def test(f_discretization=10, e_discretization=30, indep=False, selection=None, subset=None, corpus=None, smoothing=None, sensitivity=5, segmentation='reasonable'):
if not selection:
selection = (db.select())
score = db.getScore1(selection)
if not corpus:
(f, e, m) = tools.chooseFeatures()
else:
(f, e, m) = tools.loadFeatures(corpus)
if m['version'] != sf.version:
print("Scorefeatures versions don't match! Corpus version: {0} Scorefeatures version: {1}".format(m['version'], sf.version))
exit(0)
if not subset:
# Select a subset by hand
subset = selectSubset(m['featureset'])
print('\n\tPerforming {0}'.format(selection))
print('\tFeatures version: {0}'.format(m['version']))
print('\tFeatureset used: [', end=' ')
print(', '.join([m['featureset'][i] for i in subset]), end=' ')
print(']')
print('\tScorefeatures discretization: {0}\n\tExpression discretization: {1}'.format(f_discretization, e_discretization))
print('\tSensitivity: {0}'.format(sensitivity))
print('\tSmoothing: {0}'.format(smoothing))
print('\tCorpus: {0}\n'.format(corpus))
if indep:
hmm = HMM_indep(2, smoothing)
else:
hmm = HMM(2, smoothing)
trainHMM(hmm, f, e, f_discretization, e_discretization, subset=subset, ignore=selection, sensitivity=sensitivity)
#trainHMM(hmm, f, e, f_discretization, e_discretization, subset=subset)
hmm.storeInfo('hmm2.txt')
print("Loading score")
melodyscore = Score(score).melody()
melody = tools.parseScore(melodyscore)
# Segmentate the the score
print("Analysing score")
if segmentation == 'reasonable':
print('Using reasonable segmentation')
onset = structure.reasonableSegmentation(melody)
elif segmentation == 'new':
print('Using new segmentation')
onset = structure.newSegmentation(melody)
elif segmentation == 'notelevel':
print('Using notelevel segmentation')
onset = structure.noteLevel(melody)
#onset = structure.groupings(structure.list_to_tree(structure.first_order_tree(structure.onset, melody, 0.1)), 1)
#namelist = []
#for group in onset:
# namelist.append([leaf.name() for leaf in group])
#print namelist
(p, expression) = render(melodyscore, onset, hmm, f_discretization, e_discretization, subset, sensitivity=sensitivity)
print("Done, resulting expression(with a probability of {1}): {0}".format(expression, p))
performance = perform.perform(score, melodyscore, onset, expression, converter=melody)
playPerformance(selection, performance, expression, onset, '{0}_{1}_on_{2}_discret_{3}-{4}_smoothing_{5}'.format(\
selection[0], selection[1], corpus, f_discretization, e_discretization, smoothing))
def train(trainset):
expression = {}
features = {}
const = 0
Max = 0
Min = None
count = 0
print(">>> Loading scores and deviations, this will take hours and may eat all you memory")
for query in trainset:
print(">>> Loading: {0}".format(query))
score = db.getScore1(query)
deviations = db.getDeviation1(query)
alignment = Alignment(score, deviations)
melody = alignment.melody()
#segments = structure.newSegmentation(tools.parseScore(melody))
segments = structure.noteLevel(tools.parseScore(melody))
const += len(segments)
lengths = sum([len(s) for s in segments])
m = max([len(s) for s in segments])
mi = min([len(s) for s in segments])
if m > Max:
Max = m
if not Min:
Min = mi
if mi < Min:
Min = mi
print('>>> Extracting features')
expression[query] = performancefeatures.vanDerWeijExpression(alignment, segments)
features[query] = scorefeatures.vanDerWeijFeatures(melody, segments)
count += 1
print('{0}/{1} done'.format(count, len(trainset)))
print("Done, {0} segments found with an average length of: {1} (min: {2} max: {3})".format(const, lengths / float(const), Min, Max))
tools.saveFeatures(features, expression)