def __init__(self, config):
#BEGIN_CONSTRUCTOR
self.callback_url = os.environ['SDK_CALLBACK_URL']
self.shared_folder = config['scratch']
self.SSU = SequenceSetUtils(os.environ['SDK_CALLBACK_URL'])
self.MOU = MotifUtils(os.environ['SDK_CALLBACK_URL'])
self.MEU = MemeUtil(self.shared_folder)
self.dfu = DataFileUtil(self.callback_url)
self.GR = GenerateReport()
#END_CONSTRUCTOR
pass
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" (SS_ref -
optional, used for exact genome locations if possible) ->
structure: parameter "workspace_name" of String, parameter
"fastapath" of String, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long, parameter "SS_ref" of
String, parameter "obj_name" of String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
if 'motif_min_length' not in params:
params['motif_min_length'] = 8
if 'motif_max_length' not in params:
params['motif_max_length'] = 16
motMin = params['motif_min_length']
motMax = params['motif_max_length']
promoterFastaFilePath = params['fastapath']
HU = HomerUtil()
HomerMotifCommand = HU.build_homer_motif_command(
promoterFastaFilePath, motMin, motMax, params['background'])
HomerLocationCommand = HU.build_homer_location_command(
promoterFastaFilePath)
HU.run_homer_command(HomerMotifCommand)
HU.run_homer_command(HomerLocationCommand)
homer_out_path = '/kb/module/work/tmp/homer_out'
#homer_out_path = os.path.join(self.shared_folder, str(uuid.uuid4()))
homer_params = {
'ws_name': params['workspace_name'],
'path': homer_out_path + '/homerMotifs.all.motifs',
'location_path': homer_out_path + '/homer_locations.txt',
'obj_name': params['obj_name']
}
MOU = MotifUtils(self.callback_url)
dfu = DataFileUtil(self.callback_url)
locDict = {}
if 'SS_ref' in params:
get_ss_params = {'object_refs': [params['SS_ref']]}
SS = dfu.get_objects(get_ss_params)['data'][0]['data']
for s in SS['sequences']:
if s['source']['assembly_id'] != '' and s['source'][
'location'] != []:
locDict['sequence_id'] = {
'contig': s['source']['location'][0][0],
'start': str(s['source']['location'][0][1])
}
if len(locDict.keys()) > 0:
homer_params['absolute_locations'] = locDict
homer_params['min_len'] = motMin
homer_params['max_len'] = motMax
obj_ref = MOU.UploadFromHomer(homer_params)['obj_ref']
HU.write_obj_ref(homer_out_path, obj_ref)
#file = open(homer_out_path+"/homer_obj.txt","w")
#file.write(obj_ref)
#file.close()
#HERE:
#we've got object ref
#we've got html building functions
#build report, setup return,
#make report and return it
timestamp = int(
(datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()
* 1000)
timestamp = str(timestamp)
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
lineCount = 0
with open(promoterFastaFilePath, 'r') as pFile:
for line in pFile:
lineCount += 1
numFeat = lineCount / 2
with open(promoterFastaFilePath, 'r') as pFile:
fileStr = pFile.read()
promHtmlStr = '<html><body> ' + fileStr + ' </body></html>'
with open(htmlDir + '/promoters.html', 'w') as promHTML:
promHTML.write(promHtmlStr)
JsonPath = '/kb/module/work/tmp'
dfu = DataFileUtil(self.callback_url)
get_obj_params = {'object_refs': [obj_ref]}
homerMotifSet = dfu.get_objects(get_obj_params)['data'][0]['data']
mr = MakeNewReport()
mr.MakeReport(htmlDir, homerMotifSet)
try:
html_upload_ret = dfu.file_to_shock({
'file_path': htmlDir,
'make_handle': 0,
'pack': 'zip'
})
except:
raise ValueError('error uploading HTML file to shock')
reportName = 'HomerMotifFinder_report_' + str(uuid.uuid4())
reportObj = {
'objects_created': [{
'ref': obj_ref,
'description': 'Motif Set generated by Homer'
}],
'message':
'',
'direct_html':
None,
'direct_html_link_index':
0,
'file_links': [],
'html_links': [],
'html_window_height':
220,
'workspace_name':
params['workspace_name'],
'report_object_name':
reportName
}
# attach to report obj
reportObj['direct_html'] = ''
reportObj['direct_html_link_index'] = 0
reportObj['html_links'] = [{
'shock_id': html_upload_ret['shock_id'],
#'name': 'promoter_download.zip',
'name': 'index.html',
'label': 'Save promoter_download.zip'
}]
report = KBaseReport(self.callback_url, token=ctx['token'])
report_info = report.create_extended_report(reportObj)
output = {
'report_name': report_info['name'],
'report_ref': report_info['ref']
}
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" (SS_ref -
optional, used for exact genome locations if possible) ->
structure: parameter "workspace_name" of String, parameter
"fastapath" of String, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long, parameter "SS_ref" of
String, parameter "obj_name" of String, parameter "background" of
Long
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
#TODO: Things to fix in here...
