def main():
#first run the initializer to get starting centroids
filePath = '/home/mike-bowles/pyWorkspace/mapReducers/src/kMeans3/input.txt'
#mrJob = MRkMeansInit(args=[filePath])
#mrJob = MRkMeansInit(args=['-r', 'emr', filePath])
#with mrJob.make_runner() as runner:
# runner.run()
#pull out the centroid values to compare with values after one iteration
centPath = "s3://mike-mrjob/kMeans/centroids/intermediateResults.txt"
key = EMRJobRunner().get_s3_key(centPath)
centroidsJson = key.get_contents_as_string()
delta = 10
#Begin iteration on change in centroids
while delta > 0.01:
#parse old centroid values
oldCentroids = json.loads(centroidsJson)
#run one iteration
mrJob2 = MRkMeansIter(args=['-r', 'emr', filePath])
with mrJob2.make_runner() as runner:
runner.run()
#compare new centroids to old ones
centroidsJson = key.get_contents_as_string()
newCentroids = json.loads(centroidsJson)
kMeans = len(newCentroids)
delta = 0.0
for i in range(kMeans):
delta += dist(newCentroids[i],oldCentroids[i])
print delta
