def test_AttributeLookup(self):
fi= self.fi; ki= self.ki # field and keyword index
obj1= self.obj1; obj2= self.obj2
kw= OOSet((1,2,))
## Acquisition type
# implicit
self.assertEqual(fi._evaluate(obj2),'id')
self.assertEqual(fi._evaluate(obj1),None)
# "OOSet" does not support intelligent equality test
self.assertEqual(ki._evaluate(obj2).keys(),kw.keys())
self.assertEqual(ki._evaluate(obj1).keys(),kw.keys())
# none
ki.kw.AcquisitionType= 'none'
self.assertEqual(ki._evaluate(obj2),None)
self.assertEqual(ki._evaluate(obj1).keys(),[1,2])
# explicit
# same as implicit for non-methods
ki.kw.AcquisitionType= 'explicit'
self.assertEqual(fi._evaluate(obj2),'id')
self.assertEqual(fi._evaluate(obj1),None)
self.assertEqual(ki._evaluate(obj2).keys(),kw.keys())
self.assertEqual(ki._evaluate(obj1).keys(),kw.keys())
# now check methods
fi.id_.Name= 'fid'; ki.kw.Name= 'fkw'
self.assertEqual(fi._evaluate(obj2),'id')
self.assertEqual(fi._evaluate(obj1),None)
self.assertEqual(ki._evaluate(obj2),None)
self.assertEqual(ki._evaluate(obj1).keys(),kw.keys())
## call types
# call -- already checked
# return
fi.id_.CallType= 'return'
self.assertEqual(fi._evaluate(obj2), obj1.fid)
# ignore
fi.id_.CallType= 'ignore'
self.assertEqual(fi._evaluate(obj2), None)
fi.id_.CallType= 'call'
## IgnoreExceptions
fi.id_.IgnoreExceptions= 0
self.assertRaises(AttributeError,fi._evaluate,obj1)
self.assertEqual(fi._evaluate(obj2),'id')
class TaggingEngine(Persistent):
interface.implements(ITaggingEngine)
def __init__(self):
self.clear()
@property
def tagsCount(self):
return len(self.tag_oids)
@property
def itemsCount(self):
return len(self.oid_tags)
def clear(self):
self.tagsmap = LOBTree()
# index
self.tag_oids = LOBTree()
self.oid_tags = LOBTree()
# tag weight
self.weights = OOSet()
self.tag_weight = LFBTree()
def getHash(self, str):
if not str:
return 0 # empty
res = hash(str)
# NOTE: workaround for 64bit
if struct.calcsize("P") * 8 == 64:
res = ord(str[0]) << 7
for char in str:
res = c_mul(1000003, res) ^ ord(char)
res = res ^ len(str)
if res == -1:
res = -2
if res >= 2**31:
res -= 2**32
return res
def update(self, oid, tags):
self.remove(oid)
for tag in tags:
htag = self.getHash(tag)
self.tagsmap[htag] = tag
# add oid -> tag
oids = self.tag_oids.get(htag)
if oids is None:
oids = LFSet()
self.tag_oids[htag] = oids
if oid not in oids:
oids.insert(oid)
# add tag -> oid
oid_tags = self.oid_tags.get(oid)
if oid_tags is None:
oid_tags = LLSet()
self.oid_tags[oid] = oid_tags
if htag not in oid_tags:
oid_tags.insert(htag)
# culculate weight
weight = self.tag_weight.get(htag)
if weight is not None:
key = (weight, htag)
if key in self.weights:
self.weights.remove(key)
weight = float(len(oids))
self.tag_weight[htag] = weight
self.weights.insert((weight, htag))
def remove(self, oid):
for tag in self.oid_tags.get(oid, ()):
# remove oid from tag -> oids reference
oids = self.tag_oids.get(tag)
if oid in oids:
oids.remove(oid)
oids_len = float(len(oids))
# reculculate weight
weight = self.tag_weight.get(tag)
if weight is not None:
key = (weight, tag)
if key in self.weights:
self.weights.remove(key)
if oids_len:
self.tag_weight[tag] = oids_len
self.weights.insert((oids_len, tag))
# remove tag
if not oids_len:
del self.tag_oids[tag]
del self.tagsmap[tag]
if oid in self.oid_tags:
del self.oid_tags[oid]
def getItems(self, tags):
oids = self.tag_oids
weights = self.tag_weight
weight, result = 0, LFBTree()
for tag in tags:
htag = self.getHash(tag)
if htag in oids:
weight, result = weightedUnion(
oids.get(htag), result, weights.get(htag), weight)
return IFBucket(result)
def getUniqueItems(self, tags):
oids = self.tag_oids
weights = self.tag_weight
weight, result = 1.0, None
for tag in tags:
htag = self.getHash(tag)
if htag in oids:
if result is None:
weight, result = weightedUnion(
oids.get(htag), LFBTree(), weights.get(htag), weight)
else:
weight, result = weightedIntersection(
result, oids.get(htag), weight, weights.get(htag))
return IFBucket(result)
def getTagCloud(self, reverse=False):
total = len(self.oid_tags)
if not total:
return
tags = self.tagsmap
data = self.weights.keys()
percent = total / 100.0
if reverse:
first = len(data)-1
for idx in xrange(first, -1, -1):
weight, htag = data[idx]
yield weight / percent, tags.get(htag)
else:
for weight, htag in data:
yield weight / percent, tags.get(htag)
def getItemsTagCloud(self, items):
oid_tags = self.oid_tags
tags = [oid_tags[oid] for oid in items if oid in oid_tags]
if tags:
tags = multiunion(tags)
else:
return
total = len(oid_tags)
data = self.weights.keys()
weights = self.tag_weight
percent = total / 100.0
for tag in tags:
yield weights[tag] / percent, self.tagsmap.get(tag)
def getFrequency(self, tags):
tagsmap = self.tagsmap
tag_weight = self.tag_weight
for tag in tags:
yield (tag, tag_weight.get(self.getHash(tag), 0))
def __nonzero__(self):
return bool(self.tag_oids)
def __contains__(self, tag):
return self.getHash(tag) in self.tagsmap