# Use MotifUtils to parse output and create object
# create new function for report ?
if 'motif_min_length' not in params:
params['motif_min_length'] = 8
if 'motif_max_length' not in params:
params['motif_max_length'] = 16
motMin = params['motif_min_length']
motMax = params['motif_max_length']
promoterFastaFilePath = params['fastapath']
MEU = MemeUtil()
MEMEMotifCommand = MEU.build_meme_command(promoterFastaFilePath,
motMin, motMax,
params['background'])
MEU.run_meme_command(MEMEMotifCommand)
meme_out_path = '/kb/module/work/tmp/meme_out/meme.txt'
meme_params = {
'ws_name': params['workspace_name'],
'path': meme_out_path,
'obj_name': params['obj_name']
}
MOU = MotifUtils(self.callback_url)
dfu = DataFileUtil(self.callback_url)
locDict = {}
if 'SS_ref' in params:
get_ss_params = {'object_refs': [params['SS_ref']]}
SS = dfu.get_objects(get_ss_params)['data'][0]['data']
for s in SS['sequences']:
if s['source']['assembly_id'] != '' and s['source'][
'location'] != []:
locDict['sequence_id'] = {
'contig': s['source']['location'][0][0],
'start': str(s['source']['location'][0][1])
}
if len(locDict.keys()) > 0:
meme_params['absolute_locations'] = locDict
meme_params['min_len'] = motMin
meme_params['max_len'] = motMax
obj_ref = MOU.UploadFromMEME(meme_params)['obj_ref']
MEU.write_obj_ref("/kb/module/work/tmp/meme_out", obj_ref)
timestamp = int(
(datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()
* 1000)
timestamp = str(timestamp)
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
dfu = DataFileUtil(self.callback_url)
get_obj_params = {'object_refs': [obj_ref]}
memeMotifSet = dfu.get_objects(get_obj_params)['data'][0]['data']
GR = GenerateReport()
GR.MakeMotifReport(htmlDir, memeMotifSet)
try:
html_upload_ret = dfu.file_to_shock({
'file_path': htmlDir,
'make_handle': 0,
'pack': 'zip'
})
except:
raise ValueError('error uploading HTML file to shock')
#Create motif set object from MotifList
#TODO set parameters correctly
reportName = 'MEMEMotifFinder_report_' + str(uuid.uuid4())
reportObj = {
'objects_created': [{
'ref': obj_ref,
'description': 'Motif Set generated by MEME'
}],
'message':
'',
'direct_html':
None,
'direct_html_link_index':
0,
'file_links': [],
'html_links': [],
'html_window_height':
220,
'workspace_name':
params['workspace_name'],
'report_object_name':
reportName
}
# attach to report obj
#reportObj['direct_html'] = None
reportObj['direct_html'] = ''
reportObj['direct_html_link_index'] = 0
reportObj['html_links'] = [{
'shock_id': html_upload_ret['shock_id'],
#'name': 'promoter_download.zip',
'name': 'index.html',
'label': 'Save promoter_download.zip'
}]
report = KBaseReport(self.callback_url, token=ctx['token'])
report_info = report.create_extended_report(reportObj)
output = {
'report_name': report_info['name'],
'report_ref': report_info['ref']
}
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
class MotifFinderMEME:
'''
Module Name:
MotifFinderMEME
Module Description:
A KBase module: MotifFinderMEME
'''
