def folderRN(gspath, newpath, cores=1):
"""
"""
lis = lsFiles([gspath])
if lis != 0:
h.parrun(['gsutil -m mv ' + val + " " + newpath for val in lis],
cores=cores)
else:
raise ValueError('no such folder')
def recoverFiles(files, cores=1):
"""
recover a set of files in parallel that were erased
files need to have their #id appended found using ls -al file
Args:
----
files: gs paths
location: to move the files to
"""
cmd = ['gsutil mv ' + f + ' ' + f.split('#')[0] for f in files]
h.parrun(cmd, cores=cores)
def patternRN(rename_dict,
location,
wildcards,
types=[],
dryrun=True,
check_dependencies=True,
cores=1):
"""
rename/move a bunch of GCP objects found in some specific places
Args:
-----
rename_dict: dict(prevName,newName)
location:
wildcards: list[str] can be one of ['**', '.*', '*.','-.*'] if needs to be
** means any occurence of this file in any folder will change its name
.* means all file unregarding of the suffix, will rename them all a.bam [a]da.bai to b.bam, [b]da.bai
*. means all files with the suffix, will change the suffix of these files from a to b
-.* means all file unregarding of the suffix, will rename them. not just replacing the a part with a to b but the full file name [a]dea.bam to b.bam
types: Nothing yet
test: if test, just shows the command but does not run it
cores: cores tells on how many processor to parallelize the tas#k
"""
val = []
for k, v in rename_dict.items():
val.append(v)
if k in val and check_dependencies:
raise ValueError('circular dependency in the rename with key ' + k)
for k, v in rename_dict.items():
loc = location
if '**' in wildcards:
loc += '**/'
if '*.' in wildcards or '-.*' in wildcards:
loc += '*'
loc += k
if '.*' in wildcards or '-.*' in wildcards:
loc += '*'
res = os.popen('gsutil -m ls ' + loc).read().split('\n')[:-1]
print('found ' + str(len(res)) + ' files to rename')
if '-.*' in wildcards:
cmd = [
"gsutil mv " + val + " " + '/'.join(val.split('/')[:-1]) +
'/' + v + '.' + '.'.join(val.split('/')[-1].split('.')[1:])
for val in res
]
else:
cmd = ["gsutil mv " + val + " " + val.replace(k, v) for val in res]
if dryrun:
print(cmd)
else:
h.parrun(cmd, cores=cores)
def deleteJob(workspaceid, subid, taskid, deleteCurrent=False, dryrun=True):
"""
removes files generated by a job on Terra
Args:
-----
workspaceid: str wokspace name
subid: str the name of the job
taskid: str the name of the task in this job
DeleteCurrent: bool whether or not to delete files if they appear in one of the sample/samplesets/pairs data tables
dryrun: bool just plot the commands but don't execute them
"""
wm = dm.WorkspaceManager(workspaceid)
bucket = wm.get_bucket_id()
data = []
if deleteCurrent:
if dryrun:
print('gsutil -m rm gs://' + bucket + '/' + subid + '/*/' +
taskid + '/**')
else:
res = subprocess.run('gsutil -m rm gs://' + bucket + '/' + subid +
'/*/' + taskid + '/**',
shell=True,
capture_output=True)
if res.returncode != 0:
raise ValueError(str(res.stderr))
else:
res = subprocess.run('gsutil -m ls gs://' + bucket + '/' + subid +
'/*/' + taskid + '/**',
shell=True,
capture_output=True)
if res.returncode != 0 or len(str(res.stdout)) < 4:
raise ValueError(str(res.stderr))
data += str(res.stdout)[2:-1].split('\\n')[:-1]
if "TOTAL:" in data[-1]:
data = data[:-1]
sam = pd.concat(
[wm.get_samples(),
wm.get_pairs(),
wm.get_sample_sets()])
tokeep = set([
val for val in sam.values.ravel()
if type(val) is str and val[:5] == 'gs://'
])
torm = set(data) - tokeep
if dryrun:
print(torm)
else:
h.parrun(['gsutil rm ' + i for i in torm], cores=12)
async def indexBams(bucketpath, cores=4):
"""
given a bucket path, will index all .bam files without an associated index and return their paths
"""
files = gcp.lsFiles([bucketpath])
bams = [val for val in files if '.bam' in val[-4:]]
unindexed = [
val for val in bams
if val[:-4] + '.bai' not in files and val[:4] + '.bam.bai' not in files
]
print("found " + str(len(unindexed)) + " files to reindex")
h.parrun([
"export GCS_OAUTH_TOKEN=`gcloud auth application-default print-access-token` && samtools index "
+ val for val in unindexed
], cores)
return {val: val[:-4] + ".bam.bai" for val in unindexed}
def changeGSlocation(workspacefrom,
newgs,
workspaceto=None,
prevgslist=[],
index_func=None,
flag_non_matching=False,
onlysamples=[],
onlycol=[],
entity='samples',
droplists=True,
keeppath=True,
dry_run=True,
par=20):
"""
Function to move data around from one workspace to a bucket or to another workspace.
