def __init__(self, *args, **kwargs):
od = OrdDict()
for item in args:
od[None] = conversion.py2ro(item)
for k, v in kwargs.items():
od[k] = conversion.py2ro(v)
res = self._constructor.rcall(tuple(od.items()), robjects.globalenv)
self.__sexp__ = res.__sexp__
def new(cls, *args, **kwargs):
""" Constructor for the class Layer. """
for i, elt in enumerate(args):
args[i] = conversion.py2ro(elt)
for k in kwargs:
kwargs[k] = conversion.py2ro(kwargs[k])
res = cls(cls.contructor)(*args, **kwargs)
return res
def new(cls,
*args, **kwargs):
""" Constructor for the class Layer. """
for i, elt in enumerate(args):
args[i] = conversion.py2ro(elt)
for k in kwargs:
kwargs[k] = conversion.py2ro(kwargs[k])
res = cls(cls.contructor)(*args, **kwargs)
return res
def assign(self, index, value):
""" Assign a given value to a given index position in the vector """
if not (isinstance(index, rlc.TaggedList) | \
isinstance(index, rlc.ArgsDict)):
args = rlc.TaggedList([conversion.py2ro(index), ])
else:
for i in xrange(len(index)):
index[i] = conversion.py2ro(index[i])
args = index
args.append(conversion.py2ro(value))
args.insert(0, self)
res = r["[<-"].rcall(args.items())
res = conversion.ri2py(res)
return res
def new(cls, geom, geom_params, stat, stat_params, data, aesthetics,
position, params):
args = [
conversion.py2ro(x) for x in (geom, geom_params, stat, stat_params,
data, aesthetics, position, params)
]
res = cls(cls.contructor)(*args)
return res
def subset(self, *args, **kwargs):
""" Subset the "R-way.", using R's "[" function.
In a nutshell, R indexing differs from Python's on:
- indexing can be done with integers or strings (that are 'names')
- an index equal to TRUE will mean everything selected (because of the recycling rule)
- integer indexing starts at one
- negative integer indexing means exclusion of the given integers
- an index is itself a vector of elements to select
"""
args = [conversion.py2ro(x) for x in args]
for k, v in kwargs.itervalues():
args[k] = conversion.py2ro(v)
res = r["["](*([self, ] + [x for x in args]), **kwargs)
return res
def new(cls,
geom, geom_params,
stat, stat_params,
data,
aesthetics,
position,
params):
args = [conversion.py2ro(x) for x in (geom, geom_params,
stat, stat_params,
data,
aesthetics,
position,
params)]
res = cls(cls.contructor)(*args)
return res
def __setitem__(self, item, value):
robj = conversion.py2ro(value)
super(REnvironment, self).__setitem__(item, robj)
def setdim(self, value):
value = conversion.py2ro(value)
res = r["dim<-"](self, value)
#FIXME: not properly done
raise(Exception("Not yet implemented"))
for cell in all_cells_clusters23:
total_transcript_counts = cell.total_estimated_counts - cell.spikeins.est_counts.sum(
)
est_counts = cell.est_counts.transpose() #*total_transcript_counts/1e3
counts_data.loc[:, cell.id] = np.floor(est_counts)
factors = pd.DataFrame(np.zeros((1, number_of_cells), dtype='int16'),
columns=list_of_cell_names)
fact = ''
for cell in all_cells_clusters23:
factors.loc[:, cell.id] = str(cell.clusterID)
fact += str(cell.clusterID)
scde = importr("scde")
counts_data_r = conversion.py2ro(counts_data)
factors_r = conversion.py2ro(factors)
fact_r = ro.FactorVector(fact)
factors_r.colnames = list_of_cell_names
counts_data_r.colnames = list_of_cell_names
r("""counts_data_int = apply(""" + counts_data_r.r_repr() +
""",2, function(x) {storage.mode(x) = 'integer';x}) """)
r("""facts = """ + fact_r.r_repr())
r("""names(facts) = colnames(counts_data_int)""")
# r("o.ifm = scde.error.models(counts = counts_data_int, groups = facts, n.cores = 1, linear.fit=F, local.theta.fit=F, threshold.segmentation = TRUE, save.crossfit.plots = FALSE, save.model.plots = FALSE, verbose = 1)")
# r("""save(o.ifm, file = "/scratch/PI/mcovert/dvanva/sequencing/75min_scde_fit_linear.RData")""")
output = open("DMBC_Prediction.csv", "w")
training = pd.read_csv("training.csv")
test = pd.read_csv("test.csv")
training2 = pd.DataFrame.to_csv(training,
sep=",",
index=False,
line_terminator='\n')
test2 = pd.DataFrame.to_csv(test, sep=",", index=False, line_terminator='\n')
print(training2)
print(test2)
r_training = conversion.py2ro(training2)
robjects.r.assign("my_training", r_training)
robjects.r("save(my_training, file='training.RData')")
r_test = conversion.py2ro(test2)
robjects.r.assign("my_test", r_test)
robjects.r("save(my_test, file='test.RData')")
x = robjects.r('''
library(DMBC)
load("training.RData")
training_time <- get("training")
load("test.RData")
test_time <- get("test")
auc_out <- Cal_AUC(loocv(training_time))
dmbc_predict(data=training_time,testSet=test_time,auc_out=auc_out)
