def process(pdf, coarse_rotate_angle):
""" Extract images, fix skew, and get line positions, writing images and data to files. There is no return value."""
## Get images from PDF
images = extract.extractImages(pdf)
## Straighten images
if coarse_rotate_angle < 0:
coarse_rotate_angle = 360 - (abs(coarse_rotate_angle) % 360)
else:
coarse_rotate_angle = coarse_rotate_angle % 360
print "angle:" + str(coarse_rotate_angle)
images = [skew.straighten(skew.coarserotate(i, coarse_rotate_angle)) for i in images]
print "images found: " + str(len(images))
## Iterate through list of images, storing lists of corresponding line locations
## in hash table.
## Then write each image to file.
## File naming convention: <original filename><page number>.jpg
n=1
lines = {}
for i in images:
p = pdf+str(n)+".jpg"
lines[n]=getlines.getLines(i)
cv2.imwrite(p,i)
n = n+1
## Write line data in JSON format to a .txt file
## File naming convention: <original filename>.json.txt
with open ((pdf+".json.txt"),'w') as outfile:
json.dump(lines,outfile)
def process(pdf, coarse_rotate_angle):
""" Extract images, fix skew, and get line positions, writing images and data to files. There is no return value."""
## Get images from PDF
images = extract.extractImages(pdf)
## Straighten images
if coarse_rotate_angle < 0:
coarse_rotate_angle = 360 - (abs(coarse_rotate_angle) % 360)
else:
coarse_rotate_angle = coarse_rotate_angle % 360
print "angle:" + str(coarse_rotate_angle)
images = [
skew.straighten(skew.coarserotate(i, coarse_rotate_angle))
for i in images
]
print "images found: " + str(len(images))
## Iterate through list of images, storing lists of corresponding line locations
## in hash table.
## Then write each image to file.
## File naming convention: <original filename><page number>.jpg
n = 1
lines = {}
for i in images:
p = pdf + str(n) + ".jpg"
lines[n] = getlines.getLines(i)
cv2.imwrite(p, i)
n = n + 1
## Write line data in JSON format to a .txt file
## File naming convention: <original filename>.json.txt
with open((pdf + ".json.txt"), 'w') as outfile:
json.dump(lines, outfile)
def setUp(self):
#text with large amount of black space from copier error
self.image = cv2.imread("testimg/blackmarginsfixed.jpg")
self.img, self.angle = skew.straighten(
cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY))
self.lines = getlines.getLines(self.img)
def setUp(self):
#text with one-area noise (from bent corner)
self.image = cv2.imread("testimg/noisy2fixed.jpg")
self.img, self.angle = skew.straighten(
cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY))
self.lines = getlines.getLines(self.img)
def setUp(self):
#text with background noise (from newspaper)
self.image = cv2.imread("testimg/noisy1fixed.jpg")
self.img, self.angle = skew.straighten(
cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY))
self.lines = getlines.getLines(self.img)
def setUp(self):
#one picture image
self.image = cv2.imread("testimg/rotated.jpg")
self.img, self.angle = skew.straighten(
cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY))
self.lines = getlines.getLines(self.img)
def test_calculate_angle(self):
#compares calculated angle to observed angle to make sure they are equal
img, angle = skew.straighten(
cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY))
self.assertTrue(abs(angle - 0) < 2)