######## WARNING FOR GEVENT USERS ####### noqa
# Since asynchronous IO can lead to methods - even the same method -
# interrupting each other, you must be *very* careful when using global
# state. A method could easily clobber the state set by another while
# the latter method is running.
######################################### noqa
VERSION = "0.0.1"
GIT_URL = "https://github.com/kbasecollaborations/MotifFinderMEME.git"
GIT_COMMIT_HASH = "ed2e967cf8aa213388e4166525cd8ce2918d4fae"
#BEGIN_CLASS_HEADER
#END_CLASS_HEADER
# config contains contents of config file in a hash or None if it couldn't
# be found
def __init__(self, config):
#BEGIN_CONSTRUCTOR
self.callback_url = os.environ['SDK_CALLBACK_URL']
self.shared_folder = config['scratch']
self.SSU = SequenceSetUtils(os.environ['SDK_CALLBACK_URL'])
self.MOU = MotifUtils(os.environ['SDK_CALLBACK_URL'])
self.MEU = MemeUtil(self.shared_folder)
self.dfu = DataFileUtil(self.callback_url)
self.GR = GenerateReport()
#END_CONSTRUCTOR
pass
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" (SS_ref -
optional, used for exact genome locations if possible) ->
structure: parameter "workspace_name" of String, parameter
"fastapath" of String, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long, parameter "SS_ref" of
String, parameter "obj_name" of String, parameter "background" of
Long
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
if 'motif_min_length' not in params:
params['motif_min_length'] = 8
if 'motif_max_length' not in params:
params['motif_max_length'] = 16
motMin = params['motif_min_length']
motMax = params['motif_max_length']
promoterFastaFilePath = self.SSU.SeqSetToFasta({
'ws_name':
params['workspace_name'],
'SS_ref':
params['SS_ref']
})['path']
MEMEMotifCommand = self.MEU.build_meme_command(promoterFastaFilePath,
motMin, motMax,
params['background'])
meme_out_path = self.MEU.run_meme_command(MEMEMotifCommand)
meme_params = {
'ws_name': params['workspace_name'],
'format': 'MEME',
'file': {
'path': meme_out_path
},
'obj_name': params['obj_name'],
'seq_set_ref': params['SS_ref']
}
# MOU.parseMotifSet with the same parameters will
# return a dictionary of the motifset object that you save on
# your own
#
# MOU.saveMotifSet will save the object with DataFileUtils to
# whatever workspace you specify in ws_name
#
# This function will also download the sequence set as a fasta to
# unique (uuid4) file name in the scratch directory
obj_ref = self.MOU.saveMotifSet(meme_params)
timestamp = str(
int((datetime.utcnow() -
datetime.utcfromtimestamp(0)).total_seconds() * 1000))
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
get_obj_params = {'object_refs': [obj_ref]}
memeMotifSet = self.dfu.get_objects(get_obj_params)['data'][0]['data']
self.GR.MakeMotifReport(htmlDir, memeMotifSet)
try:
html_upload_ret = self.dfu.file_to_shock({
'file_path': htmlDir,
'make_handle': 0,
'pack': 'zip'
})
except Exception:
raise ValueError('Error uploading HTML file: ' + str(htmlDir) +
' to shock')
reportname = 'MEMEMotifFinder_report_' + str(uuid.uuid4())
reportobj = {
'objects_created': [{
'ref': obj_ref,
'description': 'Motif Set generated by MEME'
}],
'message':
'',
'direct_html':
None,
'direct_html_link_index':
0,
'file_links': [],
'html_links': [],
'html_window_height':
220,
'workspace_name':
params['workspace_name'],
'report_object_name':
reportname
}
# attach to report obj
reportobj['direct_html'] = ''
reportobj['direct_html_link_index'] = 0
reportobj['html_links'] = [{
'shock_id': html_upload_ret['shock_id'],
'name': 'index.html',
'label': 'Save promoter_download.zip'
}]
report = KBaseReport(self.callback_url, token=ctx['token'])
report_info = report.create_extended_report(reportobj)
output = {
'report_name': report_info['name'],
'report_ref': report_info['ref']
}
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
def ExtractPromotersFromFeatureSetandDiscoverMotifs(self, ctx, params):
"""
:param params: instance of type "extract_input" -> structure:
parameter "workspace_name" of String, parameter "genome_ref" of
String, parameter "featureSet_ref" of String, parameter
"promoter_length" of Long, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long, parameter "obj_name" of
String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN ExtractPromotersFromFeatureSetandDiscoverMotifs
BuildParams = {
'ws_name': params['workspace_name'],
'FeatureSet_ref': params['featureSet_ref'],
'genome_ref': params['genome_ref'],
'upstream_length': params['promoter_length']
}
SSret = self.SSU.buildFromFeatureSet(BuildParams)
SSref = SSret['SequenceSet_ref']
fastapath = '/kb/module/work/tmp/tmpSeqSet.fa'
newfastapath = '/kb/module/work/tmp/SeqSet.fa'
fastapath = newfastapath
FastaParams = {
'workspace_name': params['workspace_name'],
'SequenceSetRef': SSref,
'fasta_outpath': fastapath
}
output = self.BuildFastaFromSequenceSet(ctx, FastaParams)
findmotifsparams = {
'workspace_name': params['workspace_name'],
'fastapath': fastapath,
'motif_min_length': params['motif_min_length'],
'motif_max_length': params['motif_max_length'],
'SS_ref': SSref,
'obj_name': params['obj_name']
}
output = self.find_motifs(ctx, findmotifsparams)[0]
#END ExtractPromotersFromFeatureSetandDiscoverMotifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError(
'Method ExtractPromotersFromFeatureSetandDiscoverMotifs return value '
+ 'output is not type dict as required.')
