def write_zenodo_files(args, baseoutdir):
for study, dset in [v.split('/') for v in args.varvals]:
print '%-10s %15s %s' % (study, dset, baseoutdir)
metafos = heads.read_metadata(study)
for method in args.methods:
outdir = get_outdir(args, baseoutdir, varname='data', varval='zenodo/%s/%s/%s' % (study_translations.get(study, study), dset, method.replace('-default', '')))
print ' %-15s' % method
write_single_zenodo_subdir(outdir, args, study, dset, method, metafos[dset])
def get_data_plots(args, baseoutdir, methods, study, dsets):
metafos = heads.read_metadata(study)
assert len(set([metafos[ds]['locus'] for ds in dsets
])) # make sure everybody has the same locus
mfo = metafos[dsets[0]]
data_outdirs = [
heads.get_datadir(
study, 'processed', extra_str='gls-gen-paper-' + args.label) +
'/' + ds for ds in dsets
]
outdir = get_outdir(
args,
baseoutdir,
varname='data',
varval=study + '/' + '-vs-'.join(dsets)
) # for data, only the plots go here, since datascripts puts its output somewhere else
if len(dsets) > 1 and len(methods) == 1: # sample vs sample
glslabels = dsets
title = get_dset_title([metafos[ds] for ds in dsets])
if study != 'kate-qrs':
title += ' %s' % methstr(methods[0])
title_color = methods[0]
legends = get_dset_legends([metafos[ds] for ds in dsets])
legend_title = methstr(
methods[0]
) if study == 'kate-qrs' else None # for kate-qrs we need to put the subject _and_ the isotype in the title, so there's no room for the method
pie_chart_faces = False
print '%s:' % utils.color('green', methods[0]),
elif len(methods) > 1 and len(dsets) == 1: # method vs method
glslabels = methods
title = get_dset_title([mfo])
title_color = None
legends = [methstr(m) + ' only' for m in methods]
legend_title = None
pie_chart_faces = True
print '%s:' % utils.color('green', dsets[0]),
else:
raise Exception('one of \'em has to be length 1: %d %d' %
(len(methods), len(dsets)))
print '%s' % (' %s ' % utils.color('light_blue', 'vs')).join(glslabels)
make_gls_tree_plot(args,
outdir + '/' + '-vs-'.join(methods) + '/gls-gen-plots',
study + '-' + '-vs-'.join(dsets),
glsfnames=[
get_gls_fname(ddir,
meth,
locus=mfo['locus'],
data=True) for ddir in data_outdirs
for meth in methods
],
glslabels=glslabels,
locus=mfo['locus'],
title=title,
title_color=title_color,
legends=legends,
legend_title=legend_title,
pie_chart_faces=pie_chart_faces)
def get_data_pair_plots(args, baseoutdir, method, study, dsets):
mfo = heads.read_metadata(study)[dsets[0]]
assert heads.read_metadata(study)[dsets[1]]['locus'] == mfo['locus']
data_outdirs = [
heads.get_datadir(
study, 'processed', extra_str='gls-gen-paper-' + args.label) +
'/' + ds for ds in dsets
]
outdir = get_outdir(
args,
baseoutdir,
varname='data',
varval=study + '/' + '-vs-'.join(dsets)
) # for data, only the plots go here, since datascripts puts its output somewhere else
make_gls_tree_plot(args,
outdir + '/' + method + '/gls-gen-plots',
study + '-' + '-vs-'.join(dsets),
glsfnames=[
get_gls_fname(dout,
method,
locus=mfo['locus'],
data=True) for dout in data_outdirs
],
glslabels=dsets)
def get_data_plots(args, region, baseoutdir, methods, study, dsets):
metafos = heads.read_metadata(study)
assert len(set([metafos[ds]['locus'] for ds in dsets])) # make sure everybody has the same locus
mfo = metafos[dsets[0]]
data_outdirs = [heads.get_datadir(study, 'processed', extra_str=args.label) + '/' + ds for ds in dsets]
