def match_job(jobID):
try:
job = Job.objects.get(id=jobID)
required_skills = list(itertools.chain(*job.requiredjobskill_set.values_list('skill__name')))
optional_skills = list(itertools.chain(*job.optionaljobskill_set.values_list('skill__name')))
requirements = {'years': int(job.years_of_experience),
'skills': required_skills,
'optional': optional_skills}
candidates = JobSeeker.objects.filter(category=job.category)
for jobseeker in candidates:
skills = list(itertools.chain(*JobSeekerSkill.objects.filter(
seeker=jobseeker).values_list('skill__name')))
candidate = {'name': jobseeker.user.get_full_name(), 'skills': skills,
'years': int(jobseeker.years_of_experience)}
matched, score, weight = matcher(requirements, candidate)
if matched:
try:
jobmatch = JobMatch(job=job, seeker=jobseeker, score=score,
weight=weight)
jobmatch.save()
except IntegrityError, e: # Match already exists
logging.exception(e)
pass
except Exception, e:
logging.exception(e)
return False
def matchE(request):
lucene.getVMEnv().attachCurrentThread()
try:
student = {}
student['name'] = request.POST['student_name']
student['interest'] = \
request.POST['student_interest']
student['affiliation'] = \
request.POST['student_affiliation']
except KeyError:
return render_to_response('index.html', {'error_msg': 'missing field'},
context_instance=RequestContext(request))
else:
prof_matcher = matcher()
prof_list = prof_matcher.getProfMatch(student)
request.session['prof_list'] = prof_list
request.session['student'] = student
info_list = []
for i, prof in enumerate(prof_list):
score, explainList = prof_matcher.explainPos(i + 1)
info_list.append((prof, score, explainList))
for prof in prof_list:
aff_count = prof['affiliation'].count(student['affiliation'])
prof['co_count'] = aff_count
student = request.session.get('student')
return render_to_response('explain.html', {
'info_list': info_list,
'student': student
})
def matchChamp():
# getFrame("therainman")
# getFrame("guardsmanbob")
# getFrame("ms_yuyu")
# getFrame("robertxlee")
M = matcher()
print(M.matchChamp('file3.png'))
def matchE(request):
lucene.getVMEnv().attachCurrentThread()
try:
student = {}
student['name'] = request.POST['student_name']
student['interest'] = \
request.POST['student_interest']
student['affiliation'] = \
request.POST['student_affiliation']
except KeyError:
return render_to_response('index.html',
{'error_msg':'missing field'},
context_instance=RequestContext(request))
else:
prof_matcher = matcher()
prof_list = prof_matcher.getProfMatch(student)
request.session['prof_list'] = prof_list
request.session['student'] = student
info_list = []
for i,prof in enumerate(prof_list):
score,explainList = prof_matcher.explainPos(i+1)
info_list.append((prof,score,explainList))
for prof in prof_list:
print prof['name']
aff_count = prof['affiliation'].count(student['affiliation'])
prof['co_count'] = aff_count
student = request.session.get('student')
print 'in match', student, prof_list[0].get('name')
return render_to_response('explain.html', {'info_list':info_list,'student':student})
def match_candidate(jobseekerID):
try:
jobseeker = JobSeeker.objects.get(id=jobseekerID)
skills = list(itertools.chain(*jobseeker.jobseekerskill_set.all().values_list('skill__name')))
candidate = {'name': jobseeker.user.get_full_name(), 'skills': skills,
'years': int(jobseeker.years_of_experience)}
# For new candidates, do a match for jobs added in the last 30 days only
date = timezone.now() - timedelta(days=30)
jobs = Job.objects.filter(category=jobseeker.category, created_at__gte=date)
for job in jobs:
required_skills = list(itertools.chain(
*job.requiredjobskill_set.values_list('skill__name')))
optional_skills = list(itertools.chain(
*job.optionaljobskill_set.values_list('skill__name')))
requirements = {'years': int(job.years_of_experience),
'skills': required_skills,
'optional': optional_skills}
matched, score, weight = matcher(requirements, candidate)
if matched:
try:
jobmatch = JobMatch(job=job, seeker=jobseeker, score=score,
weight=weight)
jobmatch.save()
except IntegrityError, e: # Match already exists
logging.exception(e)
pass
except Exception, e:
logging.exception(e)
return False
def __init__(self, options):
self.resultes = {
'classes': {},
'methods': [],
}
self.sources = options['sources']
self.verbose = options['verbose']
self.target = os.path.abspath('./intermediate/result.json')
self.class_file_types = ['.h']
self.matcher = matcher.matcher()
def __init__(self, options):
self.sources = options['sources']
self.verbose = options['verbose']
self.scan_results_path = './intermediate/result.json'
self.scan_results = {}
self.file_types = ['.h', '.m', '.mm', '.pbxproj', '.pch']
if 'rule' in options:
self.mapper = mapper.mapper(rule_path=options['rule'])
else:
self.mapper = mapper.mapper()
self.matcher = matcher.matcher()
def clicked_go():
if (selectedDir is ""):
msgBox = QtGui.QMessageBox()
msgBox.setText("Please select output folder.")
