def isUnitSquare(rational_points):
# xxx rpoints and points must correspond xxx
points = [ p.toFloat() for p in rational_points ]
hull = ConvexHull(points)
hull_points = [ rational_points[i] for i in hull.vertices ]
line1 = Segment(Point(hull_points[0][0],hull_points[0][1]), Point(hull_points[1][0], hull_points[1][1]))
line2 = Segment(Point(hull_points[1][0],hull_points[1][1]), Point(hull_points[2][0], hull_points[2][1]))
# hull_points are guaranteed to be in counterclockwise order
angle = None
if line1.no_slope() and (not line2.no_slope() and line2.slope() == 0):
angle = pi / 2
elif line2.no_slope() and (not line1.no_slope() and line1.slope() == 0):
angle = pi / 2
else:
l1f = line1.toFloat()
l2f = line2.toFloat()
print "lines %s %s" % (l1f, l2f)
l1d = FloatPoint(l1f[1][0]-l1f[0][0],l1f[1][1]-l1f[0][1])
l2d = FloatPoint(l2f[1][0]-l2f[0][0],l2f[1][1]-l2f[0][1])
angle = atan2(l2d[1], l2d[0]) - atan2(l1d[1], l1d[0])
area = poly_area_indexed(rational_points, hull.vertices)
num_points = len(hull_points)
angle_delta = fmod(angle - (pi/2), 2*pi)
print "IsSquare: area:", area, "Num Points: ", num_points, "Angle Delta from 90 degrees, in radians: ", angle_delta
return area == 1 and num_points == 4 and abs(angle_delta) < epsilon
def _as_segment(data):
p1 = data[tags.p1]
p2 = data[tags.p2]
result = Segment(p1, p2)
if tags.color in data:
result.color = _as_color(data[tags.color])
return result
def __init__(self,
base=[0, 0, 0],
seglen=30.0,
ang=0.00643732691573, # focal length of 2m
r=5.151,
focal=200.0, conic=False):
'''
Constructor
Parameters:
base: the center point of the wide end of the segment
seglen: the axial length of each segment
ang: angle between the shell axis and the side of the front
segment
r: radius of the shell where the two segments meet
'''
if conic is False:
# Paraboloid segment
self.front = Segmentp(base=base, focal=focal, seglen=seglen, ang=ang, r1=r)
backBase = [base[0], base[1], base[2] + seglen]
# Hyperboloid segment
self.back = Segmenth(base=backBase, focal=focal, seglen=seglen, ang=ang, r0=r)
else:
self.front = Segment(base=base, seglen=seglen, ang=ang, r1=r)
backBase = [base[0], base[1], base[2]+seglen]
self.back = Segment(base=backBase, seglen=seglen, ang=3*ang, r0=r)
def create_non_intersecting_segments(pts1, pts2, segs2, angles, ref):
"""Creates list with the Segments that are visible from ref for the case
when the line segment in pts1 is in front of the line segment in pts2.
Args:
pts1 (list[np.array]): list of 2D points that are the intersection of
the rays from ref in the directions indicated by angles with the
Segment closest to ref
pts2 (list[np.array]): list of 2D points that are the intersection of
the rays from ref in the directions indicated by angles with the
Segment farthest to ref
segs2 (list[Segment]): list of Segment between the points in pts2
angles (list[float]): list of directions (radians) of the rays from ref
the angles are obtained from the two extremes of each original
Segment. The list always has 4 elements (even if repeated)
ref (np.array): reference point
"""
segments = []
segment = Segment(pts1[1], pts1[2], angles[1], angles[2], ref)
if pts1[0] is not None:
segment.merge(Segment(pts1[0], pts1[1], angles[0], angles[1], ref))
if pts1[3] is not None:
segment.merge(Segment(pts1[2], pts1[3], angles[2], angles[3], ref))
segments.append(segment)
if segs2[0] is not None:
segments = [Segment(pts2[0], pts2[1], angles[0], angles[1], ref)] + segments
if segs2[2] is not None:
segments.append(Segment(pts2[2], pts2[3], angles[2], angles[3], ref))
return segments
def row_to_object(row):
track = Track()
track.id = row.id
track.name = row.name
track.segments_old = [Segment.row_to_object(segment_row) for segment_row in row.segments_old]
track.segments_current = [Segment.row_to_object(segment_row) for segment_row in row.segments_current]
track.valid_time_start = row.valid_time_start
track.valid_time_stop = row.valid_time_stop
return track
def test_meets_conditions():
segment = Segment(['syllabic', 'voice'], ['consonantal', 'continuant'])
assert segment.meets_conditions({})
assert segment.meets_conditions({'positive': ['syllabic']})
assert segment.meets_conditions({'positive': ['syllabic', 'voice'],
'negative': ['continuant']})
assert not segment.meets_conditions({'positive': ['lateral'],
