$ chmod +x <anaconda_filename>
$ ./<anaconda_filename>
$ conda update conda
$ conda update ipython
%%bash
curl -O http://www.cs.mcgill.ca/~jfrank8/data/gait-dataset.tar.gz
tar xzf gait-dataset.tar.gz

import statsmodels.api as sm
import pandas as pd
import os

dta_path = 'gait-dataset'

files_byday = lambda dta_path, day: [os.path.join(dta_path, name) 
                                      for name in os.listdir(dta_path) 
                                          if name.endswith('.csv') and 'day'+str(day) in name]
day1 = files_byday(dta_path, 1)

dta = pd.read_csv(day1[0], header=0, sep="\t", usecols=['timestamp', 'accel_mag'])
dta.index = dta['timestamp'].astype('M8[ms]')
del dta['timestamp']

dta.plot(figsize=(12,8));

def arma_featurizer(filename, nwin=500, nparam=2):
    import numpy as np
    import pandas as pd
    import statsmodels.api as sm

    dta = pd.read_csv(filename, header=0, sep="\t", usecols=['timestamp', 'accel_mag'])
    dta.index = pd.to_datetime(dta['timestamp'], unit='ms')
    del dta['timestamp']

    nwin, nparam = 10, nparam
    X = np.zeros((nwin, nparam+1))
    for i, subseries in enumerate(np.array_split(dta, nwin)):
        model = sm.tsa.ARMA(subseries, (nparam,0)).fit()
        X[i] = model.params
        
    return X

!ipcluster start -n=2 --daemon

from IPython import parallel
rc = parallel.Client()
rc.block = True
all_engines = rc[:]

Y = all_engines.map(arma_featurizer, day1[:2])

def fit_svm(fold):
    import numpy as np
    from sklearn import svm
    from sklearn.metrics import accuracy_score

    X = np.load('instances.npy', mmap_mode='r')
    y = np.load('labels.npy', mmap_mode='r')
    train, test = fold

    clf = svm.SVC()
    clf.fit(X[train], y[train])
    ypred = clf.predict(X[test])
    return accuracy_score(y[test], ypred)

from sklearn import cross_validation

np.save('instances', np.vstack(Y))
labels = np.repeat(range(len(Y)), len(Y[0]))
np.save('labels', labels)

accuracy_scores = all_engines.map(fit_svm, cross_validation.StratifiedKFold(labels, n_folds=2))

print accuracy_scores

!ipcluster stop
$ pip install starcluster$ starcluster help
StarCluster - (http://web.mit.edu/starcluster)
Software Tools for Academics and Researchers (STAR)
Please submit bug reports to starcluster@mit.edu

cli.py:87 - ERROR - config file /home/user/.starcluster/config does not exist

Options:
--------
[1] Show the StarCluster config template
[2] Write config template to /home/user/.starcluster/config
[q] Quit

Please enter your selection:[plugin ipcluster]
setup_class = starcluster.plugins.ipcluster.IPClusterPLUGINS = ipclusterNODE_INSTANCE_TYPE = m1.small[key rllab]
KEY_LOCATION=~/.ssh/rllab.rsa
KEYNAME = rllab$ starcluster createkey mykey -o ~/.ssh/rllab.rsa$ starcluster spothistory m1.small
StarCluster - (http://star.mit.edu/cluster) (v. 0.95.2)
Software Tools for Academics and Researchers (STAR)
Please submit bug reports to starcluster@mit.edu

>>> Fetching spot history for m1.small (VPC)
>>> Current price: $0.0071
>>> Max price: $0.5000
>>> Average price: $0.0459$ starcluster start -s 2 -b 0.05 rllab$ starcluster sshmaster rllab -u sgeadmin
rc = parallel.Client('/Users/pierrelucbacon/.starcluster/ipcluster/SecurityGroup:@sc-rllab-us-east-1.json', sshkey='/Users/pierrelucbacon/.ssh/rllab.rsa') 
rc.block = True
all_engines = rc[:]

%%px
!ifconfig eth0 | grep 'inet addr'

%%px --targets=10
%%bash
curl -O http://www.cs.mcgill.ca/~jfrank8/data/gait-dataset.tar.gz
tar xzf gait-dataset.tar.gz

Y = all_engines.map(arma_featurizer, day1)

np.save('instances', np.vstack(Y))
labels = np.repeat(range(len(Y)), len(Y[0]))
np.save('labels', labels)

%%bash
starcluster put rllab -u sgeadmin instances.npy /home/sgeadmin
starcluster put rllab -u sgeadmin labels.npy /home/sgeadmin

accuracy_scores = all_engines.map(fit_svm, cross_validation.StratifiedKFold(labels, n_folds=4))

print accuracy_scores