msgBox.setWindowTitle(" ")
msgBox.exec_()
return
sortOrder = []
errorFunctions = {}
for key in classifiers:
if (classifiers[key].checkBox.isChecked()):
sortOrder.append(str(classifiers[key].label.text()))
try:
errorFunctions[key] = str(classifiers[key].lineEdit.text())
except ValueError:
try:
errorFunctions[key] = matcher.ERROR_FUNCTIONS[key]
except KeyError:
pass
print sortOrder
print errorFunctions
print("Matching...")
matcher.matcher(sortOrder=sortOrder,errorFunctions=errorFunctions,outDir=selectedDir)
print("Done!")
def __init__(self, args):
#Salvam argumentul de intrare in program (path-ul catre directorul cu poze si setam algoritm surf/sift
self.path = args
self.siftOrSurf = 'surf'
#Citim imaginile din director, le facem un resize si le memoram in images
self.images = [cv2.resize(cv2.imread(file), (700, 610)) for file in glob.glob(self.path + "*.jpg")]
#Salvam numarul de imagini
self.count = len(self.images)
print("Imagini gasite: ", self.count)
#Initializare de array-uri
self.leftList, self.rightList = [], []
#Initializam un obiect de tip matcher caruia ii pasam tipul de algoritm definit mai sus
self.matcher_obj = matcher(siftOrSurf=self.siftOrSurf)
def initialize_variables(CONS, phi, level):
"""Resample target transform and initial velocity
initialize other objects for scale level"""
# container to hold all variables
VARS = {}
# smooth and downsample the images
aaSmoother = smoother.smoother(1 + level, 0, 1, 2, CONS['spacing'],
CONS['grid'], CONS['dtype'])
fix_smooth = np.copy(aaSmoother.smooth(CONS['fixed']))
mov_smooth = np.copy(aaSmoother.smooth(CONS['moving']))
VARS['fixed'] = zoom(fix_smooth, 1. / 2**level, mode='wrap')
VARS['moving'] = zoom(mov_smooth, 1. / 2**level, mode='wrap')
# initialize a few odds and ends
shape = VARS['fixed'].shape
VARS['spacing'] = CONS['spacing'] * 2**level
VARS['warped_transform'] = np.empty(shape + (len(shape), ),
dtype=CONS['dtype'])
# initialize or resample the deformation
if phi is None:
VARS['phi'] = np.zeros(shape + (len(shape), ), dtype=CONS['dtype'])
else:
zoom_factor = np.array(shape) / np.array(phi.shape[:-1])
phi_ = [
zoom(phi[..., i], zoom_factor, mode='nearest') for i in range(3)
]
VARS['phi'] = np.ascontiguousarray(np.moveaxis(np.array(phi_), 0, -1))
# initialize the transformer
VARS['transformer'] = transformer.transformer(shape, VARS['spacing'],
CONS['dtype'])
VARS['transformer'].set_initial_transform(CONS['initial_transform'])
# initialize the smoothers
VARS['field_smoother'] = smoother.smoother(
