def classify_encapsulated(audio_summary, track_info, pickle_file):
"""Performs a similar operation to classify() and all the above functions in one step
But retrieves fit from pickle, for the GUI
"""
f = open(pickle_file, 'r')
fit = pickle.load(f)
f.close()
rv = {}
#print track_info.keys()
#print track_info[audio_summary]['title'].keys()
rv["speechiness"] = [audio_summary['audio_summary']['speechiness']]
rv["time_sig"] = [audio_summary['audio_summary']['time_signature']]
rv["bpm"] = [audio_summary['audio_summary']['tempo']]
rv["key"] = [audio_summary['audio_summary']['key']]
rv["duration"] = [audio_summary['audio_summary']['duration']]
rv["loudness"] = [audio_summary['audio_summary']['loudness']]
rv["end_of_fade_in"] = [track_info['track']['end_of_fade_in']]
rv["start_of_fade_out"] = [track_info['track']['start_of_fade_out']]
rv["bpm_range"] = [proc.bpm_range(track_info['beats'])]
rv["max_bpm_spike"] = [proc.max_bpm_spike(track_info['beats'])]
try:
rv["num_keys"] = [proc.num_keys(track_info['sections'])]
except:
rv["num_keys"] = [1]
rv["sections"] = [proc.num_sections(track_info['sections'])]
new_df = build_data_frame([rv],["Unknown"])
p = prob_category(new_df,fit)
robjects.globalenv["pred"] = p
edm_prob = robjects.default_ri2py(p.rx(1))[0]
folk_prob = robjects.default_ri2py(p.rx(2))[0]
rap_prob = robjects.default_ri2py(p.rx(3))[0]
cls = classify(new_df,fit)
return [(edm_prob,folk_prob,rap_prob),cls[0]]
def custom_ri2py(obj):
res = robjects.default_ri2py(obj)
if isinstance(res, robjects.Vector) and (len(res) == 1):
res = super(robjects.vectors.Vector, res).__getitem__(0)
if isinstance(res, rinterface.Sexp):
res = robjects.default_ri2py(res)
return res
def get_sd_ratios(what,orgn,dstn,fltnum,dfrom,dto,daysprior,cls):
curr_dt = datetime.now()
curr_date = date(curr_dt.year, curr_dt.month, curr_dt.day)
if what == 'pb':
[vals,obs,diff,dptdts] = get_fcsterr_by_dptdt_dflc_lvl('uncons',orgn,dstn,fltnum,dfrom,curr_date,daysprior,cls)
elif what == 'fb':
print "fp parameter what isn't implemented"
assert 0
elif what == 'ns':
vals = get_noshows_by_flc_lvl(orgn,dstn,fltnum,dfrom,dto,cls)
else:
print 'Unknown parameter what in get_sd_ratios'
assert 0
print vals
try:
vals_r = robjects.FloatVector(vals)
res_r = r.find_var_prop(vals_r)
res = robjects.default_ri2py(res_r)
except:
res = [0,0]
return res
def ks_p_value(l):
r_l = robjects.IntVector(l)
r_p_value = r.get_ks_p_value(r_l)
p_value = robjects.default_ri2py(r_p_value)[0]
return p_value
def cramer_von_mises_p_value(l):
r_l = robjects.IntVector(l)
r_p_value = r.get_cr_von_mis_p_value(r_l)
p_value = robjects.default_ri2py(r_p_value)[0]
return p_value
def as_dendropy_object(o, taxon_set=None):
"""
Returns a DendroPy object corresponding to the ape object ``o``. If ``o`` is
a single tree (i.e., ``phylo``), then a DendroPy Tree is returned. If ``o`` is
a list of trees (i.e., a ``multiPhylo`` object, or list of ``phylo`` objects),
then a DendroPy TreeList is returned.