# return the results
return [output]
def DiscoverMotifsFromFasta(self, ctx, params):
"""
:param params: instance of type "discover_fasta_input" -> structure:
parameter "workspace_name" of String, parameter "fasta_path" of
String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN DiscoverMotifsFromFasta
raise NotImplementedError(
'Discovery of motifs from a FASTA file is not yet implemented.')
#END DiscoverMotifsFromFasta
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method DiscoverMotifsFromFasta return value ' +
'output is not type dict as required.')
# return the results
return [output]
def DiscoverMotifsFromSequenceSet(self, ctx, params):
"""
:param params: instance of type "discover_seq_input" -> structure:
parameter "workspace_name" of String, parameter "genome_ref" of
String, parameter "SS_ref" of String, parameter "promoter_length"
of Long, parameter "motif_min_length" of Long, parameter
"motif_max_length" of Long, parameter "obj_name" of String,
parameter "background" of Long, parameter "mask_repeats" of Long,
parameter "background_group" of mapping from String to String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN DiscoverMotifsFromSequenceSet
if params['background_group'] == None:
params['background_group'] = {'background': 0}
findmotifsparams = {
'workspace_name': params['workspace_name'],
'motif_min_length': params['motif_min_length'],
'motif_max_length': params['motif_max_length'],
'SS_ref': params['SS_ref'],
'obj_name': params['obj_name']
}
if params['background_group']['background'] == 1:
findmotifsparams['background'] = 1
else:
findmotifsparams['background'] = 0
output = self.find_motifs(ctx, findmotifsparams)[0]
#END DiscoverMotifsFromSequenceSet
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError(
'Method DiscoverMotifsFromSequenceSet return value ' +
'output is not type dict as required.')
# return the results
return [output]
def status(self, ctx):
#BEGIN_STATUS
returnVal = {
'state': "OK",
'message': "",
'version': self.VERSION,
'git_url': self.GIT_URL,
'git_commit_hash': self.GIT_COMMIT_HASH
}
#END_STATUS
return [returnVal]
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" (SS_ref -
optional, used for exact genome locations if possible) ->
structure: parameter "workspace_name" of String, parameter
"fastapath" of String, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long, parameter "SS_ref" of String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
if 'motif_min_length' not in params:
params['motif_min_length'] = 8
if 'motif_max_length' not in params:
params['motif_max_length'] = 16
motMin = params['motif_min_length']
motMax = params['motif_max_length']
#promoterFastaFilePath = self.get_promoter_for_gene(ctx,params)[0]
promoterFastaFilePath = params['fastapath']
#GibbsMotifCommand = GU.build_gibbs_command(promoterFastaFilePath)
gibbsCommandList = []
for i in range(motMin, motMax + 1, 2):
gibbsCommandList.append(
GU.build_gibbs_command(promoterFastaFilePath, i))
for g in gibbsCommandList:
GU.run_gibbs_command(g)
#GU.run_gibbs_command(GibbsMotifCommand)
gibbs_out_path = '/kb/module/work/tmp/gibbs'
gibbs_params = {
'ws_name': params['workspace_name'],
'path': gibbs_out_path,
'obj_name': params['obj_name']
}
MOU = MotifUtils(self.callback_url)
dfu = DataFileUtil(self.callback_url)
locDict = {}
if 'SS_ref' in params:
get_ss_params = {'object_refs': [params['SS_ref']]}
SS = dfu.get_objects(get_ss_params)['data'][0]['data']
for s in SS['sequences']:
if s['source'] is not None:
locDict['sequence_id'] = {
'contig': s['source']['location'][0][0],
'start': str(s['source']['location'][0][1])
}
if len(locDict.keys()) > 0:
gibbs_params['absolute_locations'] = locDict
gibbs_params['min_len'] = motMin
gibbs_params['max_len'] = motMax
obj_ref = MOU.UploadFromGibbs(gibbs_params)['obj_ref']
#memeMotifList = MEU.parse_meme_output()
#HERE:
#we've got object ref
#we've got html building functions
#build report, setup return,
#make report and return it
#buildReportFromMotifSet()
timestamp = int(
(datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()
* 1000)
timestamp = str(timestamp)
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
lineCount = 0
with open(promoterFastaFilePath, 'r') as pFile:
for line in pFile:
lineCount += 1
numFeat = lineCount / 2
with open(promoterFastaFilePath, 'r') as pFile:
fileStr = pFile.read()
promHtmlStr = '<html><body> ' + fileStr + ' </body></html>'
with open(htmlDir + '/promoters.html', 'w') as promHTML:
