def assign_noz(f, nofz, **kwargs):
"""read file 'f', substitute a columns with noz(z), with z in 'f' itself, and
save in a file.
Parameters
----------
f: file object or string
file containing the catalogue
nofz: function
returns n(z) for all the z in the input file
output
------
none
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
cat = np.loadtxt(f) #read the input catalogue
cat[:, kwargs['nz_column']] = nofz(cat[:, kwargs['z_column']])
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
return None
def subtract_fromfile(f, absmin, **kwargs):
"""read file 'f', subtract absmin+offset from the first three columns and
save the file
Parameters
----------
f: file object or string
file containing the catalogue
absmin: list or array of three floats
values to subtract from the first 3 columns
output
------
none
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+offset: extra offset to add to absmin
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
cat = np.loadtxt(f) #read the input catalogue
cat[:, :3] -= absmin + kwargs['offset'] #subtract the minimum
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
return None
def rebin(fname, mbins, **kwargs):
"""
Rebin the power spectrum in file 'fn' and save it into the output file
Parameters
----------
fname: string
file name
mbins: int
number of bins to merge to create the input
accepted kwargs that affects the function:
+from_bin: skip the first 'from' bins
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(fname,
**kwargs) # create the output file name
if ofile == None:
return None
header_lines, ps = read_file(fname) #read the file
# if required to skip some line
fbin = kwargs.get('from_bin')
if fbin is not None:
ps = ps[fbin:, :]
# check that the number of remaining bins is multiple of mbins and if not
# cut the last bins
pssize = ps.shape[0]
psremaining = pssize % mbins
if psremaining != 0:
ps = ps[:-psremaining, :]
# separate the power spectrum into the various components
k, pk, pksn, nmodes = ps.T
# reshape to simplify rebinning
finalbins = ps.shape[0] // mbins
kr = k.reshape([finalbins, mbins])
pkr = pk.reshape([finalbins, mbins])
pksnr = pksn.reshape([finalbins, mbins])
nmodesr = nmodes.reshape([finalbins, mbins])
# get the values for the rebinned power spectrum
kr = np.average(kr, axis=1)
pkr = np.average(pkr, axis=1, weights=nmodesr)
pksnr = np.average(pksnr, axis=1, weights=nmodesr)
nmodesr = nmodesr.sum(axis=1)
# save the rebinned power spectrum
with open(ofile, 'w') as of:
of.writelines(header_lines)
np.savetxt(of,
np.vstack([kr, pkr, pksnr, nmodesr]).T,
delimiter='\t',
fmt=kwargs['fmt'])
def move_columns(f, from_columns, to_columns, **kwargs):
"""
Read file 'f', substitute the content of columns 'to_cols' with the one of
'from_cols'
Parameters
----------
f: file object or string
file containing the catalogue
from_columns: list of ints
list of columns to copy
to_columns: list of ints
list of columns where to copy
output
------
none
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+swap: swap the columns instead of just moving
+pandas: use pandas for the input
+chunks: chunksize in pandas.read_table
+substitute: value to substitute in 'from_columns' that are not in 'to_columns'
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
if kwargs['verbose']:
print("Processing file '{}'".format(f))
#get the columns to move or swap
substitute = None
temp_from_cols, temp_to_cols = from_columns[:], to_columns[:]
if kwargs['swap']: # swap the colums
temp_from_cols.extend(to_columns)
temp_to_cols.extend(from_columns)
else:
if kwargs['substitute'] is not None: #find the columns that need substitution
substitute = [i for i in temp_from_cols if i not in temp_to_cols]
if len(substitute) == 0:
substitute = None #set back to None if there are no columns with value to substitute
if kwargs['pandas']:
use_pandas(f, ofile, temp_from_cols, temp_to_cols, substitute, **kwargs)
else:
cat = np.loadtxt(f)
cat[:,temp_to_cols] = cat[:,temp_from_cols]
if substitute is not None:
cat[:, substitute] = kwargs['substitute']
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
def cselect(f, constraint, **kwargs):
"""read file 'f', substitute a columns with noz(z), with z in 'f' itself, and
save in a file.