"""
if o.rclass[0] == "multiPhylo":
f = tempfile.NamedTemporaryFile()
_R["write.nexus"](o, file=f.name)
return dendropy.TreeList.get_from_path(f.name, "nexus", taxon_set=taxon_set)
elif o.rclass[0] == "phylo":
f = tempfile.NamedTemporaryFile()
_R["write.nexus"](o, file=f.name)
return dendropy.Tree.get_from_path(f.name, "nexus", taxon_set=taxon_set)
elif o.rclass[0] == "list":
f = tempfile.NamedTemporaryFile()
_R["write.nexus.data"](o, file=f.name)
# print open(f.name, "r").read()
d = dendropy.DataSet.get_from_path(f.name, "nexus", taxon_set=taxon_set)
if len(d.char_matrices) == 0:
raise ValueError("No character data found")
elif len(d.char_matrices) == 1:
return d.char_matrices[0]
else:
raise ValueError("Multiple character matrices returned")
else:
return robjects.default_ri2py(o)
def ri2pandas(o):
if isinstance(o, DataFrame):
raise NotImplementedError(
"Conversion from rpy2 DataFrame to pandas' DataFrame")
else:
res = ro.default_ri2py(o)
return res
def as_dendropy_object(o, taxon_set=None):
"""
Returns a DendroPy object corresponding to the ape object ``o``. If ``o`` is
a single tree (i.e., ``phylo``), then a DendroPy Tree is returned. If ``o`` is
a list of trees (i.e., a ``multiPhylo`` object, or list of ``phylo`` objects),
then a DendroPy TreeList is returned.
"""
if o.rclass[0] == "multiPhylo":
f = tempfile.NamedTemporaryFile()
_R['write.nexus'](o, file=f.name)
return dendropy.TreeList.get_from_path(f.name,
"nexus",
taxon_set=taxon_set)
elif o.rclass[0] == "phylo":
f = tempfile.NamedTemporaryFile()
_R['write.nexus'](o, file=f.name)
return dendropy.Tree.get_from_path(f.name,
"nexus",
taxon_set=taxon_set)
elif o.rclass[0] == "list":
f = tempfile.NamedTemporaryFile()
_R['write.nexus.data'](o, file=f.name)
# print open(f.name, "r").read()
d = dendropy.DataSet.get_from_path(f.name,
"nexus",
taxon_set=taxon_set)
if len(d.char_matrices) == 0:
raise ValueError("No character data found")
elif len(d.char_matrices) == 1:
return d.char_matrices[0]
else:
raise ValueError("Multiple character matrices returned")
else:
return robjects.default_ri2py(o)
def get_fcsterr_mse(type, orgn, dstn, fltnum, dfrom, dto, daysprior, cls):
[fcst, obs, diff, dptdt_l] = get_fcsterr_by_dptdt_dflc_lvl(type, orgn, dstn, fltnum, dfrom, dto, daysprior, cls)
r_diff = robjects.FloatVector(diff)
r_mse = r.get_mse(r_diff)
mse = robjects.default_ri2py(r_mse)[0]
return mse
def ri2numpy(o):
if isinstance(o, ListVector):
res = numpy.rec.fromarrays(o, names=tuple(o.names))
elif isinstance(o, Vector) and (type(o) != Vector):
res = numpy.asarray(o)
else:
res = ro.default_ri2py(o)
return res
def f(obj):
pyobj = robjects.default_ri2py(obj)
inherits = rinterface.baseenv["inherits"]
classname = rinterface.SexpVector(["density", ],
rinterface.STRSXP)
if inherits(pyobj, classname)[0]:
pyobj = Density(pyobj)
return pyobj
def ri2numpy(o):
if isinstance(o, ListVector):
res = numpy.rec.fromarrays(o, names=tuple(o.names))
elif isinstance(o, Vector) and (type(o) != Vector):
res = numpy.asarray(o)
else:
res = ro.default_ri2py(o)
return res
def testMapperR2Python_s4(self):
robjects.r('setClass("A", representation(x="integer"))')
classname = rinterface.StrSexpVector(["A", ])
one = rinterface.IntSexpVector([1, ])
sexp = rinterface.globalenv.get("new")(classname,
x=one)
self.assertTrue(isinstance(robjects.default_ri2py(sexp),
robjects.RS4))
def shapiro_wilk_p_value(l):
# l = list(Set(l))
r_l = robjects.FloatVector(l)
r_p_value = r.get_shapiro_wilk_p_value(r_l)
