it means they still work for current versions.
English |
---|
Basic Pipeline |
Advanced Pipeline |
Vivekn Sentiment Pipeline |
Models | English | Italian |
---|---|---|
LemmatizerModel (Lemmatizer) | Download | Download |
PerceptronModel (POS) | Download | N/A |
ViveknSentimentModel (Sentiment) | Download | N/A |
SentimentDetector (Sentiment) | N/A | Download |
NerCRFModel (NER) | Download | N/A |
NerDLModel (NER) | Download | N/A |
SymmetricDeleteModel (Spell Checker) | Download | N/A |
ContextSpellCheckerModel (Spell Checker) | Download | N/A |
NorvigSweetingModel (Spell Checker) | Download | N/A |
After downloading offline models/pipelines and extracting them, here is how you can use them:
- Loading
PerceptronModel
annotator model inside Spark NLP Pipeline
val pos = annotator.PerceptronModel.load("/tmp/pos_fast_en_1.8.0_2.4_1545434653742/")
.setInputCols("document", "normal")
.setOutputCol("pos")
.asInstanceOf[PerceptronModel]
- Loading
Advanced Pipeline
val advancedPipeline = PipelineModel.load("/tmp/pipeline_advanced_en_1.8.0_2.4_1545436028146/")
// To use the loaded Pipeline for prediction
advancedPipeline.transform(predictionDF)
Check our Articles and FAQ page here
-
Q: I am getting a Java Core Dump when running OCR transformation
- A: Add
LC_ALL=C
environment variable
- A: Add
-
Q: Getting
org.apache.pdfbox.filter.MissingImageReaderException: Cannot read JPEG2000 image: Java Advanced Imaging (JAI) Image I/O Tools are not installed
when running an OCR transformation- A:
--packages com.github.jai-imageio:jai-imageio-jpeg2000:1.3.0
. This library is non-free thus we can't include it as a Spark-NLP dependency by default
- A:
Thanks in general to the community who have been lately reporting important issues and pull request with bugfixes. Community has been key in the last releases with feedback in various Spark based environments.
Here a few specific mentions for recurring feedback and slack participation
- @maziyarpanahi (https://github.com/maziyarpanahi) - For contributing with testing and valuable feedback
- @easimadi (https://github.com/easimadi) - For contributing with documentation and valuable feedback
We appreciate any sort of contributions:
- ideas
- feedback
- documentation
- bug reports
- nlp training and testing corpora
- development and testing
Clone the repo and submit your pull-requests! Or directly create issues in this repo.