def highboost():
url = req.args.get('link')
if not url:
return render_template('standard.html')
try:
# download the image from the url
res = requests.get(url)
# open the image using PIL
im = img.open(sIO(res.content))
# convert the PIL image to a numpy array and turn it into a newt image
pic = newt(array(im, dtype=double))
# do a convolution with a 17x17 disk
pic.highboost('d 17')
# revert to PIL format
pic = pic.pic - min(pic.pic)
pic = 255*pic/max(pic)
im = img.fromarray(pic.astype('uint8'))
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
except:
return redirect(url)
def srtfilt():
url = req.args.get('link')
if not url:
return render_template('standard.html')
try:
# download the image from the url
res = requests.get(url)
# open the image using PIL
im = img.open(sIO(res.content))
# shrink very large images
im.thumbnail((512,512), img.ANTIALIAS)
# convert the PIL image to a numpy array and turn it into a newt image
pic = newt(array(im, dtype=complex))
# do stuff
pic.srtfilt()
# revert to PIL format
#pic = pic.pic - min(pic.pic)
#pic = 255*pic/max(pic)
im = img.fromarray(pic.pic.astype('uint8'))
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
except:
return redirect(url)
def getHistoricalRates(indexSymbol):
req = urllib2.urlopen(source+indexSymbol)
rates = req.read()
dReader = csv.DictReader(sIO(rates))
histRates = {}
for row in dReader:
histRates[dt.datetime.strptime(row['Date'], dateFormat).date()] = (float(row['Adj Close'])/100.0)/365
return histRates
def boxesmap():
siz = double(req.args.get('size'))
siz = min([max([siz, 50]), 200])
pic = random.randint(-1, 2, size=(siz, siz))
temp = zeros(pic.shape);
value = 1
def pos(x, y):
if (x<0):
x=pic.shape[0]-1
elif (x>pic.shape[0]-1):
x=0
if (y<0):
y=pic.shape[1]-1
elif (y>pic.shape[1]-1):
y=0
return x, y
for t in range(20):
for row in range(pic.shape[0]):
for col in range(pic.shape[1]):
temp[row, col] = pic[pos(row, col-1)] +\
pic[pos(row-1, col)] + pic[pos(row, col)] + pic[pos(row+1, col)] +\
pic[pos(row, col+1)]
temp[row, col] = double(temp[row, col] > 0) - double(temp[row, col] < 0)
value = sum(pic[:]-temp[:])
for row in range(pic.shape[0]):
for col in range(pic.shape[1]):
pic[row, col] = temp[row, col]
pic = random.randint(-1, 2, size=(siz, siz, 3))
neg = 255, 255, 255
zer = 255, 134, 156
pos = 12, 163, 255
for row in range(pic.shape[0]):
for col in range(pic.shape[1]):
if (temp[row, col] ==-1):
pic[row,col] = neg
elif (temp[row, col] ==0):
pic[row, col] = zer
else:
pic[row, col] = pos
im = img.fromarray(pic.astype('uint8'))
im = im.resize((500, 500))
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
def hny():
url = req.args.get('link')
if not url:
return render_template('standard.html')
try:
# download the image from the url
res = requests.get(url)
res2 = requests.get("https://c2.staticflickr.com/4/3007/2733380075_7c8019d4eb.jpg")
# https://c2.staticflickr.com/6/5330/8808102199_bff96f1f80_o.jpg
# open the image using PIL
im = img.open(sIO(res.content))
im2 = img.open(sIO(res2.content)).resize(im.size, img.ANTIALIAS)
# convert the PIL image to a numpy array and turn it into a newt image
a = array(im, dtype=double)
b = array(im2, dtype=double)
for color in range(3):
a[:,:][:,:,color] = a[:,:][:,:,0]/3 + a[:,:][:,:,1]/3 + a[:,:][:,:,2]/3
a[:,:][:,:,color] = 4*a[:,:][:,:,color]/7 + 3*b[:,:][:,:,color]/7
pic = newt(a)
# do a convolution with a 17x17 disk
#pic.mix(array(im2, dtype=double))
