def main():
# configure logging
logging.basicConfig(level=logging.INFO)
logging.info(cf_netSDM)
# Read inputs
inputs = io_helper.fetch_data()
data = inputs['data']
normalize = get_param(inputs['parameters'], 'normalize', bool, 'True')
damping = get_param(inputs['parameters'], 'damping', float, '0.85')
data_array = np.zeros((len(data['independent'][0]['series']), len(data['independent'])))
col_number = 0
row_number = 0
for var in data['independent']:
for value in var['series']:
data_array[row_number, col_number] = value
row_number += 1
col_number += 1
row_number = 0
if normalize:
for col_number in range(data_array.shape[1]):
data_array[:, col_number] = data_array[:, col_number] / np.linalg.norm(data_array[:, col_number])
network = construct_adjacency_graph(range(data_array.shape[0]), data_array, data['dependent'][0]['series'])
propositionalized = cf_netSDM.hinmine_propositionalize(network, damping)['train_features']['data']
results_dict = {
'profile': 'tabular-data-resource',
'name': 'hinmine-features',
'data': [],
'schema': {
'fields': [],
'primaryKey': 'id'
}
}
n = propositionalized.shape[0]
for row_index in range(n):
instance = {"id": row_index}
for col_index in range(n):
instance["feature_%i" % (col_index + 1)] = propositionalized[row_index, col_index]
results_dict['data'].append(instance)
for col_index in range(n):
results_dict['schema']['fields'].append({'name': 'feature_%i' % (col_index + 1), 'type': 'float'})
io_helper.save_results(json.dumps(results_dict), '', 'text/plain')
def main():
# Configure logging
logging.basicConfig(level=logging.INFO)
# Read inputs
inputs = io_helper.fetch_data()
dep_var = inputs["data"]["dependent"][0]
inped_vars = inputs["data"]["independent"]
# Check dependent variable type (should be continuous)
if dep_var["type"]["name"] not in ["integer", "real"]:
logging.warning("Dependent variable should be continuous !")
return None
# Extract data and parameters from inputs
data = format_data(inputs["data"])
# Compute linear-regression and generate PFA output
linear_regression_results = format_output(compute_linear_regression(dep_var, inped_vars, data))
# Store results
io_helper.save_results(linear_regression_results, '', 'application/highcharts+json')
def main():
# Configure logging
logging.basicConfig(level=logging.INFO)
# Read inputs
inputs = io_helper.fetch_data()
dep_var = inputs["data"]["dependent"][0]
inped_vars = inputs["data"]["independent"]
design = get_parameter(inputs["parameters"], DESIGN_PARAM)
# Check dependent variable type (should be continuous)
if dep_var["type"]["name"] not in ["integer", "real"]:
logging.warning("Dependent variable should be continuous !")
return None
# Extract data and parameters from inputs
data = format_data(inputs["data"])
# Compute anova and generate PFA output
anova_results = format_output(compute_anova(dep_var, inped_vars, data, design).to_dict())
# Store results
io_helper.save_results(anova_results, '', 'application/highcharts+json')
def main():
# Configure logging
logging.basicConfig(level=logging.INFO)
# Read inputs
inputs = io_helper.fetch_data()
try:
dep_var = inputs["data"]["dependent"][0]
except KeyError:
logging.warning("Cannot find dependent variables data")
dep_var = []
try:
indep_vars = inputs["data"]["independent"]
except KeyError:
logging.warning("Cannot find independent variables data")
indep_vars = []
nb_bins = get_bins_param(inputs["parameters"], BINS_PARAM)
# Compute histograms (JSON formatted for HighCharts)
histograms_results = compute_histograms(dep_var, indep_vars, nb_bins)
# Store results
io_helper.save_results(histograms_results, '',
'application/highcharts+json')
import tempfile
import logging
from subprocess import call
from io_helper import io_helper
import preprocess
DEFAULT_DOCKER_IMAGE = 'python-hedwig'
if __name__ == '__main__':
# Configure logging
logging.basicConfig(level=logging.INFO)
# Read inputs
inputs = io_helper.fetch_data()
data = inputs["data"]
out_file = 'input.csv'
rules_out_file = 'rules.txt'
matrix, attributes = preprocess.to_matrix(data)
preprocess.dump_to_csv(matrix, attributes, out_file)
# Call hedwig with sensible defaults
examples_file = out_file
empty_bk = tempfile.mkdtemp()
call([
'python', '-m', 'hedwig', empty_bk, examples_file, '-f', 'csv', '-l',
'-o', rules_out_file, '--nocache'
def main():
logging.basicConfig(level=logging.INFO)
inputs = io_helper.fetch_data()
# Dependent variable for tsne this might be the labels - this is optional
labels = None
dependent = inputs["data"].get("dependent", [])
indep_vars = inputs["data"]["independent"] # For tsne the data dimensions
if not data_types_in_allowed(indep_vars, ["integer", "real"]):
logging.warning("Independent variables should be continuous !")
return None
#
data = format_independent_data(inputs["data"])
df = pd.DataFrame.from_dict(data)
source_dimensions = df.shape[1] # number of columns
num_points = df.shape[0] # number of samples/points
convdf = df.apply(lambda x: pd.to_numeric(x))
# Write the data to a temporary file
f = tempfile.NamedTemporaryFile(delete=False)
input = convdf.values.astype(np.float32)
logging.debug('input {}'.format(input))
# Get the parameters (optional)
perplexity = 30
theta = 0.5
target_dimensions = 2
iterations = 1000
do_zscore = True
dependent_is_label = True
try:
perplexity = get_parameter(inputs['parameters'], 'perplexity', perplexity)
theta = get_parameter(inputs['parameters'], 'theta', theta)
target_dimensions = get_parameter(inputs['parameters'], 'target_dimensions', target_dimensions)
iterations = get_parameter(inputs['parameters'], 'iterations', iterations)
do_zscore_str = get_parameter(inputs['parameters'], 'do_zscore', str(do_zscore))
if do_zscore_str == 'True':
do_zscore = True
elif do_zscore_str == 'False':
do_zscore = False
else:
raise ValueError
dependent_is_label_str = get_parameter(inputs['parameters'], 'dependent_is_label', str(dependent_is_label))
if dependent_is_label_str == 'True':
dependent_is_label = True
elif dependent_is_label_str == 'False':
dependent_is_label = False
else:
raise ValueError
except ValueError as e:
logging.error("Could not convert supplied parameter to value, error: ", e)
raise
except:
logging.error(" Unexpected error:", sys.exec_info()[0])
raise
# Compute results
if do_zscore:
input = scipy.stats.zscore(input)
if len(dependent) > 0 and dependent_is_label:
dep_var = dependent[0]
labels = dep_var["series"]
inputFilePath = f.name
input.tofile(inputFilePath)
f.close()
f = tempfile.NamedTemporaryFile(delete=False)
outputFilePath = f.name
f.close()
output = a_tsne(inputFilePath, outputFilePath, num_points,
source_dimensions, target_dimensions, perplexity,
theta, iterations)
logging.debug('output shape {}'.format(output.shape))
logging.debug('output {}'.format(output))
chart = generate_scatterchart(output, indep_vars, labels, perplexity, theta, iterations)
error = ''
shape = 'application/highcharts+json'
logging.debug("Highchart: %s", chart)
io_helper.save_results(chart, error, shape)
logging.info("Highchart output saved to database.")