Parameters
----------
f: file object or string
file containing the catalogue
constr: string
selection criterion
output
------
none
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+pandas: use pandas for the input
+chunks: chunksize in pandas.read_table
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
if ofile is None: #skip existing files if "skip" option used
return None
if kwargs['verbose']:
print("Processing file '{}'".format(f))
if kwargs['pandas']:
if kwargs['chunks'] is None:
cat = pd.read_table(f,
header=None,
skiprows=mf.n_lines_comments(f),
sep='\s')
cat = cat[eval(constraint)]
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
else:
chunks = pd.read_table(f,
header=None,
sep='\s',
skiprows=mf.n_lines_comments(f),
chunksize=kwargs['chunks'])
with open(ofile, 'w') as of:
for cat in chunks:
cat = cat[eval(constraint)]
np.savetxt(of, cat, fmt=kwargs['fmt'], delimiter='\t')
else:
cat = np.loadtxt(f)
cat = cat[eval(constraint), :]
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
def hist_save_return(f, readcols, hdata, **kwargs):
"""
Read the required 'col' from file 'f', do the histogram
save it to file and return the it
{common_params}
output
------
{return_values}
{common_kwargs}
{kwargs_file}
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
hist, bin_centers, bin_edges = hist_return(f, readcols, hdata, **kwargs)
# create the output 2d array
outhist = np.vstack([bin_centers, bin_edges[:-1], bin_edges[1:], hist]).T
np.savetxt(ofile, outhist, fmt=kwargs['fmt'], delimiter='\t')
return hist, bin_centers, bin_edges
def convert_save(f, distance, **kwargs ):
"""
Read file *f*, converts ra, dec, z into cartesian coordinates, computing the
comoving distance at redshift z stored in *distance*, and save to a new file
Parameters
----------
f: file object or string
file containing ra, dec and z
distance: function
function that evaluates the comoving distance at given redshift(s)
kwargs: keyword arguments
output
------
max, min: lists
maximum and minimum values of x, y and z
If kwargs['skip'] == True and the output file name already exists, a *None*
is returned
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+pandas: use pandas for the input
+chunks: chunksize in pandas.read_table
+usecols: columns to read from the input files. the first three must be ra,
dec and redshift
+negative_z: check or not for negative redshifts and perform action [None, 'skip', 'tozero']
+fmt: format of the output file
"""
if kwargs['verbose']:
print("Processing file '{}'".format(f))
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
if kwargs['pandas']:
return use_pandas(f, ofile, distance, **kwargs)
else:
return use_numpy(f, ofile, distance, **kwargs)
def hist_ndensity_save_return(f, readcols, hdata, volume_z, **kwargs):
"""
Read the required 'col' from file 'f', do the histogram of the number
density, save it to file and return it
{common_params}
volume_z: list
effective volume in each bin of the histogram.
Warning: kwargs['range'] cannot be *None* and
kwargs['nbins'] == len(volume_z). No check performed
output
------
{return_values}
{common_kwargs}
{kwargs_file}
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
hist, bin_centers, bin_edges = hist_ndensity_return(
f, readcols, hdata, volume_z, **kwargs)
# create the output 2d array
outhist = np.vstack([bin_centers, bin_edges[:-1], bin_edges[1:], hist]).T
np.savetxt(ofile, outhist, fmt=kwargs['fmt'], delimiter='\t')
return hist, bin_centers, bin_edges
def columns_operations(f, operations, to_column, **kwargs):
"""
read file, performe the desired operations between columns, save it and
write to a file.
Signature:
columns_operations("file.dat", 'c3+c6-1', 4)
#read file, add content of column 3 and column 6, subtract 1 and save
the result in column 4.
