def lollipopController(name, color = 0, size = [1, 1, 1], lockAttr = ['tr', 'ro', 'sc', 'vi'], lock = True):
#cmds.createNode('transform', n = '%s_grp' %name)
cmds.curve(n = name, d = 1, p = [ (0, 0, 0), (0, 3, 0),(0, 4, 1),(0, 5, 0), (0, 4, -1),(0, 3, 0),(0, 4, 1),(0, 4, -1)], k = [1,2,3,4,5,6,7,8])
cmds.scale(size[0],size[1],size[2])
for attr in lockAttr:
if attr == 'tr':
cmds.setAttr('%s.translateX' %name, lock = lock, keyable = False, channelBox = False)
cmds.setAttr('%s.translateY' %name, lock = lock, keyable = False, channelBox = False)
cmds.setAttr('%s.translateZ' %name, lock = lock, keyable = False, channelBox = False)
elif attr == 'ro':
cmds.setAttr('%s.rotateX' %name, lock = lock, keyable = False, channelBox = False)
cmds.setAttr('%s.rotateY' %name, lock = lock, keyable = False, channelBox = False)
cmds.setAttr('%s.rotateZ' %name, lock = lock, keyable = False, channelBox = False)
elif attr == 'sc':
cmds.setAttr('%s.scaleX' %name, lock = lock, keyable = False, channelBox = False)
cmds.setAttr('%s.scaleY' %name, lock = lock, keyable = False, channelBox = False)
cmds.setAttr('%s.scaleZ' %name, lock = lock, keyable = False, channelBox = False)
elif attr == 'vi':
cmds.setAttr('%s.visibility' %name, lock = True, keyable = False, channelBox = False)
shapeNode = cmds.listRelatives(name)
cmds.setAttr('%s.overrideEnabled' %shapeNode[0], 1)
cmds.setAttr('%s.overrideColor' %shapeNode[0], color)
util.group(name, '%s_grp' %name)
return name
def filter_train_inverses(train_graph, dset):
inverse = {
'_member_of_domain_topic': '_synset_domain_topic_of',
'_member_meronym': '_member_holonym',
'_derivationally_related_form': '_derivationally_related_form',
'_member_of_domain_region': '_synset_domain_region_of',
'_hypernym': '_hyponym',
'_member_holonym': '_member_meronym',
'_instance_hypernym': '_instance_hyponym',
'_synset_domain_topic_of': '_member_of_domain_topic',
'_hyponym': '_hypernym',
'_instance_hyponym': '_instance_hypernym',
'_synset_domain_usage_of': '_member_of_domain_usage',
'_has_part': '_part_of',
'_part_of': '_has_part',
'_synset_domain_region_of': '_member_of_domain_region',
}
def has_train_inverse(pq):
s, r, t = pq.s, pq.r[0], pq.t
if r not in inverse:
return False
try:
return s in train_graph.neighbors[t][inverse[r]]
except KeyError:
return False
grouped = util.group(dset, has_train_inverse)
easy, hard = grouped[True], grouped[False]
return easy, hard
def filter_train_inverses(train_graph, dset):
inverse = {
'_member_of_domain_topic': '_synset_domain_topic_of',
'_member_meronym': '_member_holonym',
'_derivationally_related_form': '_derivationally_related_form',
'_member_of_domain_region': '_synset_domain_region_of',
'_hypernym': '_hyponym',
'_member_holonym': '_member_meronym',
'_instance_hypernym': '_instance_hyponym',
'_synset_domain_topic_of': '_member_of_domain_topic',
'_hyponym': '_hypernym',
'_instance_hyponym': '_instance_hypernym',
'_synset_domain_usage_of': '_member_of_domain_usage',
'_has_part': '_part_of',
'_part_of': '_has_part',
'_synset_domain_region_of': '_member_of_domain_region',
}
def has_train_inverse(pq):
s, r, t = pq.s, pq.r[0], pq.t
if r not in inverse:
