def append(self, image_filename):
_hash = get_hash(os.path.join(self.image_lib_path, image_filename))
pair = _hash, image_filename
self.hashes.append(pair)
for cur_hash, f in self.hashes:
if self._resemble(_hash, cur_hash):
self.img_set.union(pair, (cur_hash, f))
self.img_set.union(pair, pair)
def query(self, image_handle=None, image_path=None, image_url=None, image_hash=None):
print 'in query'
if image_hash:
print 'image hash'
return self._query(image_hash)
if image_handle:
print 'image handle'
return self._query(get_hash(image_handle))
if image_path:
print 'image path'
return self._query(get_hash(image_path))
if image_url:
print 'image url'
print 'retrieving image from %s' % image_url
raw_image = urllib2.urlopen(image_url).read()
print 'image retrieved'
image_handle = StringIO(raw_image)
return self._query(get_hash(image_handle))
return []
def refresh(self):
self.hashes = []
self.img_set = DisjointSet()
for filename in os.listdir(self.image_lib_path):
print filename
_hash = get_hash(os.path.join(self.image_lib_path, filename))
self.hashes.append((_hash, filename.decode(filesystem_encoding).encode('utf-8')))
for pair in self.hashes:
self.img_set.union(pair, pair)
for (h1, f1), (h2, f2) in combinations(self.hashes, 2):
if self._resemble(h1, h2):
self.img_set.union((h1, f1), (h2, f2))
# This module has been deprecated.
from itertools import combinations
import os
from web.lib import pha
from web.lib.constants import IMAGE_LIBRARY_PATH, confidence
from web.lib.disjoint_set import DisjointSet
if __name__ == "__main__":
print os.getcwd()
image_lib_path = os.path.join("web", IMAGE_LIBRARY_PATH)
hashes, img_set = [], DisjointSet()
for filename in os.listdir(image_lib_path):
_hash = pha.get_hash(os.path.join(image_lib_path, filename))
hashes.append((_hash, filename.decode("gbk").encode("utf-8")))
for pair in hashes:
img_set.union(pair, pair)
for (h1, f1), (h2, f2) in combinations(hashes, 2):
cnt = bin(h1 ^ h2).count("1")
if cnt <= confidence:
img_set.union((h1, f1), (h2, f2))
print "-" * 64
print "%016x" % h1, "%016x" % h2
print "% 16s" % f1, "% 16s" % f2
print "similarity: %2.2f%%" % (((64 - cnt) / 64.0) * 100)
print "{0:064b}".format(h1 ^ h2)