promHTML.write(promHtmlStr)
JsonPath = '/kb/module/work/tmp'
dfu = DataFileUtil(self.callback_url)
get_obj_params = {'object_refs': [obj_ref]}
gibbsMotifSet = dfu.get_objects(get_obj_params)['data'][0]['data']
MakeReport(htmlDir, gibbsMotifSet)
#buildReportFromMotifSet(gibbsMotifSet,htmlDir,'gibbs')
#TODO: Here replace the makereport with a call to motifset utils
#subprocess.call(['python','/kb/module/lib/identify_promoter/Utils/makeReport.py',JsonPath + '/meme_out/meme.json',htmlDir + '/meme.html',str(numFeat)])
#fullMotifList = []
#for m in memeMotifList:
# fullMotifList.append(m)
#What needs to happen here:
#call makeLogo for each of the json outputs(capture these from somewhere)
#plt.rcParams['figure.dpi'] = 300
#htmlFiles = ['index.html','gibbs.html','homer.html']
#shockParamsList = []
#for f in htmlFiles:
# shockParamsList.append({'file_path': htmlDir + f ,'make_handle': 0, 'pack': 'zip'})
try:
html_upload_ret = dfu.file_to_shock({
'file_path': htmlDir,
'make_handle': 0,
'pack': 'zip'
})
except:
raise ValueError('error uploading HTML file to shock')
#Create motif set object from MotifList
#TODO set parameters correctly
#add narrative support to set
#MSO = {}
#MSO['Condition'] = 'Temp'
#MSO['FeatureSet_ref'] = '123'
#MSO['Motifs'] = []
#MSO['Alphabet'] = ['A','C','G','T']
#MSO['Background'] = {}
#for letter in MSO['Alphabet']:
# MSO['Background'][letter] = 0.0
#MSU.parseMotifList(fullMotifList,MSO)
#objname = 'MotifSet' + str(int((datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()*1000))
#Pass motif set into this
#save_objects_params = {}
#save_objects_params['id'] = self.ws_info[0]
#save_objects_params['id'] = long(params['workspace_name'].split('_')[1])
#save_objects_params['id'] = dfu.ws_name_to_id(params['workspace_name'])
#save_objects_params['objects'] = [{'type': 'KBaseGwasData.MotifSet' , 'data' : MSO , 'name' : objname}]
#info = dfu.save_objects(save_objects_params)[0]
#motif_set_ref = "%s/%s/%s" % (info[6], info[0], info[4])
#object_upload_ret = dfu.file_to_shock()
reportName = 'GibbsMotifFinder_report_' + str(uuid.uuid4())
reportObj = {
'objects_created': [{
'ref': obj_ref,
'description': 'Motif Set generated by Gibbs'
}],
'message':
'',
'direct_html':
None,
'direct_html_link_index':
0,
'file_links': [],
'html_links': [],
'html_window_height':
220,
'workspace_name':
params['workspace_name'],
'report_object_name':
reportName
}
# attach to report obj
#reportObj['direct_html'] = None
reportObj['direct_html'] = ''
reportObj['direct_html_link_index'] = 0
reportObj['html_links'] = [{
'shock_id': html_upload_ret['shock_id'],
#'name': 'promoter_download.zip',
'name': 'index.html',
'label': 'Save promoter_download.zip'
}]
report = KBaseReport(self.callback_url, token=ctx['token'])
#report_info = report.create({'report':reportObj, 'workspace_name':input_params['input_ws']})
report_info = report.create_extended_report(reportObj)
output = {
'report_name': report_info['name'],
'report_ref': report_info['ref']
}
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" -> structure:
parameter "workspace_name" of String, parameter "fastapath" of
String, parameter "motif_min_length" of Long, parameter
"motif_max_length" of Long, parameter "SS_ref" of String,
parameter "obj_name" of String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
if 'motif_length' not in params:
params['motif_length'] = 8
motLen = params['motif_length']
promoterFastaFilePath = params['fastapath']
MDU=MdscanUtil()
MdscanMotifCommand = MDU.build_mdscan_motif_command(promoterFastaFilePath,motLen,params['background'])
MDU.run_mdscan_command(MdscanMotifCommand)
mdscan_out_path = '/kb/module/work/tmp/mdscan_out'
mdscan_params = {'ws_name' : params['workspace_name'], 'path' : mdscan_out_path,'obj_name' : params['obj_name']}
MOU = MotifUtils(self.callback_url)
dfu = DataFileUtil(self.callback_url)
locDict = {}
obj_ref = MDU.UploadFromMdscan(self.callback_url, mdscan_params)[0]['obj_ref']
MDU.write_obj_ref(mdscan_out_path, obj_ref)
timestamp = int((datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()*1000)
timestamp = str(timestamp)
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
lineCount = 0
with open(promoterFastaFilePath,'r') as pFile:
for line in pFile:
lineCount += 1
numFeat = lineCount/2
with open(promoterFastaFilePath,'r') as pFile:
fileStr = pFile.read()
promHtmlStr = '<html><body> ' + fileStr + ' </body></html>'
with open(htmlDir + '/promoters.html','w') as promHTML:
promHTML.write(promHtmlStr)
