From 5fe3c58b4a1497f1d05459ba823e1db12d8c6e93 Mon Sep 17 00:00:00 2001
From: tqchen <tianqi.tchen@gmail.com>
Date: Mon, 22 Dec 2014 00:31:01 -0800
Subject: [PATCH] add kmeans hadoop

---
 toolkit/kmeans_hadoop.sh | 9 +++++++++
 tracker/rabit_hadoop.py  | 3 ++-
 2 files changed, 11 insertions(+), 1 deletion(-)
 create mode 100755 toolkit/kmeans_hadoop.sh

diff --git a/toolkit/kmeans_hadoop.sh b/toolkit/kmeans_hadoop.sh
new file mode 100755
index 000000000..ed576f8b9
--- /dev/null
+++ b/toolkit/kmeans_hadoop.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+if [ "$#" -lt 5 ];
+then
+    echo "Usage: <nslaves> <input_data> <ncluster> <max_iteration> <output>"
+    exit -1
+fi
+python ../rabit_hadoop.py -s $1 -i $2  -m kmeans --args "stdin "$3" "$4" stdout" -o $5
+
+ 
\ No newline at end of file
diff --git a/tracker/rabit_hadoop.py b/tracker/rabit_hadoop.py
index 79e0cb133..c81e9d344 100755
--- a/tracker/rabit_hadoop.py
+++ b/tracker/rabit_hadoop.py
@@ -35,7 +35,8 @@ if hadoop_streaming_jar != None:
   args.hadoop_streaming_jar = hadoop_streaming_jar
 
 def hadoop_streaming(nslaves, slave_args):
-  cmd = '%s jar %s -input %s -output %s -mapper \"%s stdin %d %d stdout %s\" -reducer \"/bin/cat\" -file %s -D mapred.map.tasks=%d' % (args.hadoop_binary, args.hadoop_streaming_jar, args.input, args.output, args.mapper, args.nclusters, args.iterations, ' '.join(slave_args), args.mapper, nslaves)
+  cmd = '%s jar %s -input %s -output %s -mapper \"%s %s %s\" -reducer \"/bin/cat\" -file %s -D mapred.map.tasks=%d' 
+  % (args.hadoop_binary, args.hadoop_streaming_jar, args.input, args.output, args.mapper, args.args, ' '.join(slave_args), args.mapper, nslaves)
   print cmd
   subprocess.check_call(cmd, shell = True)