'negative': ['continuant']})
def main():
# cheap hex counter simulation :)
test_input = "01 23 45 67 89 ab cd ef" # "01 b0 c4 81 a6 c5 81 a0 5f 01 b7 5f"
logic = LogicNet(int('b', 16))
for char in test_input.replace(' ', ''):
logic.read_input(int(char, 16))
segment = Segment(logic.follow_paths())
# print segment.print_segment()
segment.print_chars()
def parse_segment(self, lc):
name = strip(self.f.read(16))
vmaddr = get_int(self.f) if self.macho.is_32_bit() else get_ll(self.f)
vmsize = get_int(self.f) if self.macho.is_32_bit() else get_ll(self.f)
offset = get_int(self.f) if self.macho.is_32_bit() else get_ll(self.f)
segsize = get_int(self.f) if self.macho.is_32_bit() else get_ll(self.f)
maxprot = get_int(self.f)
initprot = get_int(self.f)
nsects = get_int(self.f)
flags = get_int(self.f)
if self.macho.is_little():
vmaddr = little(vmaddr, 'I') if self.macho.is_32_bit() \
else little(vmaddr, 'Q')
vmsize = little(vmsize, 'I') if self.macho.is_32_bit() \
else little(vmsize, 'Q')
offset = little(offset, 'I') if self.macho.is_32_bit() \
else little(offset, 'Q')
segsize = little(segsize, 'I') if self.macho.is_32_bit() \
else little(segsize, 'Q')
maxprot = little(maxprot, 'I')
initprot = little(initprot, 'I')
nsects = little(nsects, 'I')
flags = little(flags, 'I')
maxprot = dictionary.protections[maxprot & 0b111]
initprot = dictionary.protections[initprot & 0b111]
entropy = self.get_segment_entropy(offset, segsize)
segment = Segment(cmd=lc.cmd, size=lc.size, name=name,
vmaddr=vmaddr, vmsize=vmsize, offset=offset,
segsize=segsize, maxprot=maxprot, initprot=initprot,
nsects=nsects, entropy=entropy)
if self.macho.is_32_bit():
sect_size = 68
else:
sect_size = 80
for i in range(segment.nsects):
if self.f.tell() + sect_size > self.file_size:
data = {
'offset': self.f.tell(),
'file_size': self.file_size
}
a = Abnormality(title='SECTION OUT OF BOUNDS', data=data)
break
sect = self.parse_section()
segment.add_sect(sect)
self.parse_segment_flags(segment, flags)
self.macho.add_lc(segment)
def parseSegment(self, lc):
name = strip(self._f.read(16))
vmaddr = getInt(self._f) if self._macho.is32Bit() else getLL(self._f)
vmsize = getInt(self._f) if self._macho.is32Bit() else getLL(self._f)
offset = getInt(self._f) if self._macho.is32Bit() else getLL(self._f)
segsize = getInt(self._f) if self._macho.is32Bit() else getLL(self._f)
maxprot = getInt(self._f)
initprot = getInt(self._f)
nsects = getInt(self._f)
flags = getInt(self._f)
if self._macho.isLittle():
vmaddr = little(vmaddr, 'I') if self._macho.is32Bit() \
else little(vmaddr, 'Q')
vmsize = little(vmsize, 'I') if self._macho.is32Bit() \
else little(vmsize, 'Q')
offset = little(offset, 'I') if self._macho.is32Bit() \
else little(offset, 'Q')
segsize = little(segsize, 'I') if self._macho.is32Bit() \
else little(segsize, 'Q')
maxprot = little(maxprot, 'I')
initprot = little(initprot, 'I')
nsects = little(nsects, 'I')
flags = little(flags, 'I')
maxprot = dictionary.protections[maxprot & 0b111]
initprot = dictionary.protections[initprot & 0b111]
segment = Segment(cmd=lc.getCmd(), size=lc.getSize(), name=name,
vmaddr=vmaddr, vmsize=vmsize, offset=offset,
segsize=segsize, maxprot=maxprot, initprot=initprot,
nsects=nsects)
if self._macho.is32Bit():
sect_size = 68
else:
sect_size = 80
for i in range(segment.getNSects()):
if self._f.tell() + sect_size > self._file_size:
data = {
'offset': self._f.tell(),
'file_size': self._file_size
}
a = Abnormality(title='SECTION OUT OF BOUNDS', data=data)
break
sect = self.parseSection()
segment.addSect(sect)
self.parseSegmentFlags(segment, flags)
self._macho.addLC(segment)
def object_hook(d):
"""
Usage
-----
>>> import simplejson as json
>>> with open('file.json', 'r') as f:
... json.load(f, object_hook=object_hook)
"""
from segment import Segment
from timeline import Timeline
from annotation import Annotation
from transcription import Transcription
if PYANNOTE_JSON_SEGMENT in d:
return Segment.from_json(d)
if PYANNOTE_JSON_TIMELINE in d:
return Timeline.from_json(d)
if PYANNOTE_JSON_ANNOTATION in d:
return Annotation.from_json(d)
if PYANNOTE_JSON_TRANSCRIPTION in d:
return Transcription.from_json(d)
return d
def __init__(self, track, comp_location, start, orig_duration, new_duration):
"""Create a time-stetched segment.
