def compute(p, op, e, target, problem, story, order, score=None, cons=None):
if op == '=':
vec = [order, score, cons]
vec.extend(makesets.vector(p, e, problem, story, target))
op_label, op_acc, op_val = svmutils.svm_predict(
[-1], [vec], glob, '-q -b 1'
)
else:
vec = makesets.vector(p, e, problem, story, target)
op_label, op_acc, op_val = svmutils.svm_predict(
[-1], [vec], multi, '-q -b 1'
)
op_val = op_val[0]
if op == '+':
val = op_val[0]
if op == '-':
val = op_val[1]
if op == '*':
val = op_val[2]
if op == '/':
val = op_val[3]
if op == '=':
val = op_val[0]
c = makesets.combine(p[1], e[1], op)
return (val, c, op_val)
def training(trips,problem,target):
#this function take the trips and creates positive and negative training instances from them
texamples = {x:([],[]) for x in ["+","*",'/','-','=']}
for op,a,b in trips:
vec = makesets.vector(a,b,problem,target)
texamples[op][0].append(vec)
return texamples
def training(a, b, problem, story, target, j, order, score, constraints):
#this function take the trips and creates positive and negative training instances from them
if j == 0:
j = -1
vec = [j, order, score, constraints]
vec.extend(makesets.vector(a, b, problem, story, target))
return vec
def training(trips, problem, story, target):
# this function take the trips and creates positive and negative training instances from them
texamples = {x: ([], []) for x in ["+", "*", "/", "-", "="]}
for op, a, b in trips:
if op == "=":
vec = makesets.eqvector(a, b, problem, story, target)
else:
vec = makesets.vector(a, b, problem, story, target)
texamples[op][0].append(vec)
return texamples
def training(trips,problem,target):
#this function take the trips and creates positive and negative training instances from them
texamples = {x:([],[]) for x in ["+","*"]}
for op,a,b in trips:
vec = makesets.vector(a,b,problem,target)
for k in texamples:
if op == k:
texamples[k][0].append(vec)
else:
texamples[k][1].append(vec)
return texamples
def compute(p, op, e, target, problem, story, order):
vec = makesets.vector(p, e, problem, story, target)
op_label, op_acc, op_val = svmutil.svm_predict([-1], [vec], multi,
'-q -b 1')
op_val = op_val[0]
if op == '+':
val = op_val[0]
if op == '-':
val = op_val[1]
if op == '*':
val = op_val[2]
if op == '/':
val = op_val[3]
if op == '=':
val = op_val[1]
c = makesets.combine(p[1], e[1], op)
return (val, c, op_val)
def compute(p, op, e, target, problem):
vec = makesets.vector((0, p), (0, e), problem, target)
#if p.ent == e.ent and op in ['*','/']:
# val = 0
#else:
if True:
op_label, op_acc, op_val = svm_predict([-1], [vec], multi, '-q -b 1')
#pmop_label, pmop_acc, pmop_val = svm_predict([-1], [vec], pm ,'-q -b 1')
#mmop_label, mmop_acc, mmop_val = svm_predict([-1], [vec], md ,'-q -b 1')
op_val = op_val[0]
#pmop_val = pmop_val[0]
#mmop_val = mmop_val[0]
if op == '+':
val = op_val[0] #*pmop_val[0]
if op == '-':
val = op_val[1] #*pmop_val[1]
if op == '*':
val = op_val[2] #*mmop_val[0]
if op == '/':
val = op_val[3] #*mmop_val[1]
c = makesets.combine(p, e, op)
return (val, c)
def compute(p,op,e,target,problem):
vec = makesets.vector((0,p),(0,e),problem,target)
#if p.ent == e.ent and op in ['*','/']:
# val = 0
#else:
if True:
op_label, op_acc, op_val = svm_predict([-1], [vec], model ,'-q -b 1')
#sop_label, sop_acc, sop_val = svm_predict([-1], [vec], smodel ,'-q -b 1')
#print(op_label,op_acc,op_val)
op_val=op_val[0]
#sop_val=sop_val[0]
#op_val = [op_val[0]*sop_val[0],op_val[1]*sop_val[0],op_val[2]*sop_val[1],op_val[3]*sop_val[1]]
if op == '+':
val = op_val[0]
if op == '-':
val = op_val[1]
if op == '*':
val = op_val[2]
if op == '/':
val = op_val[3]
c = makesets.combine(p,e,op)
return (val,c)
def compute(p,op,e,target,problem):
vec = makesets.vector((0,p),(0,e),problem,target)
#if p.ent == e.ent and op in ['*','/']:
# val = 0
#else:
if True:
op_label, op_acc, op_val = svm_predict([-1], [vec], multi ,'-q -b 1')
#pmop_label, pmop_acc, pmop_val = svm_predict([-1], [vec], pm ,'-q -b 1')
#mmop_label, mmop_acc, mmop_val = svm_predict([-1], [vec], md ,'-q -b 1')
op_val=op_val[0]
#pmop_val = pmop_val[0]
#mmop_val = mmop_val[0]
if op == '+':
val = op_val[0]#*pmop_val[0]
if op == '-':
val = op_val[1]#*pmop_val[1]
if op == '*':
val = op_val[2]#*mmop_val[0]
if op == '/':
val = op_val[3]#*mmop_val[1]
c = makesets.combine(p,e,op)
return (val,c)
def training(a,b,problem,story,target,j,order,score,constraints,eq):
#this function take the trips and creates positive and negative training instances from them
if j == 0:
j=-1
vec = [j,order,score,constraints]
vec.extend(makesets.vector(a,b,problem,story,target))
ops = [x for x in eq if x in ['+','-','*','/']]
if len(ops)==1:
op = ops[0]
vec.append(int(op=='+'))
vec.append(int(op=='-'))
vec.append(int(op=='*'))
vec.append(int(op=='/'))
else:
vec.append(0)
vec.append(0)
vec.append(0)
vec.append(0)
if len(ops)>1:
vec.append(int('+' in ops))
vec.append(int('-' in ops))
vec.append(int('*' in ops))
vec.append(int('/' in ops))
else:
vec.append(0)
vec.append(0)
vec.append(0)
vec.append(0)
eq = eq.split(' = ')
if 'x' in eq[0]:
vec.append(1)
else:
vec.append(0)
return vec
