def PEATSAJobs(prjs, resubmit=False):
"""Submit PEATSA runs for all projects or merge results if done"""
for name in prjs:
print name
DB = PDatabase(local=os.path.join(savepath,name))
pdb = DB['wt'].Structure
PS = PEATSAPlugin()
PS.main(DB=DB)
if hasattr(DB.meta,'peatsa_jobs') and resubmit == False:
if 'mycalc' in DB.meta.peatsa_jobs:
print 'job is present'
#try to merge results
S = PEATTableModel(DB)
job,n = PS.getJob('mycalc')
PS.mergeResults(job, 'prediction', S)
DB.commit()
print 'merged results'
else:
mutlist = []
for p in DB.getRecs():
mutlist.append(DB.get(p).Mutations)
#print mutlist
pdbfile = PS.writetempPDB()
#we add source project data so exp data can be read from summary
prjdata = {'server':'enzyme.ucd.ie','username':'******',
'project':name,'password':'******','port':'8080'}
PS.submitJob(name='mycalc', pdbname=DB.meta.refprotein, pdbfile=pdbfile,
mutations=mutlist, calcs=['stability'],
meta={'protein':name,'expcol':'Exp','project':prjdata})
#required to end process
PS.jobManager.stopLogging()
DB.close()
return
def saveDBCopy(DB, filename, callback=None):
"""Save local copy of a remote or another local DB"""
import copy
total = len(DB.getRecs())
if filename == '' or filename == None:
return False
if os.path.exists(filename):
for i in ['.lock', '.index', '']:
os.remove(filename + i)
newDB = PDatabase(local=filename)
newDB = copyDB(DB, newDB)
newDB.commit()
newDB.close()
return newDB
def saveDBCopy(DB, filename, callback=None):
"""Save local copy of a remote or another local DB"""
import copy
total = len(DB.getRecs())
if filename == '' or filename == None:
return False
if os.path.exists(filename):
for i in ['.lock','.index','']:
os.remove(filename+i)
newDB = PDatabase(local=filename)
newDB = copyDB(DB, newDB)
newDB.commit()
newDB.close()
return newDB
def createProjects(files):
"""Create multiple projects at once from csv files"""
for filename in files:
print filename
name = os.path.splitext(filename)[0]
#create/open db
DB = PDatabase(local=os.path.join(savepath,name))
DB.add('wt')
#add wt pdb
stream = DBActions.fetchPDB(name)
DBActions.addPDBFile(DB, 'wt', pdbdata=stream, pdbname=name, gui=False)
DB.meta.refprotein = 'wt'
DB.meta.info['protein'] = name
#import data from csv
DB.importCSV(os.path.join(cpath,filename), namefield='Mutations')
print 'imported ok'
DB.deleteField('PDB')
DB.commit()
DB.close()
print 'done'
return
def importOldProj(datadir,local=None, server=None,
project=None, username=None):
"""Import old peat projects"""
import PEAT_DB.Database as peatDB
from PEAT_DB.PEAT_dict import PEAT_dict, sub_dict
import copy
if local != None:
newDB = PDatabase(local=local)
elif server != None:
newDB = PDatabase(server=server, username=username, port=8080,
password='******', project=project)
print newDB
PT = peatDB.Database(datadir, Tk=False)
oldDB = PT.DB
print 'got old peat_db with %s proteins' %len(PT.proteins)
print PT.DB.keys()
#import meta stuff like userfields, table
for p in newDB.meta.special:
if not p in PT.DB.keys():
continue
print 'adding',p
for k in PT.DB[p]:
newDB.meta[p][k] = copy.deepcopy(PT.DB[p][k])
newDB.meta._p_changed = 1
for p in PT.proteins:
if p in newDB.meta.special:
continue
name = oldDB[p]['Name']
rec = PEATRecord(name=name)
for col in oldDB[p].keys():
cdata = oldDB[p][col]
recdata = {}
if col == 'name':
cdata = oldDB[p]['Name']
if oldDB['userfields'].has_key(col) and oldDB['userfields'][col]['field_type'] in ekintypes:
E=EkinProject(data=cdata)
E.length = len(E.datasets)
if len(E.datasets)==0:
continue
cdata = E
if type(cdata) == sub_dict:
for k in cdata.keys():
recdata[k] = copy.deepcopy(cdata[k])
else:
recdata = cdata
if cdata != '' and cdata != None:
rec.addField(col, data=recdata)
newDB.add(p,rec)
print newDB.meta.userfields
#remove any file cols, too hard to import
for m in newDB.meta.userfields.keys()[:]:
if newDB.meta.userfields[m]['field_type'] == 'File':