JsonPath = '/kb/module/work/tmp'
dfu = DataFileUtil(self.callback_url)
get_obj_params = {'object_refs' : [obj_ref]}
mdscanMotifSet = dfu.get_objects(get_obj_params)['data'][0]['data']
mr=MakeNewReport()
mr.MakeReport(htmlDir,mdscanMotifSet)
try:
html_upload_ret = dfu.file_to_shock({'file_path': htmlDir ,'make_handle': 0, 'pack': 'zip'})
except:
raise ValueError ('error uploading HTML file to shock')
reportName = 'MdscanMotifFinder_report_'+str(uuid.uuid4())
reportObj = {'objects_created': [{'ref' : obj_ref, 'description' : 'Motif Set generated by Mdscan'}],
'message': '',
'direct_html': None,
'direct_html_link_index': 0,
'file_links': [],
'html_links': [],
'html_window_height': 220,
'workspace_name': params['workspace_name'],
'report_object_name': reportName
}
# attach to report obj
reportObj['direct_html'] = ''
reportObj['direct_html_link_index'] = 0
reportObj['html_links'] = [{'shock_id': html_upload_ret['shock_id'],
#'name': 'promoter_download.zip',
'name': 'index.html',
'label': 'Save promoter_download.zip'
}
]
report = KBaseReport(self.callback_url, token=ctx['token'])
report_info = report.create_extended_report(reportObj)
output = { 'report_name': report_info['name'], 'report_ref': report_info['ref'] }
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" -> structure:
parameter "workspace_name" of String, parameter "fastapath" of
String, parameter "prb" of Double, parameter "motif_length" of
Long, parameter "obj_name" of String, parameter "mask_repeats" of
Long
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
'''if 'motif_min_length' not in params:
params['motif_min_length'] = 8
if 'motif_max_length' not in params:
params['motif_max_length'] = 16
motMin = params['motif_min_length']
motMax = params['motif_max_length']'''
if 'motif_length' not in params:
params['motif_length'] = 20
if 'prb' not in params:
params['prb'] = 0.05
motlen = params['motif_length']
prb = params['prb']
FastaFilePath = params['fastapath']
mfu = mfmdUtil()
mfmdMotifCommand = mfu.build_mfmd_command(FastaFilePath, motlen, prb,
self.config)
mfu.run_mfmd_command(mfmdMotifCommand)
mfmd_out_path = '/kb/module/work/tmp/mfmd_out'
mfmd_params = {
'ws_name': params['workspace_name'],
'path': mfmd_out_path,
'location_path': mfmd_out_path + '/mfmd_out.txt',
'obj_name': params['obj_name']
}
MOU = MotifUtils(self.callback_url)
dfu = DataFileUtil(self.callback_url)
locDict = {}
if 'SS_ref' in params:
get_ss_params = {'object_refs': [params['SS_ref']]}
SS = dfu.get_objects(get_ss_params)['data'][0]['data']
for s in SS['sequences']:
if s['source'] is not None:
locDict['sequence_id'] = {
'contig': s['source']['location'][0][0],
'start': str(s['source']['location'][0][1])
}
if len(locDict.keys()) > 0:
mfmd_params['absolute_locations'] = locDict
mfmd_params['motlen'] = motlen
mfmd_params['prb'] = prb
obj_ref = mfu.UploadFrommfmd(self.callback_url,
mfmd_params)[0]['obj_ref']
mfu.write_obj_ref(mfmd_out_path, obj_ref)
timestamp = int(
(datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()
* 1000)
timestamp = str(timestamp)
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
lineCount = 0
dfu = DataFileUtil(self.callback_url)
get_obj_params = {'object_refs': [obj_ref]}
mfmdMotifSet = dfu.get_objects(get_obj_params)['data'][0]['data']
mr = MakeNewReport()
mr.MakeReport(htmlDir, mfmdMotifSet)
try:
html_upload_ret = dfu.file_to_shock({
'file_path': htmlDir,
'make_handle': 0,
'pack': 'zip'
})
except:
raise ValueError('error uploading HTML file to shock')
reportName = 'mfmdMotifFinder_report_' + str(uuid.uuid4())
reportObj = {
'objects_created': [{
'ref': obj_ref,
'description': 'Motif Set generated by mfmd'
}],
'message':
'',
'direct_html':
None,
'direct_html_link_index':
0,
'file_links': [],
'html_links': [],
'html_window_height':
220,
'workspace_name':
params['workspace_name'],
'report_object_name':
reportName
}
# attach to report obj
reportObj['direct_html'] = ''
reportObj['direct_html_link_index'] = 0
reportObj['html_links'] = [{
'shock_id': html_upload_ret['shock_id'],
#'name': 'promoter_download.zip',
'name': 'index.html',
'label': 'Save promoter_download.zip'
}]
report = KBaseReport(self.callback_url, token=ctx['token'])
report_info = report.create_extended_report(reportObj)
output = {
'report_name': report_info['name'],
'report_ref': report_info['ref']
}
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
class MotifFinderGibbs:
'''
Module Name:
MotifFinderGibbs
Module Description:
A KBase module: MotifFinderGibbs
'''