outdir = get_outdir(args, baseoutdir, varname='data', varval=study + '/' + '-vs-'.join(dsets)) # for data, only the plots go here, since datascripts puts its output somewhere else
title, title_color, legends, legend_title = None, None, None, None
pie_chart_faces = False
if len(dsets) > 1 and len(methods) == 1: # sample vs sample
glslabels = dsets
title = get_dset_title([metafos[ds] for ds in dsets])
if study != 'kate-qrs':
title += ' %s' % methstr(methods[0])
title_color = methods[0]
legends = get_dset_legends([metafos[ds] for ds in dsets])
legend_title = methstr(methods[0]) if study == 'kate-qrs' else None # for kate-qrs we need to put the subject _and_ the isotype in the title, so there's no room for the method
print '%s:' % utils.color('green', methods[0]),
elif len(methods) > 1 and len(dsets) == 1: # method vs method
glslabels = methods
title = get_dset_title([mfo])
title_color = None
legends = [methstr(m) + ' only' for m in methods]
legend_title = None
pie_chart_faces = len(methods) > 2 # True
print '%s:' % utils.color('green', dsets[0]),
else: # single sample plot
glslabels = dsets
print '%s' % (' %s ' % utils.color('light_blue', 'vs')).join(glslabels)
plotdir = outdir + '/' + '-vs-'.join(methods) + '/gls-gen-plots'
if args.all_regions: # NOTE not actually checking this by running... but it's the same as the gls-gen one, so it should be ok
plotdir += '/' + region
param_dirs = None
if args.add_gene_counts_to_tree_plots: # this returns 'None' for non-partis methods, which is ok for now, but I think I do usually have the parameter dir somewhere if I've run the annotation performance stuff
param_dirs = [get_param_dir(ddir, meth) for ddir in data_outdirs for meth in methods]
make_gls_tree_plot(args, region, plotdir, study + '-' + '-vs-'.join(dsets),
glsfnames=[get_gls_fname(region, ddir, meth, locus=mfo['locus'], data=True) for ddir in data_outdirs for meth in methods],
glslabels=glslabels,
locus=mfo['locus'],
title=title,
title_color=title_color,
legends=legends,
legend_title=legend_title,
pie_chart_faces=pie_chart_faces,
param_dirs=param_dirs)
def get_data_plots(args, baseoutdir, method):
for var in args.varvals:
study, dset = var.split('/')
mfo = heads.read_metadata(study)[dset]
data_outdir = heads.get_datadir(
study, 'processed',
extra_str='gls-gen-paper-' + args.label) + '/' + dset
outdir = get_outdir(
args, baseoutdir, varname='data', varval=study + '/' + dset
) # for data, only the plots go here, since datascripts puts its output somewhere else
make_gls_tree_plot(args,
outdir + '/' + method + '/gls-gen-plots',
study + '-' + dset,
glsfnames=[
get_gls_fname(data_outdir,
method,
locus=mfo['locus'],
data=True)
],
glslabels=['data'])
# ],
'crotty-fna' : [
['RUj15_ALN-FNA_week3_groupD', 'RUj15_L-ILN-FNA_week3_groupD', 'RUj15_R-ILN-FNA_week3_groupD'],
['RUj15_ALN-FNA_week3_groupD', 'RUj15_L-ILN-FNA_week15_groupD', 'RUj15_R-ILN-FNA_week15_groupD'],
['ROp15_R-ILN-FNA_week3_groupC', 'ROp15_L-ILN-FNA_week3_groupC', 'ROp15_ALN-FNA_week3_groupC'],
['ROp15_R-ILN-FNA_week9_groupC', 'ROp15_L-ILN-FNA_week9_groupC', 'ROp15_ALN-FNA_week9_groupC'],
['ROp15_R-ILN-FNA_week15_groupC', 'ROp15_L-ILN-FNA_week15_groupC', 'ROp15_ALN-FNA_week15_groupC'],
['ROp15_R-ILN-FNA_week21_groupC', 'ROp15_L-ILN-FNA_week21_groupC', 'ROp15_ALN-FNA_week21_groupC'],
['RJk15_L-ILN-FNA_week3_groupB', 'RJk15_R-ILN-FNA_week3_groupB', 'RJk15_ALN-FNA_week3_groupB'],