newDB.deleteField(m)
newDB.commit(user='******', note='import')
newDB.close()
print 'import done'
return
def importOldProj(datadir,
local=None,
server=None,
project=None,
username=None):
"""Import old peat projects"""
import PEAT_DB.Database as peatDB
from PEAT_DB.PEAT_dict import PEAT_dict, sub_dict
import copy
if local != None:
newDB = PDatabase(local=local)
elif server != None:
newDB = PDatabase(server=server,
username=username,
port=8080,
password='******',
project=project)
print newDB
PT = peatDB.Database(datadir, Tk=False)
oldDB = PT.DB
print 'got old peat_db with %s proteins' % len(PT.proteins)
print PT.DB.keys()
#import meta stuff like userfields, table
for p in newDB.meta.special:
if not p in PT.DB.keys():
continue
print 'adding', p
for k in PT.DB[p]:
newDB.meta[p][k] = copy.deepcopy(PT.DB[p][k])
newDB.meta._p_changed = 1
for p in PT.proteins:
if p in newDB.meta.special:
continue
name = oldDB[p]['Name']
rec = PEATRecord(name=name)
for col in oldDB[p].keys():
cdata = oldDB[p][col]
recdata = {}
if col == 'name':
cdata = oldDB[p]['Name']
if oldDB['userfields'].has_key(col) and oldDB['userfields'][col][
'field_type'] in ekintypes:
E = EkinProject(data=cdata)
E.length = len(E.datasets)
if len(E.datasets) == 0:
continue
cdata = E
if type(cdata) == sub_dict:
for k in cdata.keys():
recdata[k] = copy.deepcopy(cdata[k])
else:
recdata = cdata
if cdata != '' and cdata != None:
rec.addField(col, data=recdata)
newDB.add(p, rec)
print newDB.meta.userfields
#remove any file cols, too hard to import
for m in newDB.meta.userfields.keys()[:]:
if newDB.meta.userfields[m]['field_type'] == 'File':
newDB.deleteField(m)
newDB.commit(user='******', note='import')
newDB.close()
print 'import done'
return
def summarise(projects):
summDB = PDatabase(local='summary.fs')
C = CorrelationAnalyser()
figs = []
for f in range(4):
figs.append(plt.figure())
gs = gridspec.GridSpec(5, 5, wspace=0.3, hspace=0.5)
i=0
data=[]
print 'processing %s projects' %len(projects)
for p in projects:
print 'structure:',p
DB = PDatabase(local=os.path.join(savepath,p))
S = PEATTableModel(DB)
try:
exp,pre = S.getColumns(['Exp','prediction'],allowempty=False)
errs = [j[0]-j[1] for j in zip(exp,pre)]
except:
print 'no results'
continue
#DB.close()
#add link to proj
summDB.add(p)
summDB.addField('project',fieldtype='Project')
summDB[p]['project'] = {'server':'enzyme.ucd.ie','username':'******',
'project':p,'password':'******','port':'8080'}
print summDB.isChanged()
#stats
cc,rmse,meanerr = C.getStats(pre,exp)
#ttest for mean errs 0
ttp = round(stats.ttest_1samp(errs, 0)[1],2)
#normality of errs
w,swp = C.ShapiroWilk(errs)
x={'name':p,'mutants':len(pre),'rmse':rmse,'corrcoef':cc,'meanerr':meanerr,
'ttest':ttp,'shapirowilk':swp}
'''ax = figs[0].add_subplot(gs[0, i])
C.plotCorrelation(pre,exp,title=p,ms=2,axeslabels=False,ax=ax)
ax = figs[1].add_subplot(gs[0, i])
C.showHistogram([pre,exp],title=p,labels=['pre','exp'],ax=ax)
ax = figs[2].add_subplot(gs[0, i])
C.plotNorm(errs,title=p,lw=1,ax=ax)
#qqplot
ax = figs[3].add_subplot(gs[0, i])
C.QQplot(errs,title=p,ax=ax)'''
#get PDB info
parser = PDBParser()
descr = parser.getDescription(p)
x.update(descr)
data.append(x)
i+=1
summDB.importDict(data)
print summDB.isChanged()
summDB.commit()
#add all peatsa jobs to summary proj also
'''print 'adding peatsa job info'
PS = PEATSAPlugin()
PS.main(DB=summDB)
#summDB.meta.peatsa_jobs = None
#from ZODB.PersistentMapping import PersistentMapping
#summDB.meta.peatsa_jobs = PersistentMapping()
PS.checkJobsDict()
PS.jobManager.stopLogging()
for p in projects:
#print summDB.meta
DB = PDatabase(local=os.path.join(savepath,p))
job = DB.meta.peatsa_jobs['mycalc']
summDB.meta.peatsa_jobs[p] = job
print job
#DB.close()
print summDB.isChanged()
print summDB.meta.peatsa_jobs
summDB.commit()'''
#for i in range(len(figs)):
# figs[i].savefig('fig%s.png' %i)
#plt.show()
return