# revert to PIL format
pic = pic.pic - min(pic.pic)
pic = 255*pic/max(pic)
im = img.fromarray(pic.astype('uint8'))
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
except:
return redirect(url)
def hny():
url = req.args.get('link')
if not url:
return render_template('standard.html')
try:
# download the image from the url
res = requests.get(url)
res2 = requests.get("http://www.themarysue.com/wp-content/uploads/2012/08/c6bfb4fac68932e833f917cd45ad2ff9.jpeg")
# open the image using PIL
im = img.open(sIO(res.content))
im2 = img.open(sIO(res2.content)).resize(im.size, img.ANTIALIAS)
# convert the PIL image to a numpy array and turn it into a newt image
a = array(im, dtype=double)
b = array(im2, dtype=double)
for color in range(3):
a[:,:][:,:,color] = a[:,:][:,:,0]/3 + a[:,:][:,:,1]/3 + a[:,:][:,:,2]/3
a[:,:][:,:,color] = 4*a[:,:][:,:,color]/7 + 3*b[:,:][:,:,color]/7
pic = newt(a)
# do a convolution with a 17x17 disk
#pic.mix(array(im2, dtype=double))
# revert to PIL format
pic = pic.pic - min(pic.pic)
pic = 255*pic/max(pic)
im = img.fromarray(pic.astype('uint8'))
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
except:
return redirect(url)
def asf():
url = req.args.get('link')
if not url:
return render_template('scientific.html')
try:
# download the image from the url
res = requests.get(url)
# open the image using PIL
im = img.open(sIO(res.content))
# convert the PIL image to a numpy array and turn it into a newt image
pic = newt(array(im, dtype=double))
# do a convolution with a 17x17 disk
pic.dhat('g 7')
# revert to PIL format
z = zeros(pic.pic[:,:,0].shape)
red = dstack((pic.pic[:,:,0],z,z))
green = dstack((z,pic.pic[:,:,1],z))
blue = dstack((z,z,pic.pic[:,:,2]))
red = 255*red/max(red)
green = 255*green/max(green)
blue = 255*blue/max(blue)
pic = vstack((red,green,blue))
im = img.fromarray(pic.astype('uint8'))
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
except:
return redirect(url)
def readStocksFromFile(stockCount, markIndex, sortBy, reverse):
if markIndex=="SP":
f = open("S&P500.csv")
else:
f = open("DOW30.csv")
vals = f.read()
dReader = csv.DictReader(sIO(vals))
stocks = [(row['Symbol'], row[sortBy]) for row in dReader]
stocks = sorted(stocks, key=itemgetter(1), reverse=reverse)
if stockCount < len(stocks):
return stocks[:stockCount]
else:
return stocks
def laplacian():
url = req.args.get('link')
if not url:
return render_template('scientific.html')
try:
wordcloud = wc().generate(url)
im = wordcloud.to_image()
# save the new image
buff = sIO()
im.save(buff, 'JPEG', quality=90)
buff.seek(0)
return send_file(buff, mimetype='image/jpeg')
except:
return redirect(url)
def parse_rdf(string, model=None, context="none"):
if model == None:
model = bound_graph()
model.parse(sIO(string))
return model
def getHistoricalPrices(stockSymbol): req = urllib2.urlopen(source+stockSymbol) prices = req.read() dReader = csv.DictReader(sIO(prices)) histPrices = [fin.AssetPrice(dt.datetime.strptime(row['Date'], dateFormat).date(), row['Open'], row['High'], row['Low'], row['Close'], row['Volume'], row['Adj Close']) for row in dReader] return histPrices
def parse_rdf(string, model=None, context="none"):
if model == None:
model = bound_graph()
model.parse(sIO(string))
return model
def url_to_pdf(url):
'''Treat give url as a pdf, to be fed into pdfminer functions'''
open = urlopen(Request(url)).read()
return sIO(open)