can also work on dataframes containing lists of paths
Args:
-----
workspacefrom: the workspace name where the data is
newgs: the newgs bucket where to copy the data in
workspaceto: if we should have these new samples and columns added to another workspace instead \
of just updating the same one (usefull to copy one workspace to another)
prevgslist: if providded, will only move files that are in the set of google bucket listed here
index_func: *WIP* unused
flag_non_matching: if set to true and prevgslist is set to some value, will return a list of samples that were not
matched to anything in the prevgslist
onlycol: do this only on a subset of columns in terra workspace
entity: the entity in the terra workspace on which to do this
droplists: if set to true remove all columns containing list of paths (list of path are not uploaded well in terra)
keeppath: if set to true, will keep the full object path and just change the bucket
dry_run: if set to true will not update anything on Terra but just return the result
par: on how many processor do the gs copy commands.
Returns:
-------
torename: the pandas.df containing the new paths
flaglist: the samples that were non matching (if flag_non_matching is set to true)
"""
flaglist = []
wmfrom = dm.WorkspaceManager(workspacefrom)
a = wmfrom.get_entities(entity)
if len(onlysamples) > 0:
a = a[a.index.isin(onlysamples)]
print("using the data from " + workspacefrom + " " + entity + " list")
if len(a) == 0:
raise ValueError('no ' + entity)
if onlycol:
a = a[onlycol]
todrop = set()
torename = {}
print(
'this should only contains gs:// paths otherwise precise columns using \"onlycol\"'
)
for col in a.columns.tolist():
val = []
for k, prev in a[col].iteritems():
if type(prev) is str:
new = prev
if newgs not in new:
if len(prevgslist) > 0:
for prevgs in prevgslist:
new = new.replace(prevgs, newgs)
if flag_non_matching:
if new == prev:
flaglist.append(prev)
if not keeppath:
new = newgs + new.split('/')[-1]
else:
new = newgs + '/'.join(new.split('/')[3:])
else:
print("sample " + str(k) + " was already in the new gs")
val.append(new)
# IN CASE WE HAVE A LIST
if type(prev) is list:
if droplists:
todrop.add(k)
continue
ind = []
for prevname in prev:
newname = prevname
if newgs not in newname:
if len(prevgslist) > 0:
for prevgs in prevgslist:
new = new.replace(prevgs, newgs)
if flag_non_matching:
if new == prev:
flaglist.append(prev)
if not keeppath:
new = newgs + new.split('/')[-1]
else:
new = newgs + '/'.join(new.split('/')[3:])
else:
print("sample " + str(k) +
" was already in the new gs")
ind.append(newname)
val.append(ind)
torename.update({col: val})
if not dry_run:
if keeppath:
h.parrun([
'gsutil mv ' + a.iloc[i][col] + ' ' + v
for i, v in enumerate(val)
],
cores=20)
else:
gcp.mvFiles(a[col].tolist(), newgs)
else:
if keeppath:
print([
'gsutil mv ' + a.iloc[i][col] + ' ' + v
for i, v in enumerate(val)
])
else:
print("mv " + str(a[col].tolist()) + " " + newgs)
torename = pd.DataFrame(data=torename,
index=[i for i in a.index.tolist() if i != 'nan'])
if workspaceto is not None:
wmto = dm.WorkspaceManager(workspaceto)
if not dry_run:
wmto.disable_hound().update_entity_attributes(entity, torename)
return torename, flaglist
async def getSpikeInControlScales(refgenome, fastq=None, fastQfolder='', mapper='bwa', pairedEnd=False, cores=1,
pathtosam='samtools', pathtotrim_galore='trim_galore', pathtobwa='bwa',
totrim=True, tomap=True, tofilter=True, results='res/', toremove=False):
"""
Will extract the spikeInControls from a fastq file (usefull for, let say ChIPseq data with spike ins)
Count based sequencing data is not absolute and will be normalized as each sample will be sequenced at a specific depth.