It acts like a :py:class:`radiotool.composer.Segment` but you
can specify the target duration. The segment will then
resample its frames to meet this duration.
:param track: Track to slice
:type track: :py:class:`radiotool.composer.Track`
:param float comp_location: Location in composition to play this segment (in seconds)
:param float start: Start of segment (in seconds)
:param float orig_duration: Original duration of segment (in seconds)
:param float new_duration: Target (stretched) duration of segment (in seconds)
"""
Segment.__init__(self, track, comp_location, start, new_duration)
self.orig_duration = int(orig_duration * self.samplerate)
def token_to_segment(token, segment_list, diacritic_list):
'''Converts a string token in IPA to Segment object, given
a list of dictionaries representing segments and the same representing
diacritics.'''
diacritic_strings = [segment['IPA'] for segment in diacritic_list]
# Isolate the base IPA segment string
base_string = ''.join(filter(lambda x: x not in diacritic_strings,
token))
# Isolate an iterable of diacritics present
diacritics = [diacritic for diacritic in diacritic_list
if diacritic['IPA'] in token]
# Initialise the base Segment
segment = Segment.from_dictionary(find_segment(base_string,
segment_list))
# Add each diacritic feature to the segment
for diacritic in diacritics:
diacritic_segment = Segment(diacritic['applies'].get('positive', []),
diacritic['applies'].get('negative', []))
segment = segment + diacritic_segment
return segment
def __init__(self,
base = [0,0,0],
seglen = 30.0,
focal = 200.0,
radii = [5.151,4.9,4.659,4.429,4.21,4.0,3.799],
angles = None
):
'''
Constructor
Parameters:
base: the center point of the wide end of the segment
seglen: the axial length of each segment
focal: the focal length, measured from the center of the module
radii: a list of radii, one for each shell from biggest to smallest
angles: optional parameter to overwrite the shell angles computed by constructor
'''
if angles is None:
angles = calcShellAngle(radii,focal)
elif len(radii) != len(angles):
raise ValueError('number of radii and number of angles do not match')
self.shells = []
for i,r in enumerate(radii):
self.shells.append(Shell(base=base, seglen=seglen, ang=angles[i], r=r))
# inner core (blocks rays going through center of module)
r0 = self.shells[-1].back.r0
r1 = r0 - seglen * tan(4*angles[-1])
ang = atan((r0-r1)/(2*seglen))
self.core = Segment(base=base, seglen=2*seglen, ang=ang, r0=r0)
self.coreFaces = [Circle(center=base,normal=[0,0,1],radius=r0),
Circle(center=[base[0],base[1],base[2]+2*seglen],normal=[0,0,-1],radius=r1)]
def test_initialisation():
feature_dictionary = {'stress': '+', 'long': '-', 'continuant': '0',
'IPA': 'b'}
segment = Segment.from_dictionary(feature_dictionary)
assert segment.positive == ['stress']
assert segment.negative == ['long']
def from_json(cls, data):
uri = data.get(PYANNOTE_URI, None)
modality = data.get(PYANNOTE_MODALITY, None)
annotation = cls(uri=uri, modality=modality)
for s, track, label in data[PYANNOTE_JSON_ANNOTATION]:
segment = Segment.from_json(s)
annotation[segment, track] = label
return annotation
def __init__(self,
base = [0,0,0],
seglen = 30.0,
ang = 0.00643866102405, # focal length of 2m
r = 5.151
):
'''
Constructor
Parameters:
base: the center point of the wide end of the segment
seglen: the axial length of each segment
ang: angle between the shell axis and the side of the front segment
r: radius of the shell where the two segments meet
'''
self.front = Segment(base=base, seglen=seglen, ang=ang, r1=r)
backBase = [base[0], base[1], base[2]+seglen]
self.back = Segment(base=backBase, seglen=seglen, ang=3*ang, r0=r)
def test_addition():
feature_dictionary = {'stress': '+', 'syllabic': '-', 'continuant': '0',
'IPA': 'b'}
segment = Segment.from_dictionary(feature_dictionary)
syllabic_diacritic = Segment(['syllabic'], ['voice'])
addition = segment + syllabic_diacritic
assert addition.positive == ['stress', 'syllabic']
assert addition.negative == ['voice']
def from_json(cls, data):
uri = data.get(PYANNOTE_URI, None)
modality = data.get(PYANNOTE_MODALITY, None)
annotation = cls(uri=uri, modality=modality)
for one in data[PYANNOTE_JSON_CONTENT]:
segment = Segment.from_json(one[PYANNOTE_SEGMENT])
track = one[PYANNOTE_TRACK]
label = one[PYANNOTE_LABEL]
annotation[segment, track] = label
return annotation
def row_to_object(row):
run = Run()
run.id = row.id
run.user_id = row.run_info.user_id
run.user_bt_name = row.run_info.user_bt_name
run.segment = Segment.row_to_object(row.run_info.segment)
print 'type of time start cell'
print str(type(row.run_info.time_start))
run.time_start = row.run_info.time_start
run.time_stop = row.run_info.time_stop
run.time_span_ms = row.run_info.time_span_ms
return run
def benchmark_match_accuracy(segments, diacritics, filename):
'''Convert all given segments to feature strings, then convert back to
segments. Use the given feature string file. Return the percentage accuracy
of the conversion.
'''
feature_strings = load_feature_strings(path.join(base_directory, 'engine',
'data', filename))
print('Loaded {0} feature strings'.format(len(feature_strings)))
base_matches = []
matches = []
deparse.initialise_cache()
for segment in segments:
base_segment = Segment.from_dictionary(segment)
base_matches.append((segment['IPA'],
deparse.segment_match(feature_strings,
base_segment)))
matches.append((segment['IPA'], deparse.segment_match(feature_strings,
base_segment)))
for diacritic in diacritics:
IPA_representation = segment['IPA'] + diacritic['IPA']
if base_segment.meets_conditions(diacritic['conditions']):
diacritic_segment = base_segment + Segment(diacritic['applies'].get('positive', []),
diacritic['applies'].get('negative', []))
matches.append((IPA_representation,
deparse.segment_match(feature_strings,
diacritic_segment)))
print('Calculating base accuracy...')
base_successes = 0
for match in base_matches:
if match[0] == match[1]:
base_successes += 1
else:
print('\tExpected {0}, deparsed {1}'.format(match[0], match[1]))
print('Calculating diacritic accuracy...')