p_value = robjects.default_ri2py(r_p_value)[0]
return p_value
def f(obj):
pyobj = robjects.default_ri2py(obj)
inherits = rinterface.baseenv["inherits"]
classname = rinterface.SexpVector([
"density",
], rinterface.STRSXP)
if inherits(pyobj, classname)[0]:
pyobj = Density(pyobj)
return pyobj
def convert_xts_to_df(o):
"""
Will convert xts objects to DataFrame
"""
dates = o.do_slot('index')
index = convert_posixct_to_index(dates)
res = robjects.default_ri2py(o)
df = _convert_Matrix(res)
df.index = index
return df
def convert_xts_to_df(o):
"""
Will convert xts objects to DataFrame
"""
dates = o.do_slot('index')
dates = np.array(dates, dtype=np.dtype("M8[s]"))
res = robjects.default_ri2py(o)
df = rcom.convert_robj(res)
df.index = dates
return df
def get_ljung_box_pval(type, orgn, dstn, fltnum, dfrom, dto, daysprior, cls):
[fcst, obs, diff, dptdt_l] = get_fcsterr_by_dptdt_dflc_lvl(type, orgn, dstn, fltnum, dfrom, dto, daysprior, cls)
if len(diff) != 0:
r_diff = robjects.FloatVector(diff)
r_pval = r.get_ljung_box_pval(r_diff)
pval = robjects.default_ri2py(r_pval)[0]
else:
pval = 0
return pval
def convert_xts_to_df(o):
"""
Will convert xts objects to DataFrame
"""
dates = o.do_slot('index')
dates = np.array(dates, dtype=np.dtype("M8[s]"))
res = robjects.default_ri2py(o)
df = rcom.convert_robj(res)
df.index = dates
return df
def testMapperR2Python_s4(self):
robjects.r('setClass("A", representation(x="integer"))')
classname = rinterface.StrSexpVector([
"A",
])
one = rinterface.IntSexpVector([
1,
])
sexp = rinterface.globalenv.get("new")(classname, x=one)
self.assertTrue(isinstance(robjects.default_ri2py(sexp), robjects.RS4))
def my_ri2py(o):
res = None
try:
rcls = o.do_slot("class")
except LookupError:
rcls = [None]
if isinstance(o, ri.SexpVector):
if 'xts' in rcls:
res = convert_xts_to_df(o)
if res is None:
res = robjects.default_ri2py(o)
return res
def get_esp_params(flights,dows,sh_id,pool_ids,clss):
sizes = []
alphas = []
for flight in flights:
orgn = flight[0]
dstn = flight[1]
fltnum = flight[2]
sh = splitHistory(sh_id)
for dow in dows:
for pool_id in pool_ids:
for cls in clss:
for daysprior in get_dayspriors():
fdr = fcstDataReader(orgn,dstn,fltnum,\
cls,daysprior,dow,sh)
vals = fdr.get_fcst_minus_booked_vals('uncons',pool_id)
sample_size = get_sample_size(vals)
if sample_size is not None:
sizes.append(sample_size)
vals_r = robjects.FloatVector(vals)
vals_r_ts = r.ts(vals_r,frequency=len(vals))
hw_res = r.HoltWinters(vals_r,beta=False,gamma=False)
alpha = robjects.default_ri2py(hw_res[2])
alphas.append(alpha)
lag = int(ceil(mean(sizes)))
alpha = float(mean(alphas))
return [lag, alpha]
def ri2py_list(obj):
if 'data.frame' in obj.rclass:
# R "factor" vectors will not convert well by default
# (will become integers), so we build a temporary list o2
# with the factors as strings.
o2 = list()
# An added complication is that the conversion defined
# in this module will make __getitem__ at the robjects
# level return numpy arrays
for column in rinterface.ListSexpVector(obj):
if 'factor' in column.rclass:
levels = tuple(column.do_slot("levels"))
column = tuple(levels[x-1] for x in column)
o2.append(column)
names = obj.do_slot('names')
if names is NULL:
res = numpy.rec.fromarrays(o2)
else:
res = numpy.rec.fromarrays(o2, names=tuple(names))
else:
# not a data.frame, yet is it still possible to convert it
res = ro.default_ri2py(obj)
return res
def ri2py_list(obj):
if 'data.frame' in obj.rclass:
