安装SparkNLP后，无法导入SparkNLP

Question

以下在Cloudera CDSW群集网关上成功运行。

import pyspark
from pyspark.sql import SparkSession
spark = (SparkSession
            .builder
            .config("spark.jars.packages","JohnSnowLabs:spark-nlp:1.2.3")
            .getOrCreate()
         )

产生此输出。

Ivy Default Cache set to: /home/cdsw/.ivy2/cache
The jars for the packages stored in: /home/cdsw/.ivy2/jars
:: loading settings :: url = jar:file:/opt/cloudera/parcels/SPARK2-2.2.0.cloudera1-1.cdh5.12.0.p0.142354/lib/spark2/jars/ivy-2.4.0.jar!/org/apache/ivy/core/settings/ivysettings.xml
JohnSnowLabs#spark-nlp added as a dependency
:: resolving dependencies :: org.apache.spark#spark-submit-parent;1.0
    confs: [default]
    found JohnSnowLabs#spark-nlp;1.2.3 in spark-packages
    found com.typesafe#config;1.3.0 in central
    found org.fusesource.leveldbjni#leveldbjni-all;1.8 in central
downloading http://dl.bintray.com/spark-packages/maven/JohnSnowLabs/spark-nlp/1.2.3/spark-nlp-1.2.3.jar ...
    [SUCCESSFUL ] JohnSnowLabs#spark-nlp;1.2.3!spark-nlp.jar (3357ms)
downloading https://repo1.maven.org/maven2/com/typesafe/config/1.3.0/config-1.3.0.jar ...
    [SUCCESSFUL ] com.typesafe#config;1.3.0!config.jar(bundle) (348ms)
downloading https://repo1.maven.org/maven2/org/fusesource/leveldbjni/leveldbjni-all/1.8/leveldbjni-all-1.8.jar ...
    [SUCCESSFUL ] org.fusesource.leveldbjni#leveldbjni-all;1.8!leveldbjni-all.jar(bundle) (382ms)
:: resolution report :: resolve 3836ms :: artifacts dl 4095ms
    :: modules in use:
    JohnSnowLabs#spark-nlp;1.2.3 from spark-packages in [default]
    com.typesafe#config;1.3.0 from central in [default]
    org.fusesource.leveldbjni#leveldbjni-all;1.8 from central in [default]
    ---------------------------------------------------------------------
    |                  |            modules            ||   artifacts   |
    |       conf       | number| search|dwnlded|evicted|| number|dwnlded|
    ---------------------------------------------------------------------
    |      default     |   3   |   3   |   3   |   0   ||   3   |   3   |
    ---------------------------------------------------------------------
:: retrieving :: org.apache.spark#spark-submit-parent
    confs: [default]
    3 artifacts copied, 0 already retrieved (5740kB/37ms)
Setting default log level to "ERROR".
To adjust logging level use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).

但是，当我尝试导入John Snow Labs上的Pyspark上所述的SparkNLP ...

import sparknlp
# or 
from sparknlp.annotator import *

我明白了：

ImportError: No module named sparknlp
ImportError: No module named sparknlp.annotator 

使用SparkNLP我需要做什么？当然，这可以对任何火花包进行推广。

Answer 1

您可以使用命令中的PySpark中使用SparkNLP软件包：

pyspark --packages JohnSnowLabs:spark-nlp:1.3.0

，但这并不能告诉Python在哪里可以找到绑定。按照此处的类似报告的说明，可以通过将JAR目录添加到您的pythonpath：

来修复：

export PYTHONPATH="~/.ivy2/jars/JohnSnowLabs_spark-nlp-1.3.0.jar:$PYTHONPATH"

或

import sys, glob, os
sys.path.extend(glob.glob(os.path.join(os.path.expanduser("~"), ".ivy2/jars/*.jar")))

Answer 2

我弄清楚了。正确加载的JAR文件仅是编译的Scala文件。我仍然必须将包含包装代码的Python文件放在我可以从中导入的位置。一旦我这样做，一切都很好。

Answer 3

感谢粘土。以下是我设置pythonpath的方式：

git clone --branch 3.0.3 https://github.com/JohnSnowLabs/spark-nlp
export PYTHONPATH="./spark-nlp/python:$PYTHONPATH"

然后它对我有用，因为我的./spark-nlp/python文件夹现在包含难以捉摸的SparkNLP模块。

pyspark --packages com.johnsnowlabs.nlp:spark-nlp_2.12:3.0.3
>>> import sparknlp
>>>