######## WARNING FOR GEVENT USERS ####### noqa
# Since asynchronous IO can lead to methods - even the same method -
# interrupting each other, you must be *very* careful when using global
# state. A method could easily clobber the state set by another while
# the latter method is running.
######################################### noqa
VERSION = "0.0.1"
GIT_URL = ""
GIT_COMMIT_HASH = ""
#BEGIN_CLASS_HEADER
#END_CLASS_HEADER
# config contains contents of config file in a hash or None if it couldn't
# be found
def __init__(self, config):
#BEGIN_CONSTRUCTOR
self.callback_url = os.environ['SDK_CALLBACK_URL']
self.shared_folder = config['scratch']
self.SSU = SequenceSetUtils(os.environ['SDK_CALLBACK_URL'])
self.MOU = MotifUtils(os.environ['SDK_CALLBACK_URL'])
#self.HU = GibbsUtil(self.shared_folder)
self.dfu = DataFileUtil(self.callback_url)
self.GR = GenerateReport()
#END_CONSTRUCTOR
pass
def find_motifs(self, ctx, params):
"""
:param params: instance of type "find_motifs_params" (SS_ref -
optional, used for exact genome locations if possible) ->
structure: parameter "workspace_name" of String, parameter
"fastapath" of String, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long, parameter "SS_ref" of String
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN find_motifs
if 'motif_min_length' not in params:
params['motif_min_length'] = 8
if 'motif_max_length' not in params:
params['motif_max_length'] = 16
motMin = params['motif_min_length']
motMax = params['motif_max_length']
promoterFastaFilePath = params['fastapath']
GU=GibbsUtil()
gibbsCommandList = []
for i in range(motMin,motMax+1,2):
gibbsCommandList.append(GU.build_gibbs_command(promoterFastaFilePath,i))
for g in gibbsCommandList:
GU.run_gibbs_command(g)
gibbs_out_path = '/kb/module/work/tmp/gibbs'
#gibbs_params = {'ws_name' : params['workspace_name'], 'path' : gibbs_out_path,'obj_name' : params['obj_name']}
gibbs_params = {
'ws_name': params['workspace_name'],
'format': 'GIBBS',
'file': {'path': gibbs_out_path},
'obj_name': params['obj_name'],
'seq_set_ref': params['SS_ref']
}
#MOU = MotifUtils(self.callback_url)
#dfu = DataFileUtil(self.callback_url)
locDict = {}
if 'SS_ref' in params:
get_ss_params = {'object_refs' : [params['SS_ref']]}
SS = self.dfu.get_objects(get_ss_params)['data'][0]['data']
for s in SS['sequences']:
if s['source'] is not None:
locDict['sequence_id'] = {'contig' : s['source']['location'][0][0],'start':str(s['source']['location'][0][1])}
if len(locDict.keys()) > 0:
gibbs_params['absolute_locations'] = locDict
gibbs_params['min_len'] = motMin
gibbs_params['max_len'] = motMax
#obj_ref = MOU.UploadFromGibbs(gibbs_params)['obj_ref']
obj_ref = self.MOU.saveMotifSet(gibbs_params)
GU.write_obj_ref(gibbs_out_path, obj_ref)
timestamp = int((datetime.utcnow() - datetime.utcfromtimestamp(0)).total_seconds()*1000)
timestamp = str(timestamp)
htmlDir = self.shared_folder + '/html' + timestamp
os.mkdir(htmlDir)
lineCount = 0
with open(promoterFastaFilePath,'r') as pFile:
for line in pFile:
lineCount += 1
numFeat = lineCount/2
with open(promoterFastaFilePath,'r') as pFile:
fileStr = pFile.read()
promHtmlStr = '<html><body> ' + fileStr + ' </body></html>'
with open(htmlDir + '/promoters.html','w') as promHTML:
promHTML.write(promHtmlStr)
JsonPath = '/kb/module/work/tmp'
dfu = DataFileUtil(self.callback_url)
get_obj_params = {'object_refs' : [obj_ref]}
gibbsMotifSet = dfu.get_objects(get_obj_params)['data'][0]['data']
#g=GenerateReport()
GR.GenerateMotifReport(htmlDir,gibbsMotifSet)
try:
html_upload_ret = dfu.file_to_shock({'file_path': htmlDir ,'make_handle': 0, 'pack': 'zip'})
except:
raise ValueError ('error uploading HTML file to shock')
reportName = 'GibbsMotifFinder_report_'+str(uuid.uuid4())