# R "factor" vectors will not convert well by default
# (will become integers), so we build a temporary list o2
# with the factors as strings.
o2 = list()
# An added complication is that the conversion defined
# in this module will make __getitem__ at the robjects
# level return numpy arrays
for column in rinterface.ListSexpVector(obj):
if 'factor' in column.rclass:
levels = tuple(column.do_slot("levels"))
column = tuple(levels[x-1] for x in column)
o2.append(column)
names = obj.do_slot('names')
if names is NULL:
res = numpy.rec.fromarrays(o2)
else:
res = numpy.rec.fromarrays(o2, names=tuple(names))
else:
# not a data.frame, yet is it still possible to convert it
res = ro.default_ri2py(obj)
return res
def run_test(self):
''' take the R globalEnv located HyperGParam obj and run
HyperGTest, returns the HyperGResult obj '''
return rob.default_ri2py(rob.r.hyperGTest(rob.r[self.par_name]))
def pd_ri2py(o):
res = None
try:
rcls = o.do_slot("class")
except LookupError, le:
rcls = [None]
if isinstance(o, SexpVector):
if 'xts' in rcls:
print 'xts'
res = convert_xts_to_df(o)
if 'POSIXct' in rcls:
res = convert_posixct_to_index(o)
if res is None:
res = robjects.default_ri2py(o)
return res
def convert_xts_to_df(o):
"""
Will convert xts objects to DataFrame
"""
dates = o.do_slot('index')
dates = np.array(dates, dtype=np.dtype("M8[s]"))
res = robjects.default_ri2py(o)
df = rcom.convert_robj(res)
df.index = dates
return df
def convert_posixct_to_index(o):
baseenv_ri = rinterface.baseenv
globalenv_ri = rinterface.globalenv
def my_ri2py(o):
res = None
try:
rcls = o.do_slot("class")
except LookupError, le:
rcls = [None]
if isinstance(o, SexpVector):
if 'xts' in rcls:
res = convert_xts_to_df(o)
if res is None:
res = robjects.default_ri2py(o)
return res
def convert_xts_to_df(o):
"""
Will convert xts objects to DataFrame
"""
dates = o.do_slot('index')
dates = np.array(dates, dtype=np.dtype("M8[s]"))
res = robjects.default_ri2py(o)
df = rcom.convert_robj(res)
df.index = dates
return df
robjects.conversion.ri2py = my_ri2py
def ri2py_sexp(obj):
if (obj.typeof in _vectortypes) and (obj.typeof != VECSXP):
res = numpy.asarray(obj)
else:
res = ro.default_ri2py(obj)
return res
def run_test(self): ''' take the R globalEnv located HyperGParam obj and run HyperGTest, returns the HyperGResult obj ''' return rob.default_ri2py(rob.r.hyperGTest(rob.r[self.par_name]))
def testMapperR2Python_string(self):
sexp = rinterface.globalEnv.get("letters")
ob = robjects.default_ri2py(sexp)
self.assertTrue(isinstance(ob,
robjects.RVector))
def testMapperR2Python_environment(self):
sexp = rinterface.globalenv.get(".GlobalEnv")
self.assertTrue(isinstance(robjects.default_ri2py(sexp),
robjects.Environment))
def testMapperR2Python_function(self):
sexp = rinterface.globalenv.get("plot")
ob = robjects.default_ri2py(sexp)
self.assertTrue(isinstance(ob,
robjects.Function))
def testMapperR2Python_boolean(self):
sexp = rinterface.globalenv.get("T")
ob = robjects.default_ri2py(sexp)
self.assertTrue(isinstance(ob,
robjects.Vector))
def ri2py_sexp(obj):
if (obj.typeof in _vectortypes) and (obj.typeof != VECSXP):
res = numpy.asarray(obj)
else:
res = ro.default_ri2py(obj)
return res
def ri2pandas(o):
if isinstance(o, DataFrame):
raise NotImplementedError("Conversion from rpy2 DataFrame to pandas' DataFrame")
else:
res = ro.default_ri2py(o)
return res
def testMapperR2Python_boolean(self):
sexp = rinterface.globalenv.get("T")
ob = robjects.default_ri2py(sexp)
self.assertTrue(isinstance(ob, robjects.Vector))
def testMapperR2Python_environment(self):
sexp = rinterface.globalenv.get(".GlobalEnv")
self.assertTrue(
isinstance(robjects.default_ri2py(sexp), robjects.Environment))
def testMapperR2Python_function(self):
sexp = rinterface.globalenv.get("plot")
ob = robjects.default_ri2py(sexp)
self.assertTrue(isinstance(ob, robjects.Function))