Parameters
----------
f: file object or string
file containing the catalogue
operations: string
columns with operations to perform
to_columns int
save there the result of operations
output
------
none
accepted kwargs that affects the function
+substitute: substitute the content of the columns involved with operations
with this value, if not None
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+pandas: use pandas for the input
+chunks: chunksize in pandas.read_table
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
if kwargs['verbose']:
print("Processing file '{}'".format(f))
pattern = re.compile(r"c(\d+?)") #re pattern with the columns name
if kwargs['pandas']: #expression to evaluate to execute the operation
to_evaluate = pattern.sub("cat[\\1]", operations)
else:
to_evaluate = pattern.sub("cat[:,\\1]", operations)
#columns used in the operation
columns_read = [int(s) for s in pattern.findall(operations)]
# read the input catalogue
if kwargs['pandas']:
if kwargs['chunks'] is None:
cat = pd.read_table(f,
header=None,
skiprows=mf.n_lines_comments(f),
sep='\s')
new_column = eval(to_evaluate) # do the operation
if kwargs['substitute'] is not None:
cat[columns_read] = kwargs['substitute']
cat[to_column] = new_column #copy the result of the operation
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
else:
chunks = pd.read_table(f,
header=None,
sep='\s',
skiprows=mf.n_lines_comments(f),
chunksize=kwargs['chunks'])
with open(ofile, 'w') as of:
for cat in chunks:
new_column = eval(to_evaluate) # do the operation
if kwargs['substitute'] is not None:
cat[columns_read] = kwargs['substitute']
cat[to_column] = new_column #copy the result of the operation
np.savetxt(of, cat, fmt=kwargs['fmt'], delimiter='\t')
else:
cat = np.loadtxt(f)
new_column = eval(to_evaluate) # do the operation
if kwargs['substitute'] is not None:
cat[:, columns_read] = kwargs['substitute']
cat[:, to_column] = new_column #copy the result of the operation
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
def downsample_noz(f, interpf, zcol, **kwargs):
"""
read the input files, downsample it and save it
Parameters
----------
f: file object or string
file containing the catalogue
interpf: callable
function that returns the fraction used for the downsample
zcol: int
column in the input file to use for the downsampling
output
------
accepted kwargs that affects the function
+substitute: substitute the content of the columns involved with operations
with this value, if not None
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+pandas: use pandas for the input
+chunks: chunksize in pandas.read_table
+fmt: format of the output file
"""
ofile = mf.create_ofile_name(f, **kwargs) # create the output file name
if kwargs['verbose']:
print("Processing file '{}'".format(f))
# read the input catalogue
if kwargs['pandas']:
if kwargs['chunks'] is None:
cat = pd.read_table(f,
header=None,
skiprows=mf.n_lines_comments(f),
sep='\s')
to_keep = cat[zcol] # save the column to use for the selection
to_keep = np.random.rand(to_keep.size) <= interpf(
to_keep) #random select objects to keep
np.savetxt(ofile, cat[to_keep], fmt=kwargs['fmt'], delimiter='\t')
else:
chunks = pd.read_table(f,
header=None,
sep='\s',
skiprows=mf.n_lines_comments(f),
chunksize=kwargs['chunks'])
with open(ofile, 'w') as of:
for cat in chunks:
to_keep = cat[
zcol] # save the column to use for the selection
to_keep = np.random.rand(to_keep.size) <= interpf(
to_keep) #random select objects to keep
np.savetxt(of,
cat[to_keep],
fmt=kwargs['fmt'],
delimiter='\t')
else:
cat = np.loadtxt(f)
to_keep = cat[:, zcol] # save the column to use for the selection
to_keep = np.random.rand(to_keep.size) <= interpf(
to_keep) #random select objects to keep
np.savetxt(ofile, cat[to_keep, :], fmt=kwargs['fmt'], delimiter='\t')
def substitute_from_file(fname, col, replacementa, **kwargs):
"""
Read file 'fname' and replace 'column' matching 'replacement_arr[:,0]' with
'replacement_arr[:,1]'
Parameters
----------