To figure out what was the actual sample concentration, we use Spike In control
You should have FastQfolder/[NAME].fastq & BigWigFolder/[NAME].bw with NAME being the same for the same samples
Args:
-----
refgenome: str the file path to the indexed reference genome
FastQfolder: str the folder path where the fastq files are stored (should be named the same as files in BigWigFolder)
BigWigFolder: str the folder path where the bigwig files are stored (should be named the same as files in FastQfolder)
mapper: str flag to 'bwa', ...
pairedEnd: Bool flat to true for paired end sequences. if true, You should have FastQfolder/[NAME]_1|2.fastq
Returns:
--------
dict(file,float) the scaling factor dict
"""
if len(fastQfolder) > 0:
print('using all files from folder')
fastqs = os.listdir(fastQfolder)
fastqs = [i for i in fastqs if '.fq.gz' ==
i[-6:] or '.fastq.gz' == i[-9:]]
fastqs.sort()
if pairedEnd and (tomap or totrim):
print("need to be name_*1, name_*2")
fastqs = [i for i in h.grouped(fastqs, 2)]
elif fastq is None:
raise ValueError('you need input files')
else:
if type(fastq) is list:
print('your files need to be all in the same folder')
fastQfolder = '/'.join(fastq[0].split('/')[:-1]) + '/'
if not totrim and not tomap:
fastqs = [f.split('/')[-1] for f in fastq]
else:
print("need to be name_*1, name_*2")
fastqs = [[f[0].split('/')[-1], f[1].split('/')[-1]]
for f in h.grouped(fastq, 2)]
else:
fastQfolder = '/'.join(fastq.split('/')[:-1]) + '/'
fastqs = [fastq.split('/')[-1]]
print(fastqs)
if not totrim:
print("you need to have your files in the " + results + " folder")
if totrim and tomap:
print("\n\ntrimming\n\n")
if pairedEnd:
cmds = []
rm = []
for file in fastqs:
cmd = pathtotrim_galore + ' --paired --fastqc --gzip ' + fastQfolder + \
file[0] + ' ' + fastQfolder + file[1] + " -o " + results
if toremove:
rm.append('rm ' + fastQfolder +
file[0] + ' ' + fastQfolder + file[1])
cmds.append(cmd)
print(cmds)
h.parrun(cmds, cores, add=rm)
fastqs = [[file[0].split('.')[
0] + '_val_1.fq.gz', file[1].split('.')[0] + '_val_2.fq.gz'] for file in fastqs]
if tomap:
print("\n\nmapping\n\n")
if pairedEnd:
cmds = []
rm = []
for file in fastqs:
cmd = pathtobwa + ' mem ' + refgenome + ' ' + results + file[0] + ' ' + results +\
file[1] + ' | ' + pathtosam + ' sort - -o ' + \
results + file[0].split('.')[0] + '.sorted.bam'
if toremove:
rm.append('rm ' + results +
file[0] + ' ' + results + file[1])
cmds.append(cmd)
h.parrun(cmds, cores, add=rm)
fastqs = [file[0].split('.')[0] + '.sorted.bam' for file in fastqs]
if tofilter:
print("\n\nfiltering\n\n")
cmds = []
rm = []
h.parrun([pathtosam + ' index ' + results + file.split('.')
[0] + '.sorted.bam' for file in fastqs], cores)
h.parrun([pathtosam + ' flagstat ' + results + file.split('.')[0] + '.sorted.bam > ' +
results + file.split('.')[0] + '.sorted.bam.flagstat' for file in fastqs], cores)
h.parrun([pathtosam + ' idxstats ' + results + file.split('.')[0] + '.sorted.bam > ' +
results + file.split('.')[0] + '.sorted.bam.idxstat' for file in fastqs], cores)
fastqs = [file.split('.')[0] + '.sorted.bam' for file in fastqs]
else:
print("files need to be named: NAME.sorted.bam")
fastqs = [file for file in fastqs if '.sorted.bam' == file[-11:]]
mapped = {}
norm = {}
unique_mapped = {}
print("\n\ncounting\n\n")
for file in fastqs:
mapped[file.split('.')[0]] = int(os.popen(pathtosam + ' view -c -F 0x004 -F 0x0008 -f 0x001 -F 0x0400 -q 1 ' + results +
file + ' -@ ' + str(cores)).read().split('\n')[0])
# unique_mapped[file.split('.')[0]] = int(re.findall("Mapped reads: (\d+)", os.popen('bamtools stats -in '+results +
# file + '.sorted.bam').read())[0])
nbmapped = np.array([i for i in mapped.values()])
nbmapped = np.sort(nbmapped)[0] / nbmapped.astype(float)
for i, val in enumerate(mapped.keys()):
norm[val] = nbmapped[i]
return norm, mapped, # unique_mapped
