def run_sim(rootv, mname, params, n, minsz=0.0):
simvs = []
simvs.append({"size": rootv, "mult_factor": 0.0, "depth": 0.0})
rgen = rng.Rangen(mname, params)
rvs = rgen.rvs(n)
r = rng.get_random()
for i in xrange(n):
l = len(simvs)
fchosen = r.randint(0, l - 1)
s = simvs[fchosen]
mf = rvs[i]
ns = max(s["size"] * mf, minsz)
nmf = s["mult_factor"] * mf
ndepth = s["depth"] + 1
nsv = {"size": ns, "mult_factor": nmf, "depth": ndepth}
simvs.append(nsv)
fsz = []
mfs = []
dp = []
for simv in simvs:
fsz.append(simv["size"])
mfs.append(simv["mult_factor"])
dp.append(simv["depth"])
return (np.array(fsz), np.array(mfs), np.array(dp))
def run_sim(rootv, mname, params, n, minsz = 0.0):
simvs = []
simvs.append({"size": rootv, "mult_factor": 0.0, "depth": 0.0})
rgen = rng.Rangen(mname, params)
rvs = rgen.rvs(n)
r = rng.get_random()
for i in xrange(n):
l = len(simvs)
fchosen = r.randint(0, l-1)
s = simvs[fchosen]
mf = rvs[i]
ns = max(s["size"]*mf, minsz)
nmf = s["mult_factor"]*mf
ndepth = s["depth"] + 1
nsv = {"size": ns, "mult_factor": nmf, "depth": ndepth}
simvs.append(nsv)
fsz=[]
mfs=[]
dp=[]
for simv in simvs:
fsz.append(simv["size"])
mfs.append(simv["mult_factor"])
dp.append(simv["depth"])
return (np.array(fsz), np.array(mfs), np.array(dp))
def get_root_from_data(data_file):
"""
Get a random root value from the points specified by the data file
"""
fp_df = os.path.join(os.environ["TRACE_DB_LOC"], data_file)
vals = util.read_data(fp_df)
n = len(vals)
r = rng.get_random()
i_rootv = r.randint(0, n - 1)
rootv = vals[i_rootv]
return rootv
def get_root_from_data(data_file): """ Get a random root value from the points specified by the data file """ fp_df = os.path.join(os.environ["TRACE_DB_LOC"], data_file) vals = util.read_data(fp_df) n = len(vals) r = rng.get_random() i_rootv = r.randint(0, n-1) rootv = vals[i_rootv] return rootv
def recursive_model(**kwargs):
"""
Recursive forest file model simulator
Params:
file = data file to pick initial roots
nroots = number of roots
d1 = model for roots
d1params = (param1, param2,...) parameters for d1
d2 = model for children nodes
d2params = (param1, param2, ...) parameters for d2
g = probability of a new file
nu = probability of a deletion
n = number of iterations
minsize = minimum file size
"""
sroots = None
d1f = None
d2f = None
g = None
nu = None
minsize = 0.0
nroots = kwargs["nroots"]
if "minsize" in kwargs:
minsize = kwargs["minsize"]
if "file" in kwargs:
fname = kwargs["file"]
fl = util.read_data(fname)
d1f = DFile(flist=fl, minsize=0.0)
else:
d1model = kwargs["d1"]
d1params = kwargs["d1params"]
d1f = DFile(d=d1model, plist=d1params, minsize=0.0)
d2model = kwargs["d2"]
d2params = kwargs["d2params"]
d2f = DFile(d=d2model, plist=d2params)
gf = rng.get_random() # random.Random()
nf = rng.get_random() # random.Random()
fpick = rng.get_random() # random.Random()
g = kwargs["g"]
nu = kwargs["nu"]
n = kwargs["n"]
sroots = d1f.random(nroots)
simvs = []
for sroot in sroots:
s = {"size": sroot, "deleted": False, "mult_factor": 0.0, "depth": 0}
simvs.append(s)
for i in xrange(nroots, n+1):
gvar = gf.random()
if gvar <= g:
a = d1f.random(1)
sa = {"size": a, "deleted": False, "mult_factor": 0.0, "depth": 0}
simvs.append(sa)
else:
idx = fpick.randint(0, len(simvs)-1)
nvar = nf.random()
if nvar <= nu:
simvs.pop(idx)
else:
mf = d2f.random(1)
idx = fpick.randint(0, len(simvs)-1)
pick = simvs[idx]
ns = max(pick["size"]*mf[0], minsize)
ndel = False
nmf = pick["mult_factor"] * mf[0]
ndepth = pick["depth"] + 1
simvs.append({"size": ns, "deleted": ndel, "mult_factor": nmf, "depth": ndepth})
sizev=[]
mfv=[]
depthv=[]
for simv in simvs:
sizev.append(simv["size"])
mfv.append(simv["mult_factor"])
depthv.append(simv["depth"])
return (np.array(sizev), np.array(mfv), np.array(depthv))