CONS['field_abcd'][0] * 2**level, *CONS['field_abcd'][1:],
VARS['spacing'], shape, CONS['dtype'])
VARS['grad_smoother'] = smoother.smoother(CONS['grad_abcd'][0] * 2**level,
*CONS['grad_abcd'][1:],
VARS['spacing'], shape,
CONS['dtype'])
# initialize the matcher
VARS['matcher'] = matcher.matcher(VARS['fixed'], VARS['moving'],
CONS['lcc_radius'])
return VARS
def main():
categorys = ['102803_ctg1_4188_-_ctg1_4188','102803_ctg1_2088_-_ctg1_2088','102803_ctg1_5988_-_ctg1_5988','102803_ctg1_5088_-_ctg1_5088','102803_ctg1_1288_-_ctg1_1288','102803_ctg1_4288_-_ctg1_4288',
'102803_ctg1_4688_-_ctg1_4688','102803_ctg1_2488_-_ctg1_2488','102803_ctg1_3288_-_ctg1_3288','102803_ctg1_5288_-_ctg1_5288','102803_ctg1_5188_-_ctg1_5188','102803_ctg1_1388_-_ctg1_1388',
'102803_ctg1_4788_-_ctg1_4788','102803_ctg1_2188_-_ctg1_2188','102803_ctg1_6088_-_ctg1_6088','102803_ctg1_1199_-_ctg1_1199','102803_ctg1_2288_-_ctg1_2288','102803_ctg1_4988_-_ctg1_4988',
'102803_ctg1_1988_-_ctg1_1988','102803_ctg1_4388_-_ctg1_4388','102803_ctg1_5788_-_ctg1_5788','102803_ctg1_4888_-_ctg1_4888','102803_ctg1_2588_-_ctg1_2588','102803_ctg1_3188_-_ctg1_3188',
'102803_ctg1_1488_-_ctg1_1488','102803_ctg1_2688_-_ctg1_2688','102803_ctg1_5588_-_ctg1_5588','102803_ctg1_5888_-_ctg1_5888','102803_ctg1_1688_-_ctg1_1688','102803_ctg1_4588_-_ctg1_4588',
'102803_ctg1_5388_-_ctg1_5388','102803_ctg1_5488_-_ctg1_5488','102803_ctg1_4488_-_ctg1_4488','102803_ctg1_1588_-_ctg1_1588','102803_ctg1_2388_-_ctg1_2388','102803_ctg1_5688_-_ctg1_5688',
'102803_ctg1_6399_-_ctg1_6399','102803_ctg1_2788_-_ctg1_2788']
categorys = ['102803_ctg1_1199_-_ctg1_1199'] #需要修改
categorys = ['102803_ctg1_2288_-_ctg1_2288','102803_ctg1_4988_-_ctg1_4988',
'102803_ctg1_1988_-_ctg1_1988','102803_ctg1_4388_-_ctg1_4388','102803_ctg1_5788_-_ctg1_5788','102803_ctg1_4888_-_ctg1_4888','102803_ctg1_2588_-_ctg1_2588','102803_ctg1_3188_-_ctg1_3188',
'102803_ctg1_1488_-_ctg1_1488','102803_ctg1_2688_-_ctg1_2688','102803_ctg1_5588_-_ctg1_5588','102803_ctg1_5888_-_ctg1_5888','102803_ctg1_1688_-_ctg1_1688','102803_ctg1_4588_-_ctg1_4588',
'102803_ctg1_5388_-_ctg1_5388','102803_ctg1_5488_-_ctg1_5488','102803_ctg1_4488_-_ctg1_4488','102803_ctg1_1588_-_ctg1_1588','102803_ctg1_2388_-_ctg1_2388','102803_ctg1_5688_-_ctg1_5688',
'102803_ctg1_6399_-_ctg1_6399','102803_ctg1_2788_-_ctg1_2788']
categorys = ['102803_ctg1_5688_-_ctg1_5688']
username = '******'
pwd = 'nan18756072542'
WBLogin = weiboLogin.weiboLogin()
if WBLogin.login(username,pwd)==1:
print 'Login success!'