fname: file object or string
file containing the catalogue
col: integer
column to use
replacementa: 1D or 2D array
function that execute the replacement
output
------
none
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+pandas: use pandas for the input
+chunks: chunksize in pandas.read_table
+fmt: format of the output file
"""
def replace(x, repl_array):
"""
if 'repl_array' is a 1D array, return repl_array[x]
if 'repl_array is a Nx2 array, returns repl_array[:,1] where x matches repl_array[:,0]
"""
if repl_array.ndim == 1:
if repl_array.dtype != int:
print("Warning: 'x' will be casted to integer")
return repl_array[x.astype(np.int)]
else:
if x.dtype == int and repl_array[:, 1].dtype == float:
_x = x.astype(np.float)
else:
_x = x.copy()
for (r0, r1) in repl_array:
_x[x == r0] = r1
return _x
if kwargs['verbose']:
print("Processing file '{}'".format(fname))
ofile = mf.create_ofile_name(fname,
**kwargs) # create the output file name
if kwargs['pandas']:
if kwargs['chunks'] is None:
cat = pd.read_table(fname,
header=None,
skiprows=mf.n_lines_comments(fname),
sep='\s')
cat[col] = replace(cat[col], replacementa)
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
else:
chunks = pd.read_table(fname,
header=None,
sep='\s',
skiprows=mf.n_lines_comments(fname),
chunksize=kwargs['chunks'])
with open(ofile, 'w') as fo: #open the output file
for cat in chunks: #loop over the chunks
cat[col] = replace(cat[col], replacementa)
np.savetxt(fo, cat, fmt=kwargs['fmt'], delimiter='\t')
else:
cat = np.loadtxt(fname)
cat[:, col] = replace(cat[:, col], replacementa)
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
def fits2ascii(fname, selected_columns, operations=None, **kwargs):
"""
Read some columns a fits file 'fname' and save the output into an ascii table.
Operations between columns permitted. The output file has the order given
in columns
Signature:
fits2ascii( "file.fits", 0 ) #read and save the given column
fits2ascii( "file.fits", [0,4,5,3] ) #read and save the given columns (in that order)
fits2ascii( "file.fits", ['a', 'b', 'n', 'c'] ) #as before but with column name
fits2ascii( "file.fits", ['a', 'b+n-1', 'c'] ) #save a 3 columns ascii table with 'a' in the first,
the result of the fitsio.read_column('b')+fitsio.read_column('n')-1 in the second and 'c' in the third
Parameters
----------
fname: string
file containing the catalogue
selected_columns: integer, string, list of integers or strings.
columns to read. If strings, operation permitted: in this case
the *name* of the column, *not* the number, must be given
operations: regular expression pattern
pattern containing the list of accepted operators. If None no operation
done
output
------
none
accepted kwargs that affects the function
+verbose: verbose mode [True|False]
+replace: replace string *replace[0]* with *replace[1]* in f.name
+insert: insert string *insert[0]* before *insert[1]* in f.name
+skip: existing file names skipped [True|False]
+overwrite: existing file names overwritten [True|False]
+fmt: format of the output file
"""
if kwargs['verbose']:
print("extracting columns '{0}' from file '{1}'".format(
selected_columns, fname))
ofile = mf.create_ofile_name(fname,
**kwargs) # create the output file name
# check if there are operations to execute
check_operations = []
if operations is None:
check_operations = False
else:
for c in selected_columns:
try:
n_matches = len(operations.findall(c))
except TypeError: # The number of the column given
check_operations.append(False)
else: #If a string is given
if n_matches > 0:
check_operations.append(True)
else:
check_operations.append(False)
# If there are no operations to perform, open the file, read the columns
# and save them
if sum(check_operations) == 0:
cat = fitsio.read(fname, columns=selected_columns)[selected_columns]
# If there are operations involved
else:
with fitsio.FITS(fname) as fits:
cat = read_with_operations(fits[1], selected_columns, operations)
np.savetxt(ofile, cat, fmt=kwargs['fmt'], delimiter='\t')