return False
try:
return s in train_graph.neighbors[t][inverse[r]]
except KeyError:
return False
grouped = util.group(dset, has_train_inverse)
easy, hard = grouped[True], grouped[False]
return easy, hard
def flattenDependencyList(dependencies):
flatlist = []
for file in dependencies.keys():
if dependencies[file]:
filename = os.path.splitext(os.path.split(file)[1])[0]
deplist = sorted(dependencies[file], key=str.lower)
for depgroup in group(deplist, DEPENDENCIES_PER_LINE):
flatlist.append("$(O)\\%s.obj: %s" % (filename, " ".join(depgroup)))
return flatlist
def flattenDependencyList(dependencies):
flatlist = []
for file in dependencies.keys():
if dependencies[file]:
filename = os.path.splitext(os.path.split(file)[1])[0]
deplist = sorted(dependencies[file], key=str.lower)
for depgroup in group(deplist, DEPENDENCIES_PER_LINE):
flatlist.append("$(O)\\%s.obj: %s" % (filename, " ".join(depgroup)))
return flatlist
def flattenDependencyList(dependencies):
flatlist = []
for file in dependencies.keys():
if dependencies[file]:
opath = getObjectPath(file)
filename = os.path.splitext(os.path.split(file)[1])[0]
# TODO: normalizing paths already in prependPath makes getObjectPath fail under cygwin
deplist = sorted(dependencies[file], key=lambda s: str.lower(s.replace("/", "\\")))
for depgroup in group(deplist, DEPENDENCIES_PER_LINE):
flatlist.append("%s\\%s.obj: %s" % (opath, filename, " ".join(depgroup)))
return flatlist
def flattenDependencyList(dependencies):
flatlist = []
for file in dependencies.keys():
if dependencies[file]:
opath = getObjectPath(file)
filename = os.path.splitext(os.path.split(file)[1])[0]
# TODO: normalizing paths already in prependPath makes getObjectPath fail under cygwin
deplist = sorted(dependencies[file], key=lambda s: str.lower(s.replace("/", "\\")))
for depgroup in group(deplist, DEPENDENCIES_PER_LINE):
flatlist.append("%s\\%s.obj: %s" % (opath, filename, " ".join(depgroup)))
return flatlist
def compute_best_thresholds(examples, debug=False):
# per-relation thresholds
ex_by_rel = util.group(examples, lambda q: q.r[0])
thresholds = {}
for relation, examples_r in util.verboserate(ex_by_rel.items()):
if debug:
print relation
scores = [ex.score for ex in examples_r]
labels = [ex.label for ex in examples_r]
thresholds[relation] = util.best_threshold(scores, labels, debug)
return thresholds
def compute_best_thresholds(examples, debug=False):
# per-relation thresholds
ex_by_rel = util.group(examples, lambda q: q.r[0])
thresholds = {}
for relation, examples_r in util.verboserate(ex_by_rel.items()):
if debug:
print relation
scores = [ex.score for ex in examples_r]
labels = [ex.label for ex in examples_r]
thresholds[relation] = util.best_threshold(scores, labels, debug)
return thresholds
def segmented_evaluation(file_path, categorize=None):
queries = []
with open(file_path, 'r') as f:
for line in util.verboserate(f):
items = line.split('\t')
s, r, t = items[0], tuple(items[1].split(',')), items[2]
q = PathQuery(s, r, t)
quantile_str = items[3]
q.quantile = float(quantile_str)
q.num_candidates = int(items[4])
queries.append(q)
def single_relation(query):
if len(query.r) != 1:
return False