for category in categorys:
i = 1
while True:
print u'正在获取第' + str(i) + '页内容、、、'
page01 = getWeiboPage.getWeiboPage(category,i).get_firstpage()
if page01 == 0:
break
else:
#matcher.matcher(page01).pageAnalyse()
matcher.matcher(page01).insertContents()
time.sleep(5)
page02 = getWeiboPage.getWeiboPage(category,i).get_secondpage()
if page02 == 0:
break
else:
#matcher.matcher(page02).pageAnalyse()
matcher.matcher(page02).insertContents()
time.sleep(10)
page03 = getWeiboPage.getWeiboPage(category,i).get_thirdpage()
if page03 == 0:
break
else:
#matcher.matcher(page03).pageAnalyse()
matcher.matcher(page03).insertContents()
time.sleep(30)
i += 1
time.sleep(60)
else:
print 'Login error!'
exit()
def main(args):
data_path = data_dir = os.path.dirname(__file__) + "./data/" + args.data
result_path = data_dir = os.path.dirname(__file__) + "./data/result"
#----------------------------------------------
#
# step 1: preprocess data
# (1) load data
# (2) get feature
# (3) find knn for each feature
#----------------------------------------------
processer_obj = preprocesser(data_path)
#----------------------------------------------
#
# step 2: find relation between all images
# (1) construct candidate map
# (2) compute homography matrix
#----------------------------------------------
matcher_obj = matcher(processer_obj.img_pts_dict, args.candidate,
args.lowe_ratio, args.ransac_th)
#----------------------------------------------
#
# step 3: stich all images and make panorama
# (1) select best candidate based on given img
# (2) stich each by each
# (3) apply ROI on panorama image
#----------------------------------------------
stitcher_obj = sticher(matcher_obj, args.roi_improve)
pano_img, roi_pano_img = stitcher_obj.img_pano(processer_obj.images, 0)
pano_img = pano_img.astype(np.uint8)
roi_pano_img = roi_pano_img.astype(np.uint8)
plt.figure()
plt.subplot(2, 1, 1)
plt.imshow(pano_img)
plt.subplot(2, 1, 2)
plt.imshow(roi_pano_img)
plt.show()
cv2.imwrite(result_path + "/" + args.data + "raw.png", pano_img[..., ::-1])
cv2.imwrite(result_path + "/" + args.data + "roi.png",
roi_pano_img[..., ::-1])
def match(request):
if request.method == "POST":
form = QueryForm(request.POST)
if form.is_valid():
name = form.cleaned_data['name']
interest = form.cleaned_data['interest']
affiliation = form.cleaned_data['affiliation']
try:
result = Result.objects.get(stuname=name,
stuaffiliation=affiliation,
stuinterest=interest)
except Result.DoesNotExist:
lucene.getVMEnv().attachCurrentThread()
student = {}
student['name'] = name
student['interest'] =interest
student['affiliation'] = affiliation
algo_id = 1
prof_matcher = matcher()
if random.randint(1,1000) & 1:
boost = {'interest':1.0, 'processed_aff':2.0}
else:
boost = {'interest':2.0, 'processed_aff':1.0}
algo_id = 2
try:
prof_result = prof_matcher.getProfMatch(student,
fieldList = ["interest","processed_aff"],
boosts=boost)
except:
messages.error(request, "Sorry, but I can't recognize your query.")
return render_to_response('index.html',
{'form': form},
context_instance=
RequestContext(request))
if not prof_result:
messages.error(request, "Can't found "
"enough experts to show to you, please adjust "
"the query.")
