def em_predict() :
""" used to assigned label of cluster the Image
coming from client side and return the result
object contain image feature, responsibility
and assigned cluster
Request Method :
POST
Body :
type = 'application/json'
value (str) : base64 encoded value of image
filename (str) : name of Image file
filetype (str) : Image extension
k (int) : number of clusters (explicitly type cast to int)
Return :
Json Predicted responsibilities for all clusters along with the file details
"""
try :
req = request.get_data().decode('utf-8')
img = json.loads(req)
img_base64 = img['value']
img_name = img['filename']
filetype = img['filetype']
k = int(img['k'])
path = os.path.join(CONSTANT.TEMP_FILE_PATH, img_name)
path = spr.decode_base64(img_base64, path)
if path is not None :
emobj = get_model(k)
result = emobj.predict_data(img_name, filetype, path, img_base64)
return result.to_json(orient='records')
except Exception as e:
LOGGER.LOG(e) # log exception
return {}
def getfirstndataresponsibility():
"""
return all the image with parameters like
responsibility image parameters assigned clusters
Method :
GET
Arguments :
k (int) : number of clusters
n (int) : number of responsibilities
Return :
JSON With cluster name to top n requested training data responsibility with Details
"""
try:
k = request.args.get('k', type = int)
n = request.args.get('n', type = int)
emobj = get_model(k)
temp = emobj.get_first_n_data_responsibility(n, to_json=True)
for i in temp.keys() :
temp[i] = ast.literal_eval(temp[i]) # convert string list to list
return json.loads(json.dumps(temp))
except Exception as e:
LOGGER.LOG(e) # Log Exception
return {}
def changek():
"""
change value of k and retrain the model
Method :
POST
Arguments :
k (int) : number of clusters
Return :
JSON {'res': (bool)status if successfully create model for changed k or not}
"""
try:
k = request.args.get('k')
if k in dic_model :
del dic_model[k]
emObj = get_model(k)
if emObj is None :
return jsonify({'res':False})
else :
return jsonify({'res':True})
except Exception as e:
LOGGER.LOG(e) # Log Exception
return {}
def getclusterparameter():
""" get all cluster trained parameter like weights
means covarience metrix responsibilities for
each cluster
Method :
GET
Arguments :
k (int) : number of cluster (explicitly casted to int)
Return :
JSON contain {weights: weight of cluster, means: mean points of cluster, covariances: covariance metrix}
"""
try:
k = request.args.get('k', type = int)
emobj = get_model(k)
param = emobj.em_parameters
param['weights'] = param['weights'].tolist()
param['means'] = param['means'].tolist()
param['covariances'] = param['covariances'].tolist()
del param['responsibility']
return json.loads(json.dumps(param))
except Exception as e:
LOGGER.LOG(e) # Log Exception
return {}
def setclustername():
"""
set cluster number to name mapping
Method :
POST
Body :
type = 'application/json'
k (int) : cluster number (explicitly casted)
mappedey (dictionary) : cluster number to name mapping
Return :
JSON {'res' : (bool)saved status}
"""
try:
req = request.get_data().decode('utf-8')
data = json.loads(req)
k = data['k']
dic = data['mappedKey']
emobj = get_model(k)
emobj.set_cluster_name(dic)
return emobj.set_cluster_name(dic)
except Exception as e:
LOGGER.LOG(e) # Log Exception
return {}
def getclustername() :
"""
get cluster number to name mapping
Method :
GET
Arguments :
k (int) : number of clusters
Return :
cluster number to name mapping json
"""
try :
k = request.args.get('k', type = int)
emobj = get_model(k)
return emobj.cluster_name
except Exception as e:
LOGGER.LOG(e) # Log Exception
return {}
def getsupportedimagesextension() :
"""
Used to get all the image file extension
supported for image processing in EM algorithms
Method :
GET
Arguments :
k (int) : number of cluster
Return :
Array of string having supported extensions
"""
try :
k = request.args.get('k', type = int)
emobj = get_model(k)
return jsonify(emobj.IMAGE_EXT_SUPPORTED)
except Exception as e:
LOGGER.LOG(e) # Log Exception
return {}
def getfirstnheterogeneity():
"""
return value of k and its assiciate
heterogeneity to obtain optimum value of k
Method :
GET
Arguments :
k (int) : number of clusters
n (int) : number of First n heterogeneity
Return :
JSON of number of cluster to heterogeneity mapped pair
"""
try:
k = request.args.get('k', type = int)
n = request.args.get('n', type = int)
emobj = get_model(k)
return emobj.get_first_n_heterogeneity(n, seed=CONSTANT.SEED)
except Exception as e:
LOGGER.LOG(e)
return {}