['RJk15_L-ILN-FNA_week9_groupB', 'RJk15_R-ILN-FNA_week9_groupB'],
['RJk15_L-ILN-FNA_week15_groupB', 'RJk15_R-ILN-FNA_week15_groupB', 'RJk15_ALN-FNA_week15_groupB'],
['RJk15_L-ILN-FNA_week21_groupB', 'RJk15_R-ILN-FNA_week21_groupB', 'RJk15_ALN-FNA_week21_groupB'],
]
}
default_varvals['data'] = ':'.join([study + '/' + heads.full_dataset(heads.read_metadata(study), dset) for study in default_varvals['data'] for dset in default_varvals['data'][study]])
for study in all_data_groups:
for idp in range(len(all_data_groups[study])):
all_data_groups[study][idp] = [heads.full_dataset(heads.read_metadata(study), ds) for ds in all_data_groups[study][idp]]
# ----------------------------------------------------------------------------------------
parser = argparse.ArgumentParser()
parser.add_argument('action', choices=['mfreq', 'nsnp', 'multi-nsnp', 'prevalence', 'n-leaves', 'weibull', 'alcluster', 'gls-gen', 'data'])
parser.add_argument('--methods', default='partis') # not using <choices> 'cause it's harder since it's a list
parser.add_argument('--method-vs-method', action='store_true')
parser.add_argument('--sample-vs-sample', action='store_true')
parser.add_argument('--v-genes', default='IGHV4-39*01')
parser.add_argument('--locus', default='igh')
parser.add_argument('--all-regions', action='store_true') # it'd be nicer to just have an arg for which region we're running on, but i need a way to keep the directory structure for single-region plots the same as before I generalized to d and j
parser.add_argument('--varvals')
parser.add_argument('--n-event-list', default='1000:2000:4000:8000') # NOTE modified later for multi-nsnp also NOTE not used for gen-gset
parser.add_argument('--gls-gen-events', type=int, default=50000)
'RJk15_L-ILN-FNA_week3_groupB', 'RJk15_R-ILN-FNA_week3_groupB',
'RJk15_ALN-FNA_week3_groupB'
],
['RJk15_L-ILN-FNA_week9_groupB', 'RJk15_R-ILN-FNA_week9_groupB'],
[
'RJk15_L-ILN-FNA_week15_groupB', 'RJk15_R-ILN-FNA_week15_groupB',
'RJk15_ALN-FNA_week15_groupB'
],
[
'RJk15_L-ILN-FNA_week21_groupB', 'RJk15_R-ILN-FNA_week21_groupB',
'RJk15_ALN-FNA_week21_groupB'
],
]
}
default_varvals['data'] = ':'.join([
study + '/' + heads.full_dataset(heads.read_metadata(study), dset)
for study in default_varvals['data']
for dset in default_varvals['data'][study]
])
for study in all_data_groups:
for idp in range(len(all_data_groups[study])):
all_data_groups[study][idp] = [
heads.full_dataset(heads.read_metadata(study), ds)
for ds in all_data_groups[study][idp]
]
# ----------------------------------------------------------------------------------------
parser = argparse.ArgumentParser()
parser.add_argument('action',
choices=[
'mfreq', 'nsnp', 'multi-nsnp', 'prevalence',
'n-leaves', 'weibull', 'alcluster', 'gls-gen', 'data'
import colored_traceback.always
sys.path.insert(1, './python')
import utils
import glutils
sys.path.insert(1, './datascripts')
import heads
scolors = {
'ok': 'DarkSeaGreen',
'missing': 'IndianRed',
'spurious': 'IndianRed',
'data': 'LightSteelBlue',
'both': 'LightGrey',
}
metafos = heads.read_metadata('kate-qrs')
for ds in metafos:
if 'LN1' in ds or 'LN2' in ds:
scolors[ds] = '#85ad98' # green
elif 'LN4' in ds or 'LN3' in ds:
scolors[ds] = '#94a3d1' # blue
faces = {
'missing': ete3.CircleFace(10, 'white'),
'spurious': ete3.CircleFace(10, 'black')
}
def get_cmdfos(cmdstr, workdir, outfname):
return [{'cmd_str': cmdstr, 'workdir': workdir, 'outfname': outfname}]