r = query.r[-1]
if inverted(r):
return False
return r
# group queries
if categorize is None:
categorize = single_relation
groups = util.group(queries, categorize)
print 'computing grouped stats'
stats = defaultdict(dict)
for key, queries in util.verboserate(groups.iteritems()):
scores = [q.quantile for q in queries]
score = np.nanmean(scores)
def inv_sigmoid(y):
return -np.log(1.0 / y - 1)
score2 = inv_sigmoid(score)
total = len(scores)
nontrivial = np.count_nonzero(~np.isnan(scores))
stats[key] = {
'score': score,
'score2': score2,
'total_eval': total,
'nontrivial_eval': nontrivial
}
stats.pop(False, None)
return pd.DataFrame(stats).transpose()
def drawCostellationsFigures(self): from reportlab.lib.colors import Color self.c.setStrokeColor(Color( 0.1,0.2, 0.7, alpha=0.5)) self.c.setLineWidth(0.2) figures=catalogues.CostellationFigures() costellations=set(map(lambda x:x[0],figures)) for costellation in costellations: data=filter(lambda x:x[0]==costellation,figures)[0] data=list(util.group(data[2:],2)) for s in data: star1=self.H.search(s[0]) star2=self.H.search(s[1]) if star1!=None and star2!=None: costellation_line=((star1[4]*180/pi,star1[5]*180/pi),(star2[4]*180/pi,star2[5]*180/pi)) self.drawLine(costellation_line)
def multiperson(img, func, mode):
if mode == 'multi':
scales = [2., 1.8, 1.5, 1.3, 1.2, 1., 0.7, 0.5, 0.25]
else:
scales = [1]
height, width = img.shape[0:2] # 582,800； 这里假设height,width中最大的为800，以下的shape都是依据800计算的
center = (width/2, height/2)
dets, tags = None, []
for idx, i in enumerate(scales):
scale = max(height, width)/200
inp_res = int((i * max(512, max(height, width)) + 3)//4 * 4)
res = (inp_res, inp_res) # 800,800
inp = crop(img, center, scale, res)
tmp = func([inp, inp[:,::-1]])
det = tmp['det'][0,:,:] + tmp['det'][1,:,::-1][:,:,flipRef] #det shape [200,200,17]
if idx == 0:
dets = det
mat = get_transform(center, scale, res)[:2] # shape:[2,3] ;mat=[1 0 0;0 1 109],表示没有缩放，只有y轴方向偏移109=(800-582)/2
mat = np.linalg.pinv(np.array(mat).tolist() + [[0,0,1]])[:2] # 计算仿射变换矩阵mat的伪逆
else:
dets = dets + resize(det, dets.shape[0:2])
if abs(i-1)<0.5:
res = dets.shape[0:2] # res shape:[200,200]
tags += [resize(tmp['tag'][0,:,:,:], res), resize(tmp['tag'][1,:,::-1][:,:,flipRef], res)]
tags = np.concatenate([i[:,:,:,None] for i in tags], axis=3) #tags shape:[200,200,17,2]
dets = dets/len(scales)/2
import cv2
cv2.imwrite('det.jpg', (tags.mean(axis=3).mean(axis=2) *255).astype(np.uint8))
grouped = group(dets, tags, 30)
# grouped shape:[30,17,5] 表示检测到30个人，每个人17个关键点，每个关键点前三个数表示x，y，prediction
grouped[:,:,:2] = kpt_affine(grouped[:,:,:2], mat)
# 筛选并整合人体关键点信息
persons = []
for val in grouped: # val为某一个人的关键点信息
if val[:, 2].max()>0: # 某个人的17个关键点中最大的prediction必须大于0
tmp = {"keypoints": [], "score":float(val[:, 2].mean())} # 将17个关键点的平均值作为score分数值
for j in val: # j表示17个关键点中的某一个
if j[2]>0.: # 关键点的prediction必须大于0，否则认为检测错误，记为[0,0,0]
tmp["keypoints"]+=[float(j[0]), float(j[1]), float(j[2])]
else:
tmp["keypoints"]+=[0, 0, 0]
persons.append(tmp)
return persons # 返回满足要求的所有人
def multiperson(img, func, mode):
if mode == 'multi':
scales = [2., 1.8, 1.5, 1.3, 1.2, 1., 0.7, 0.5, 0.25]
else:
scales = [1]
