Jad2192 update xgb (#4)

* update xgb to 1.5.2 in testing

* add `allowNonZeroForMissing`

* update version

README.md

@@ -6,6 +6,6 @@ Spark users can use XGBoost for classification and regression tasks in a distrib
 
 This repository contains the Python wrapper components from that PR. By building and installing the appropriate .whl (see 'releases' in this repository), PySpark users can directly use this wrapper with the XGBoost library.
 
-Tested with  XGBoost 1.3.1
+Tested with  XGBoost 1.5.2
 
 See the notebook in `/examples` for a practical illustration of usage.

setup.py

@@ -24,7 +24,7 @@
 
 setup(
     name='spark-xgboost',
-    version='0.90.2',
+    version='1.0.0',
     description='spark-xgboost is the PySpark package for XGBoost',
 
     long_description=long_description,

sparkxgb/__init__.py

@@ -17,4 +17,4 @@
 from sparkxgb.xgboost import XGBoostClassifier, XGBoostRegressor, XGBoostClassificationModel, XGBoostRegressionModel
 
 __all__ = ["XGBoostClassifier", "XGBoostRegressor", "XGBoostClassificationModel", "XGBoostRegressionModel"]
-__version__ = "0.90"
+__version__ = "1.0.0"

sparkxgb/testing/utils.py

@@ -10,7 +10,7 @@ def default_session(conf=None):
     spark_major_version = pyspark.__version__[0]
     scala_version = SPARK_SCALA_MAPPING[spark_major_version]
     mvn_group = "ml.dmlc"
-    xgb_version = "1.3.1"
+    xgb_version = "1.5.2"
     xgboost4j_coords = f"{mvn_group}:xgboost4j_{scala_version}:{xgb_version}"
     xgboost4j_spark_coords = f"{mvn_group}:xgboost4j-spark_{scala_version}:{xgb_version}"
 

sparkxgb/tests/classifier_test.py

@@ -58,7 +58,8 @@ def test_binary_classifier_args(self):
             objective="binary:logistic",
             numRound=5,
             numWorkers=2,
-            killSparkContextOnWorkerFailure=False
+            killSparkContextOnWorkerFailure=False,
+            allowNonZeroForMissing=True,
         )
 
         xgb = (

sparkxgb/xgboost.py

@@ -25,6 +25,7 @@ class XGBoostClassifier(XGboostEstimator):
 
     @keyword_only
     def __init__(self,
+                 allowNonZeroForMissing=False,
                  alpha=0.0,
                  baseMarginCol=None,
                  baseScore=0.5,
@@ -91,6 +92,7 @@ def __init__(self,
 
     @keyword_only
     def setParams(self,
+                  allowNonZeroForMissing=False,
                   alpha=0.0,
                   baseMarginCol=None,
                   baseScore=0.5,
@@ -180,6 +182,7 @@ class XGBoostRegressor(XGboostEstimator):
 
     @keyword_only
     def __init__(self,
+                 allowNonZeroForMissing=False,
                  alpha=0.0,
                  baseMarginCol=None,
                  baseScore=0.5,
@@ -247,6 +250,7 @@ def __init__(self,
 
     @keyword_only
     def setParams(self,
+                  allowNonZeroForMissing=False,
                   alpha=0.0,
                   baseMarginCol=None,
                   baseScore=0.5,