successes = len([match for match in matches if match[0] == match[1]])
return (base_successes / len(base_matches)), (successes / len(matches))
def test_setters():
feature_dictionary = {'stress': '+', 'long': '-', 'continuant': '0',
'IPA': 'b'}
segment = Segment.from_dictionary(feature_dictionary)
segment.add_positive('long')
assert segment.positive == ['stress', 'long']
assert segment.negative == []
segment.add_negative('stress')
assert segment.positive == ['long']
assert segment.negative == ['stress']
segment.add_negative('stress')
assert segment.positive == ['long']
assert segment.negative == ['stress']
def _travel_path(origin_id, start_id): path = [] branch_encounted = False head = start_id while True: path.append(head) b = Segment.by_id(head) if not b: b = Block.by_id(head) if origin_id == head: break if b.has_return(): break if b.branch() and not b.branch().has_return(): branch_encounted = True break head = b.passthru().id return path, branch_encounted
def _new_segment(segment, segments, included_in, members): id = segment.id segments.append(segment.id) m = [] if segment.type == Segment.BRANCH: m = segment.left[:] m.extend(segment.right[:]) elif segment.type == Segment.LOOP: m = segment.body[:] sm = [] for block_id in m: if Segment.by_id(block_id): if block_id not in included_in: # temporary included_in[block_id] = id sm.extend(members[block_id]) sm.remove(block_id) # remove dup m.append(id) m.extend(sm) members[id] = m
track_features_small = pickle.load(open('trackToFeatureTuples.p', 'rb'))
clusters = km.kmeans(track_features_small, 10)
track_features_full = j.constructNamesToDict()
cluster_num = 0
for cluster in clusters:
# list of Segment objects
cluster_30segments = []
starts = []
for track in cluster:
segments = track_features_full[track]["segments"]
segment30 = []
for i in xrange(len(segments)):
s = Segment(track, i, segments[i])
segment30.append(s)
if (i + 1) % 30 == 0:
cluster_30segments.append(segment30)
if i + 1 == 30:
starts.append(segment30)
segment30 = []
start = random.choice(starts)
sp = mashupSearchProblem(cluster_30segments, start, 5, sc.segmentsCost)
ucs_alg = ucs.UniformCostSearch()
ucs_alg.solve(sp)
mashup = [s for s in start] + [s for a in ucs_alg.actions for s in a]
segment_titles = []
def from_json(cls, data):
uri = data.get(PYANNOTE_URI, None)
segments = [Segment.from_json(s) for s in data[PYANNOTE_JSON_CONTENT]]
return cls(segments=segments, uri=uri)
def from_json(cls, data):
segments = [Segment.from_json(s) for s in data[PYANNOTE_JSON_TIMELINE]]
uri = data.get(PYANNOTE_URI, None)
return cls(segments=segments, uri=uri)
class Shell:
'''
A shell consists of two segments, one in the front and one behind. The base location
of the front segment has a smaller z value than that of the back segement. Thus, rays
coming from sources in the negative z range will enter the front end first.
'''
def __init__(self,
base = [0,0,0],
seglen = 30.0,
ang = 0.00643866102405, # focal length of 2m
r = 5.151
):
'''
Constructor
Parameters:
base: the center point of the wide end of the segment
seglen: the axial length of each segment
ang: angle between the shell axis and the side of the front segment
r: radius of the shell where the two segments meet
'''
self.front = Segment(base=base, seglen=seglen, ang=ang, r1=r)
backBase = [base[0], base[1], base[2]+seglen]
self.back = Segment(base=backBase, seglen=seglen, ang=3*ang, r0=r)
def getSurfaces(self):
'''
Returns a list of surfaces
'''
return [self.front,self.back]
def plot2D(self, axes, color = 'b'):
'''
Plots a 2d cross section of the shell
'''
self.front.plot2D(axes, color)
self.back.plot2D(axes, color)
# plot rays
ang = self.front.ang
r = self.back.r1
z = self.back.base[2]
d = r*tan((pi/2)-4*ang)
axes.plot((2*z,2*z+d),(r,0),'y:')
axes.plot((2*z,2*z+d),(-r,0),'y:')
def plot3D(self, axes, color = 'b'):
'''
Generates a 3d plot of the shell in the given figure
'''
self.front.plot3D(axes,color)
self.back.plot3D(axes,color)
def targetFront(self,a,b):
'''
Takes two list arguments of equal size, the elements of which range from 0 to 1.
Returns an array of points that exist on the circle defined by the wide end of
the shell.
'''
return self.front.targetFront(a,b)
def targetBack(self,a,b):
'''
Takes two list arguments of equal size, the elements of which range from 0 to 1.
Returns an array of points that exist on the circle defined by the small end of
the shell.
