def classifyForShot(block, context, classifiers, scaler):
"""
Constructs a feature-line from the block and the history in context. There needs to
be a correctly constructed context for this function to work. The feature-line is
scaled using the scaler. Then the distribution for all classifiers is compted and
a single mixed distribution is returned.
"""
feature_line = getFeatureLine(context, block, True, -1)
feature_line.pop()
feature_line = scaler.transform(feature_line)
return distributionOfClassification(feature_line, classifiers,
dist=[0, 0, 0, 0, 0, 0, 0])
def classifyForCut(block, context, classifiers, scaler):
"""
Constructs a feature-line from the block and the history in context. There needs
to be a correctly constructed context for this function to work. The feature-line
is scaled using the scaler. Then the distribution for all classifiers is compted
and a single mixed distribution is returned. This function is for computing the
decision if there should be a cut or not.
"""
if len(context["BygoneBlocks"]) >= 1:
last_shot_id = context["BygoneBlocks"][-1][-1].shotId
else: last_shot_id = 0
feature_line = getFeatureLine(context, block, True, last_shot_id)
feature_line.pop()
feature_line = scaler.transform(feature_line)
return distributionOfClassification(feature_line, classifiers, dist=[0, 0])
def onlineFeatureLineCreator(filename, use_classified_shot = False, use_history = True):
# load context and complete beatlist from file
context, beatList = getContextAndBeatListFromFile(filename)
Features.initializeContextVars(context)
blockList = coalesceBeats(beatList)
context["BygoneBlocks"] = []
for block in blockList:
shot_true = block[-1].shot
# get current feature line and true shot class
featureLine = getFeatureLine(context, block, True, -1)
features = np.array(featureLine[:-1], dtype=np.float64)
shot_classified = yield features, shot_true
# update block and lastShotId
if use_classified_shot:
for beat in block:
beat.shot = shot_classified
if use_history:
context["BygoneBlocks"].append(block)
