def getTranscribedPatterns(ptr = ['TA', 'TA', 'KI', 'TA']):
"""
Gets the transcribed sequences for the pattern (ptr) provided bsed on the onset information in the transcribed data.
"""
result = []
l = len(ptr)
# Getting the masterdata
config = ut.loadConfig('/home/swapnil/SMC/MasterThesis/gitPercPatterns/code/rlcs/config')
transFolder = config['transFolder']
#transFolder = '/home/swapnil/Desktop/temp'
lblDir = config['lblDir']
onsDir = config['onsDir']
masterData = ut.getAllSylbData(tPath = transFolder, lblDir = lblDir, onsDir = onsDir)
# Find the start and the end point of patters in
for comp in masterData:
compName = comp[2]
print 'Working for composition:' + compName
transScore = comp[0][0]
transOnset = comp[0][1]
origScore = comp[1][0]
origOnset = comp[1][1]
# Get the starting indices for the pattern in the composition comp
ptrStartIndices = ut.getSubstringPos(origScore, ptr)
# Get the dictionaries of set for the indices of patterns in the ground truth
dictPtrIndices = populateDictPtrIndices(ptrStartIndices, l)
# Get the closest set onsets for the pattern in the transcription
ptrIndicesInTrans = getIndicesInTrans(origOnset, transOnset, dictPtrIndices)
ptrnsInTrans = getPatternsInTrans(transScore, ptrIndicesInTrans)
result.append((compName, ptrnsInTrans))
return result
baseline = True
#similarityListFile = os.path.join(sylbSimFolder,'simMatList.txt')
#similarityList = [line.strip().split('.')[0] for line in open(similarityListFile)]
# For correctness check
#similarityList = ['TablaDB_3_kmeans_mahal'] # looks promising ....
#similarityList = ['KLMonteCarlo-5','KLMonteCarlo-6', 'KLMonteCarlo-7'] # looks promising ....
#similarityList = ['KLGaussApprox-3', 'KLGaussApprox-4', 'KLGaussApprox-5', 'KLGaussApprox-6', 'KLGaussApprox-7'] # looks promising ....
similarityList = ['binaryDistance'] # results without similarity
ignrSimList = ['TablaDB_10_GMM_euclidean']
simObject = None
masterData = ut.getAllSylbData(tPath = transFolder, lblDir = lblDir, onsDir = onsDir)
#simDict = ut.getSimilarityDict('/home/swapnil/SMC/MasterThesis/sylbSimilarity/TablaDBstrokes','/home/swapnil/SMC/MasterThesis/sylbSimilarity/results_mat/TablaDB_6_kmeans_euclidean.mat')
def getAccuracies(payload, tres = 70.0, fp = None):
#fo.write('run.getAccuracies::')
totalRelvRetrieved = 0
totalRetrieved = 0
totalRelevant = 0
totalRelvRetrievedInGt = 0
ptrInTransLen = [] # List for the length of all the patterns that are the candidate patterns
for inst in payload:
retrieved = inst[0]