'''
return self.back.targetBack(a,b)
def createSegment(self, vt1, vt2):
segment = Segment(self)
segment.setPos(vt1, vt2)
return segment
def __init__(self, host, port):
self.load()
self.segmentor= Segment(host, port)
def set_process(sentences, delimiter='<!>', line_break='\n', option=None):
dep_out = False
phr_out = False
relaxed_out = False
laa_out = False
collx_out = False
use_collx = False
phrases = False
print_tree = False
phr_verbs_path = ''
pos_tags = True
desamb_hacks = False
grammar = os.path.join(set_path, 'grammar.set')
lw_collx = os.path.join(set_path, 'data', 'lemma-word')
ll_collx = os.path.join(set_path, 'data', 'lemma-lemma')
lt_collx = os.path.join(set_path, 'data', 'lemma-tag')
ls_collx = os.path.join(set_path, 'data', 'lemma-semclass')
ss_collx = os.path.join(set_path, 'data', 'semclass-semclass')
if option == '-d': dep_out = True
elif option == '-p': phr_out = True
elif option == '--cout': collx_out = True
elif option == '--phrases': phrases = True
elif option == '--laa': laa_out = True
phr_verbs = {}
if phr_verbs_path:
for line in open(phr_verbs_path):
llist = line.split()
if llist[0] not in phr_verbs: phr_verbs[llist[0]] = []
phr_verbs[llist[0]].append(tuple(llist[1:]))
grammar = Grammar(grammar)
if use_collx:
parser = Parser(grammar, lw_collx_root=lw_collx,
ll_collx_root=ll_collx, lt_collx_root=lt_collx,
ls_collx_root=ls_collx, ss_collx_root=ss_collx)
else:
parser = Parser(grammar)
def sigusr1_handler(*args):
#open('/tmp/set-sigusr1-last-sentence', 'w').write(''.join(lines))
sys.exit(1)
signal.signal(signal.SIGUSR1, sigusr1_handler)
chunks = sentences.split(delimiter)
signal.alarm(len(chunks) * 10)
f = TextFile()
for sentence in chunks:
sentence_chunks = [w.replace('\t', ' ')
for w in sentence.split(line_break) if w and w[0] != "<"]
if len(sentence_chunks) == 0:
continue
s = Segment(sentence_chunks, pos_tags=pos_tags,
desamb_hacks=desamb_hacks, phr_verbs=phr_verbs)
parser.parse(s)
if dep_out: s.print_dep_tree(root=False, file=f)
if laa_out: s.print_laa_tree(file=f)
if phr_out: s.print_phr_tree(file=f)
if relaxed_out: s.print_relaxed_tree(file=f)
if collx_out: s.print_collx_out(file=f)
if print_tree: s.print_tree(file=f)
if phrases: s.print_phrases(file=f)
f.write('\n')
return f.get()
class Classfier():
segmentor = None
features={}
categorys={}
model={}
def __init__(self, host, port):
self.load()
self.segmentor= Segment(host, port)
def __del__(self):
if self.segmentor:
self.segmentor = None
self.features = None
self.categorys = None
self.model = None
def load(self):
feature_file='thrift/etc/feature.txt'
category_file='thrift/etc/catid.txt'
model_file='thrift/data/knn.model'
# load centoroid
for line in open(model_file):
line = line.strip()
if not line:
break
index, rest=line.split('|')
centoroid = {}
for part in rest.split(" "):
k, v = part.split(':')
centoroid[int(k)] = float(v)
self.model[int(index)] = centoroid
# load features
for line in open(feature_file):
line=line.strip()
parts = line.split("\t")
feature = {
"index": int(parts[0]),
"idf" : float(parts[2])
}
word = parts[1]
self.features[word] = feature
# load category
for line in open(category_file):
line = line.strip()
catid, leaf, root = line.split("\t")
self.categorys[int(catid)] = leaf
def predict(self, content):
features = self.text2Feature(content)
return self.maxProbCat(features);
def segment(self, content):
self.segmentor.connect()
result = self.segmentor.segment(content)
self.segmentor.close()
return self.segmentor.parse(result)
def similarity(self, docFeatures, centoroid):
value = 0.0
for k,v in enumerate(docFeatures):
item = docFeatures[v]
if (centoroid.has_key(v)):
value += item['weight'] * centoroid[v]
return value
def text2Feature(self, content):
keywords = self.segment(content)
#print keywords
features = {}
for wordTf in keywords:
word, tf = wordTf.split(':')
tf = float(tf)
if self.features.has_key(word):
feature = self.features[word]
index = feature["index"]
idf = feature["idf"]
features[index] = {"word":word, "weight":tf*idf}
#print features
return features
def maxProbCat(self, features):
maxScore = 0.0
probCat = -1
#docDotProduct = math.sqrt(sum(f.weight * f.weight for f in features))
#if docDotProduct == 0:
#return probCat
for k, v in enumerate(self.model):
score = self.similarity(features, self.model[v]);
#print self.categorys[v], int(score * 10000)
if score > maxScore:
maxScore = score
probCat = v
label = self.categorys[probCat]
return label