return render_to_response('index.html',
{'form': form},
context_instance=
RequestContext(request))
prof_list = []
for result in prof_result:
name = result['name']
interest = result['interest']
print name, interest
professor = Professor.objects.get(name__icontains=name.split(' ')[0],
interest=interest)
prof_list.append(professor.id)
print prof_list
result = Result(stuinterest=student['interest'],
stuname=student['name'], stuaffiliation=
student['affiliation'], date=datetime.now(),
pos1id=prof_list[0], pos2id=prof_list[1],
pos3id=prof_list[2],
algoid=Algo.objects.get(pk=algo_id))
result.save()
request.session['result_id'] = result.id
return HttpResponseRedirect(reverse('kdd_matcher:results'))
else:
return render_to_response('index.html',
{'form': form},
context_instance
=RequestContext(request))
def match(request):
if request.method == "POST":
form = QueryForm(request.POST)
if form.is_valid():
name = form.cleaned_data['name']
interest = form.cleaned_data['interest']
affiliation = form.cleaned_data['affiliation']
try:
result = Result.objects.get(stuname=name,
stuaffiliation=affiliation,
stuinterest=interest)
except Result.DoesNotExist:
lucene.getVMEnv().attachCurrentThread()
student = {}
student['name'] = name
student['interest'] = interest
student['affiliation'] = affiliation
algo_id = 1
prof_matcher = matcher()
if random.randint(1, 1000) & 1:
boost = {'interest': 1.0, 'processed_aff': 2.0}
else:
boost = {'interest': 2.0, 'processed_aff': 1.0}
algo_id = 2
try:
prof_result = prof_matcher.getProfMatch(
student,
fieldList=["interest", "processed_aff"],
boosts=boost)
except:
messages.error(request,
"Sorry, but I can't recognize your query.")
return render_to_response(
'index.html', {'form': form},
context_instance=RequestContext(request))
if not prof_result:
messages.error(
request, "Can't found "
"enough experts to show to you, please adjust "
"the query.")
return render_to_response(
'index.html', {'form': form},
context_instance=RequestContext(request))
prof_list = []
for result in prof_result:
name = result['name']
print name
interest = result['interest']
print interest
try:
professor = Professor.objects.get(
name__icontains=name.split(' ')[0],
interest=interest)
except:
try:
professor = Professor.objects.get(
name__icontains=name.split(' ')[-1],
interest=interest)
except:
continue
prof_list.append(professor.id)
if len(prof_list) < 3:
messages.error(
request, "Can't found "
"enough experts to show to you, please adjust "
"the query.")
return render_to_response(
'index.html', {'form': form},
context_instance=RequestContext(request))
result = Result(stuinterest=student['interest'],
stuname=student['name'],
stuaffiliation=student['affiliation'],
date=datetime.now(),
pos1id=prof_list[0],
pos2id=prof_list[1],
pos3id=prof_list[2],
algoid=Algo.objects.get(pk=algo_id))
result.save()
request.session['result_id'] = result.id
return HttpResponseRedirect(reverse('kdd_matcher:results'))
else:
return render_to_response('index.html', {'form': form},
context_instance=RequestContext(request))
def __init__(self):
self.matcher = matcher()
self.understanding = understanding()
import editdistance
import core
import analyze
import matcher
import sys
sample_size = 1000
print('Loading catalogs...')
amzn = core.amazon_catalog()
goog = core.google_catalog()
print('Performing compare all match (edit distance)...')
compare_all_edit_match = matcher.matcher(amzn,goog,editdistance.eval, matcher.all)
print('Compare All Matcher (Edit Distance) Performance: ' + str(core.eval_matching(compare_all_edit_match)))
print('Performing compare all match (jaccard distance)...')
compare_all_jaccard_match = matcher.matcher(amzn,goog,analyze.jaccard_calc, matcher.all)
print('Compare All Matcher (Jaccard Distance) Performance: ' + str(core.eval_matching(compare_all_jaccard_match)))
print('Performing random sample match (edit distance)...')
compare_all_edit_match = matcher.matcher(amzn,goog,editdistance.eval, matcher.random_sample, sample_size)
print('Random Sample Matcher (Edit Distance) Performance: ' + str(core.eval_matching(compare_all_edit_match)))
print('Performing random sample match (jaccard distance)...')