height, width = img.shape[0:2]
center = (width/2, height/2)
dets, tags = None, []
for idx, i in enumerate(scales):
scale = max(height, width)/200
inp_res = int((i * max(512, max(height, width)) + 3)//4 * 4)
res = (inp_res, inp_res)
inp = crop(img, center, scale, res)
tmp = func([inp, inp[:,::-1]])
det = tmp['det'][0,:,:] + tmp['det'][1,:,::-1][:,:,flipRef]
if idx == 0:
dets = det
mat = get_transform(center, scale, res)[:2]
mat = np.linalg.pinv(np.array(mat).tolist() + [[0,0,1]])[:2]
else:
dets = dets + resize(det, dets.shape[0:2])
if abs(i-1)<0.5:
res = dets.shape[0:2]
tags += [resize(tmp['tag'][0,:,:,:], res), resize(tmp['tag'][1,:,::-1][:,:,flipRef], res)]
tags = np.concatenate([i[:,:,:,None] for i in tags], axis=3)
dets = dets/len(scales)/2
#import cv2
#cv2.imwrite('det.jpg', (tags.mean(axis=3).mean(axis=2) *255).astype(np.uint8))
grouped = group(dets, tags, 30)
grouped[:,:,:2] = kpt_affine(grouped[:,:,:2], mat)
persons = []
for val in grouped:
if val[:, 2].max()>0:
tmp = {"keypoints": [], "score":float(val[:, 2].mean())}
for j in val:
if j[2]>0.:
tmp["keypoints"]+=[float(j[0]), float(j[1]), float(j[2])]
else:
tmp["keypoints"]+=[0, 0, 0]
persons.append(tmp)
return persons
def drawCostellationsFigures(self):
from reportlab.lib.colors import Color
self.c.setStrokeColor(Color(0.1, 0.2, 0.7, alpha=0.5))
self.c.setLineWidth(0.2)
figures = catalogues.CostellationFigures()
costellations = set(map(lambda x: x[0], figures))
for costellation in costellations:
data = filter(lambda x: x[0] == costellation, figures)[0]
data = list(util.group(data[2:], 2))
for s in data:
star1 = self.H.search(s[0])
star2 = self.H.search(s[1])
if star1 != None and star2 != None:
costellation_line = ((star1[4] * 180 / pi,
star1[5] * 180 / pi),
(star2[4] * 180 / pi,
star2[5] * 180 / pi))
self.drawLine(costellation_line)
def segmented_evaluation(file_path, categorize=None):
queries = []
with open(file_path, 'r') as f:
for line in util.verboserate(f):
items = line.split('\t')
s, r, t = items[0], tuple(items[1].split(',')), items[2]
q = PathQuery(s, r, t)
quantile_str = items[3]
q.quantile = float(quantile_str)
q.num_candidates = int(items[4])
queries.append(q)
def single_relation(query):
if len(query.r) != 1:
return False
r = query.r[-1]
if inverted(r):
return False
return r
# group queries
if categorize is None:
categorize = single_relation
groups = util.group(queries, categorize)
print 'computing grouped stats'
stats = defaultdict(dict)
for key, queries in util.verboserate(groups.iteritems()):
scores = [q.quantile for q in queries]
score = np.nanmean(scores)
def inv_sigmoid(y):
return -np.log(1.0 / y - 1)
score2 = inv_sigmoid(score)
total = len(scores)
nontrivial = np.count_nonzero(~np.isnan(scores))
stats[key] = {'score': score, 'score2': score2, 'total_eval': total, 'nontrivial_eval': nontrivial}
stats.pop(False, None)
return pd.DataFrame(stats).transpose()
def parseFlopRoundLevel(flopcards):
"Return a value indicating how high the hand ranks"
# counts元组保存每种牌型值的个数
# points元组保存不同牌型值，并且按照大小排序（count值越大越优先）
# Eg. '7 T 7 9 7' => counts=(3,1,1) points=(7,10,9)
groups = group([card.point for card in flopcards])
(counts, points) = unzip(groups)
