def main():
#recordAudio()
recordInfo()
fullData = dict()
for folder in os.listdir(
os.path.join(os.path.dirname(__file__), 'Pictures')):
fullData[folder] = (detectVideoEmotion(os.path.join(
'Pictures', folder)))
with open('appFaceData.json', 'w') as fp:
fp.write(json.dumps(fullData))
print(fullData)
scores = dict()
for folder in os.listdir(os.path.join(os.path.dirname(__file__), 'Audio')):
strList = (detectSpeech(
os.path.join(os.path.dirname(__file__), 'Audio', folder)))
sentiments = []
for item in strList:
#print(item)
sentiments.append(detectSpeechEmotion(item))
sum = 0
for item in sentiments:
sum = sum + item
sum = sum / len(sentiments)
goodpart, badpart = os.path.splitext(folder)
#print(goodpart)
scores[(goodpart)] = sum
#print(sum)
sortedSpeechData = dict()
for k in sorted(scores.keys()):
sortedSpeechData[k] = scores[k]
with open("appSpeechData.json", "w") as fp:
fp.write(json.dumps(sortedSpeechData))
#print(sortedSpeechData)
graphs.makeGraph()
def cosineTranscripts(dataset, queryTranscript, queryName, reference, k):
matrix, query = cosineComparison.calculateComparison(dataset, queryTranscript, k)
scores = cosineComparison.scoreDocuments(query, matrix, reference)
xAxis = [pair[0] for pair in scores]
values = [pair[1] for pair in scores]
graphs.makeGraph(xAxis, "scores", "Transcript Cosine. Reference = {0} K = {1}".\
format(queryName,k), values)
def graphByMonthGraph(dicForm, title, context):
monthScore = np.ones(
[12]).tolist() # value is on dissimilarity so start at one
months = generateMonths()
for date in dicForm:
month = date.month - 1
monthScore[month] -= dicForm[date] # subtract dissimilarlity
graphs.makeGraph(months, "accumulative scores", title, monthScore, context)
def KlPmTranscripts(dataset, queryTranscript, queryName, reference):
#print dataset
matrix, query = cosineComparison.matrixQueryNoTransformation(dataset, queryTranscript)
#print(matrix)
scores = cosineComparison.KLdivergence(query, matrix, reference) #jointEntropy
scores.sort(key=lambda x: x[0])
print scores
xAxis = [pair[0] for pair in scores]
values = [pair[1] for pair in scores]
graphs.makeGraph(xAxis, "scores", "transcript budget log for {0}".format(queryName), values, queryName)
def KlHansard(dataset, queryTranscript, queryName, reference, smoothing):
# This is for when i have a specific target year to compare others to
#print dataset
#print(matrix)
scores = cosineComparison.KLdivergence(queryTranscript, dataset, reference,
smoothing) #jointEntropy
print scores
xAxis = [pair[0] for pair in scores]
values = [pair[1] for pair in scores]
graphs.makeGraph(
xAxis, "scores",
"KL Hansard for {0} Smoothing - {1}".format(queryName, smoothing),
values, queryName)
def KLBackwardsGraph(dataset, queryTranscript, queryName, reference,
smoothing):
# Want to compare all years to specific year. All years is reference to predict specific.
#print dataset
#print(matrix)
scores = cosineComparison.KLBackwardsDivergence(queryTranscript, dataset,
reference,
smoothing) #jointEntropy
print scores
xAxis = [pair[0] for pair in scores]
values = [pair[1] for pair in scores]
graphs.makeGraph(
xAxis, "scores",
"KL Hansard for {0} Smoothing - {1}".format(queryName, smoothing),
values, queryName)
def KlHansardIterative(dataset, reference, smoothing):
# I want to compare the previous year to current year itteratively
print reference
print len(dataset)
#print dataset[0]
klValues = []
yearsInvestigated = []
for i in range(1, len(dataset)):
priorBow = dataset[i - 1]
currentBow = dataset[i]
currentYear = reference[i]
yearsInvestigated.append(currentYear)
klValues.append(cosineComparison.KLbow(currentBow, priorBow,
smoothing))
print klValues
graphs.makeGraph(
yearsInvestigated, "scores",
"KL Iterative Opposition - Smooth = {0}".format(smoothing), klValues,
"not sure what this is")
def graphByMonthYearGraph(dicForm, title, context):
maxYear, minYear = calculateMinMaxYear(dicForm.keys())
numYears = maxYear - minYear + 1
print maxYear
print minYear
print numYears
monthScore = np.ones([12 * (numYears)]).tolist()
print "len is {0}".format(len(monthScore))
xLabel = numYears * generateMonths()
for date in dicForm:
#print date
yearOffset = date.year - minYear
monthOffset = date.month - 1
indexOffset = yearOffset * 12 + monthOffset
monthScore[indexOffset] += (
1 - dicForm[date]
) # attempt to make high look good. Note we need to normalise by number in that month
graphs.makeGraph(xLabel, "accumulative scores", title, monthScore, context)
def createGraph(referenceBudget, yearAsString):
referenceBudget = budget2006
budgets.remove(referenceBudget)
# budgets = [budget[0] for budget in budgetList]
# names = [[budget[1] for budget in budgetList]]
reference, documents = cosineComparison.arrayToBow(budgets, bowDirectory)
_, queryBow = cosineComparison.arrayToBow([referenceBudget], bowDirectory)
matrix, query = cosineComparison.matrixQueryNoTransformation(documents, queryBow[0])
#print matrix
scores = cosineComparison.inverseKLdivergence(query, matrix, reference) #jointEntropy
values = [pair[1] for pair in scores]
xAxis = [pair[0][0:4] for pair in scores]
graphs.makeGraph(xAxis, "scores", "budget log for {0}".format(yearAsString), values)
def makeGraphs():
course = request.args.get("course")
makeGraph(str(course))
return render_template("graph.html", courseNum=course)
partyFunction = hansardHandler.filenameToPartyInCharge
_, hansardBows, hansardReference = hansardHandler.budgetToBow(
initialYear, finalYear, None, partyFunction, True, True, False,
hansardSource)
#print hansardBows
#print reference
_, transcriptBows, transcriptReference = transcriptHandler.getTranscriptsBudgetDateTechnique(
initialYear, finalYear, None, None, transcriptSource, False, False, True)
#print transcriptBows
#print transcriptReference
klValues = []
xAxis = []
for i in range(len(transcriptBows)):
klValues.append(cosineComparison.KLbow(hansardBows[i], transcriptBows[i]))
print klValues
for i in range(len(klValues)):
print "{0} : {1}".format(transcriptReference[i], klValues[i])
xAxis.append(transcriptReference[i])
#xAxis = [pair[0] for pair in transcriptReference[i]]
graphs.makeGraph(xAxis, "scores", "KlHansardTranscripts", klValues)