compare_all_jaccard_match = matcher.matcher(amzn,goog,analyze.jaccard_calc, matcher.random_sample, sample_size)
print('Random Sample Matcher (Jaccard Distance) Performance: ' + str(core.eval_matching(compare_all_jaccard_match)))
def register(args):
CONS = initialize_constants(args)
level = len(CONS['iterations']) - 1
print(args)
print(args, file=CONS['log'])
# record initial energy
# TODO: include initial transform in this energy calculation
ff, mm, rad = CONS['fixed'], CONS['moving'], CONS['lcc_radius']
mat = matcher.matcher(ff, mm, rad)
energy = mat.lcc(ff, mm, rad)
message = 'initial energy: ' + str(energy)
print(message)
print(message, file=CONS['log'])
# multiscale loop
start_time = time.clock()
lowest_phi = 0
for local_iterations in CONS['iterations']:
# initialize level
phi_ = None if level == len(CONS['iterations']) - 1 else lowest_phi
VARS = initialize_variables(CONS, phi_, level)
iteration, backstep_count, converged = 0, 0, False
local_step = CONS['gradient_step']
lowest_energy = 0
# loop for current level
while iteration < local_iterations and not converged:
t0 = time.clock()
# compute the residual
warped = VARS['transformer'].apply_transform(
VARS['moving'],
VARS['spacing'],
VARS['phi'],
initial_transform=True) # should check args.initial_transform
energy, residual = VARS['matcher'].lcc_grad(
VARS['fixed'], warped, CONS['lcc_radius'], VARS['spacing'])
residual = VARS['grad_smoother'].smooth(residual)
max_residual = np.linalg.norm(residual, axis=-1).max()
residual *= VARS['spacing'].min() / max_residual
# monitor the optimization
if energy > (1 - CONS['tolerance']) * lowest_energy:
VARS['phi'] = np.copy(lowest_phi)
local_step *= 0.5
backstep_count += 1
iteration -= 1
VARS['field_smoother'] = smoother.smoother(
CONS['field_abcd'][0] * 2**level / 4**backstep_count,
*CONS['field_abcd'][1:], VARS['spacing'],
VARS['fixed'].shape, CONS['dtype'])
if backstep_count >= max(local_iterations // 10, 5):
converged = True
else:
if energy < lowest_energy:
lowest_energy, lowest_phi = energy, np.copy(VARS['phi'])
backstep_count = max(0, backstep_count - 1)
# the gradient descent update
residual *= -local_step
for i in range(3):
VARS['warped_transform'][
..., i] = VARS['transformer'].apply_transform(
VARS['phi'][..., i], VARS['spacing'], residual)
VARS['phi'] = VARS['warped_transform'] + residual
VARS['phi'] = VARS['field_smoother'].smooth(VARS['phi'])
iteration += 1
# record progress
message = 'it: ' + str(iteration) + \
', en: ' + str(energy) + \
', time: ' + str(time.clock() - t0) + \
', bsc: ' + str(backstep_count)
print(message)
print(message, file=CONS['log'])
level -= 1
message = 'total optimization time: ' + str(time.clock() - start_time)
print(message)
print(message, file=CONS['log'])
if args.final_lcc is not None or \
args.warped_image is not None:
warped = VARS['transformer'].apply_transform(
CONS['moving'],
CONS['spacing'],
lowest_phi,
initial_transform=True) # should check args.initial_transform
# write the warped image
if args.warped_image is not None:
inout.write_image(warped, args.warped_image)
# write the final lcc
if args.final_lcc is not None:
final_lcc = VARS['matcher'].lcc(CONS['fixed'],
warped,
CONS['lcc_radius'],
mean=False)
inout.write_image(final_lcc, args.final_lcc)
# write the deformation field
output = lowest_phi
if args.compose_output_with_it:
output += VARS['transformer'].Xit - VARS['transformer'].X
inout.write_image(output, args.output)