# 对于顺子(A,2,3,4,5), 规定其值为(1,2,3,4,5)
if points == (14, 5, 4, 3, 2):
points = (5, 4, 3, 2, 1)
# 判断是否为顺子:
# 五张牌数值各不同，同时最大牌与最小牌相差4
straight = (len(points) == 5) and (max(points) - min(points) == 4)
# 判断是否为同花：
# 五张牌花色相同
flush = len(set([card.color for card in flopcards])) == 1
# 这里我们判断9种牌型：同花顺、四条、葫芦、同花、顺子、三条、两对、一对、高牌
level = (9 if straight and flush else 8 if (4, 1) == counts else 7 if
(3, 2) == counts else 6 if flush else 5 if straight else 4 if
(3, 1, 1) == counts else 3 if (2, 2, 1) == counts else 2 if
(2, 1, 1, 1) == counts else 1)
'''
# 打印该五张牌的信息
print 'All five cards information:'
for card in flopcards:
print getColorByIndex(card.color) + '-' + getPointByIndex(card.point)
# 打印该五张牌的牌型有多少种大小
print 'Points Count: ', len(points)
# 计算牌型价值
'''
value = computeCardsValue(level, points)
print 'Cards Value: ', value
return value, level
def gen_c_code(langs_ex, strings_dict, file_name, lang_index):
# This is just to make the order the same as the old code that was parsing
# just one translation file, to avoid a diff in generated c code when switching
# from the old to the new method
langs = [cols[0] for cols in lang_index]
assert DEFAULT_LANG == langs[0]
langs_count = len(langs)
translations_count = len(strings_dict)
langs_c = [c_escape(tmp) for tmp in langs]
langs_c = ",\n ".join([", ".join(ten) for ten in group(langs_c, 10)])
keys = strings_dict.keys()
keys.sort(cmp=key_sort_func)
lines = []
for lang in langs:
if DEFAULT_LANG == lang:
trans = keys
else:
trans = get_trans_for_lang(strings_dict, keys, lang)
lines.append("")
lines.append(" /* Translations for language %s */" % lang)
lines += [" %s," % c_escape(t) for t in trans]
translations = "\n".join(lines)
langs_ex.sort(lang_sort_func)
lang_ids = make_lang_ids(langs_ex, lang_index)
lang_layouts = make_lang_layouts(lang_index)
lang_names = [
'{ "%s", %s, %s, %d },' %
(lang[0], c_escape(lang[1]), lang_ids[lang[0]], lang_layouts[lang[0]])
for lang in langs_ex
]
lang_names = "\n ".join(lang_names)
file_content = TRANSLATIONS_TXT_C % locals()
file(file_name, "wb").write(file_content)
def createFastSelector(fullList, nameList, funcName, type): cases = ["case %s:" % value for (name, value) in fullList if name in nameList] return unTab(Template_Selector % (funcName, type, "\n ".join([" ".join(part) for part in group(cases, 4)])))
def createTypeEnum(list, type, default): list = sorted(list, key=lambda a: a[0]) parts = group([item[1] for item in list] + [default], 5) return unTab(Template_Enumeration % (type, ",\n ".join([", ".join(part) for part in parts])))
def readTLE(self,url):
url=url
f=urllib.urlopen(url)
data=f.read().split('\r\n')
s=util.group(data,3)
return s
def createFastSelector(fullList, nameList, funcName, type):
cases = [
"case %s:" % value for (name, value) in fullList if name in nameList
]
return unTab(Template_Selector % (funcName, type, "\n ".join(
[" ".join(part) for part in group(cases, 4)])))
def createTypeEnum(list, type, default):
list = sorted(list, key=lambda a: a[0])
parts = group([item[1] for item in list] + [default], 5)
return unTab(Template_Enumeration %
(type, ",\n ".join([", ".join(part) for part in parts])))