reportObj = {'objects_created': [{'ref' : obj_ref, 'description' : 'Motif Set generated by Gibbs'}],
'message': '',
'direct_html': None,
'direct_html_link_index': 0,
'file_links': [],
'html_links': [],
'html_window_height': 220,
'workspace_name': params['workspace_name'],
'report_object_name': reportName
}
# attach to report obj
reportObj['direct_html'] = ''
reportObj['direct_html_link_index'] = 0
reportObj['html_links'] = [{'shock_id': html_upload_ret['shock_id'],
#'name': 'promoter_download.zip',
'name': 'index.html',
'label': 'Save promoter_download.zip'
}
]
report = KBaseReport(self.callback_url, token=ctx['token'])
report_info = report.create_extended_report(reportObj)
output = { 'report_name': report_info['name'], 'report_ref': report_info['ref'] }
#END find_motifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method find_motifs return value ' +
'output is not type dict as required.')
# return the results
return [output]
def BuildFastaFromSequenceSet(self, ctx, params):
"""
:param params: instance of type "BuildSeqIn" -> structure: parameter
"workspace_name" of String, parameter "SequenceSetRef" of String,
parameter "fasta_outpath" of String
:returns: instance of type "BuildSeqOut" -> structure: parameter
"fasta_outpath" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN BuildFastaFromSequenceSet
dfu = DataFileUtil(self.callback_url)
get_objects_params = {'object_refs' : [params['SequenceSetRef']]}
SeqSet = dfu.get_objects(get_objects_params)['data'][0]['data']
outFile = open(params['fasta_outpath'],'w')
for s in SeqSet['sequences']:
sname = '>' + s['sequence_id'] + '\n'
outFile.write(sname)
sseq = s['sequence'] + '\n'
outFile.write(sseq)
outFile.close()
output = {'fasta_outpath' : params['fasta_outpath']}
#END BuildFastaFromSequenceSet
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method BuildFastaFromSequenceSet return value ' +
'output is not type dict as required.')
# return the results
return [output]
def ExtractPromotersFromFeatureSetandDiscoverMotifs(self, ctx, params):
"""
:param params: instance of type "extract_input" -> structure:
parameter "workspace_name" of String, parameter "genome_ref" of
String, parameter "featureSet_ref" of String, parameter
"promoter_length" of Long, parameter "motif_min_length" of Long,
parameter "motif_max_length" of Long
:returns: instance of type "extract_output_params" -> structure:
parameter "report_name" of String, parameter "report_ref" of String
"""
# ctx is the context object
# return variables are: output
#BEGIN ExtractPromotersFromFeatureSetandDiscoverMotifs
SSU = SequenceSetUtils(self.callback_url)
BuildParams = {'ws_name' : params['workspace_name'], 'FeatureSet_ref' : params['featureSet_ref'], 'genome_ref' : params['genome_ref'], 'upstream_length' : params['promoter_length']}
SSret = SSU.buildFromFeatureSet(BuildParams)
SSref = SSret['SequenceSet_ref']
fastapath = '/kb/module/work/tmp/tmpSeqSet.fa'
FastaParams = {'workspace_name' : params['workspace_name'] , 'SequenceSetRef' : SSref , 'fasta_outpath' : fastapath}
output = self.BuildFastaFromSequenceSet(ctx,FastaParams)
newfastapath = '/kb/module/work/tmp/SeqSet.fa'
fu=FastaUtils()
fu.RemoveRepeats(fastapath,newfastapath)
findmotifsparams= {'workspace_name' : params['workspace_name'],'fastapath':fastapath,'motif_min_length':params['motif_min_length'],'motif_max_length':params['motif_max_length'],'SS_ref':SSref,'obj_name':params['obj_name']}
output = self.find_motifs(ctx,findmotifsparams)[0]
#END ExtractPromotersFromFeatureSetandDiscoverMotifs
# At some point might do deeper type checking...
if not isinstance(output, dict):
raise ValueError('Method ExtractPromotersFromFeatureSetandDiscoverMotifs return value ' +
'output is not type dict as required.')
# return the results
return [output]