def load_data():
try:
file_name = dict(request.form)["builtin"][0]
except Exception as e:
file_name = None
if file_name:
try:
data, names = convert_csv("data/{}".format(file_name))
except IOError:
return jsonify({"error": "No built in data file '{}'.".format(file_name)})
else:
csv_file = dict(request.files)["file"][0]
file_name = csv_file.filename
file_name = "{}-{}".format(str(uuid.uuid4()), file_name)
# save the csv to data dir
csv_file.save("data/{}".format(file_name))
# then just convert it as usual
data, names = convert_csv("data/{}".format(file_name))
# update the dataset session var
session["dataset"] = file_name
# return data
return jsonify({"variable_names": data.columns.values.tolist()})
def cluster():
try:
method_name = request.args.get("method")
method = settings.CLUSTER_METHODS[method_name]
except KeyError:
methods = ",".join(settings.CLUSTER_METHODS.keys())
error = "parameter 'method' must be one of the following: {}".format(methods)
return jsonify({"error": error})
try:
num_clusters = int(request.args.get("clusters"))
if num_clusters < 1:
raise TypeError
except (TypeError, ValueError):
error = "parameter 'clusters' must be an integer number of clusters > 0"
return jsonify({"error": error})
return_X_data = bool(request.args.get("data"))
# Get dataset - either from session or from DEFAULT_DATASET location
dataset = session.get("dataset", settings.DEFAULT_DATASET)
# Convert csv to dataframe
X, names = convert_csv("data/{}".format(dataset))
cluster_vals = method(X, num_clusters)
# TODO: check numpy array and cast to list if needed
data = {"clusters": {num_clusters: cluster_vals.tolist()}, "names": names}
if return_X_data:
data["variables"] = {var: X[var].tolist() for var in X}
return jsonify(data)
