def rankScore(self, scorelist): ### Returns rank of scores as list
'''
Returns rank of scores as list.
>> scorelist = list of scores
<< ranklist:list of ranks (0 = Lowest, 1 = Highest)
'''
if self.opt['Rank']:
return rje.rankList(scorelist)
else:
return [0] * len(scorelist)
def rankScore(self,scorelist): ### Returns rank of scores as list
'''
Returns rank of scores as list.
>> scorelist = list of scores
<< ranklist:list of ranks (0 = Lowest, 1 = Highest)
'''
if self.opt['Rank']:
return rje.rankList(scorelist)
else:
return [0] * len(scorelist)
try:
### NewRanks ###
scores = []
objlist = objlist[0:]
for obj in objlist[0:]:
score = obj.getData(dkey,dlist,case,str=False,default=default,dp=-1)
if convert:
try:
score = float(score)
except:
score = default
if score == None:
objlist.remove(obj)
else:
scores.append(score)
newranks = rje.rankList(scores,rev,absolute,lowest)
### Rerank ###
rankdict = {}
for i in range(len(objlist)):
obj = objlist[i]
r = newranks[i]
if rankdict.has_key(r):
rankdict[r].append(obj)
else:
rankdict[r] = [obj]
if addstat:
obj.stat[addstat] = r
obj.dict['Data'][addstat] = r
newlist = []
for r in rje.sortKeys(rankdict):
def orderHits(self,seq,hits,hitseq): ### Returns orderd hits and fuller hitdict with GABLAM Stats etc.
'''
Returns orderd hits and fuller hitdict with GABLAM Stats etc.
>> seq:Query Sequence Object (name,sequence)
>> hits:List of hit entries
>> hitseq:Dictionary of {Hit:Sequence}
<< Tuple of (hits,hitdict)
'''
try:### ~ [0] Order hits by BLAST Rank ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
sorted = hits[0:]
for hit in hits: sorted[hit['Rank']-1] = hit
### ~ [1] Screen Species ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
hits = sorted
hits.reverse()
mapspec = None
qryspec = rje_sequence.specCodeFromName(seq[0])
if self.getStr('MapSpec').lower() == 'self': mapspec = qryspec
elif self.getStr('MapSpec').lower() not in ['any','none']: mapspec = self.getStr('MapSpec').upper()
for hit in hits[0:]:
#self.debug(hit)
#self.debug('%s vs %s' % (rje_sequence.specCodeFromName(hitseq[hit['Hit']][0]), mapspec))
if mapspec and rje_sequence.specCodeFromName(hitseq[hit['Hit']][0]) != mapspec:
hits.remove(hit)
hitseq.pop(hit['Hit'])
elif not mapspec and rje_sequence.specCodeFromName(hitseq[hit['Hit']][0]) == qryspec:
hits.remove(hit)
hits.append(hit) # => Move hits from same species to top of list
hits.reverse()
### ~ [2] Convert GABLAM Stats to Percentages and move to hitdict ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
self.debug(self.db())
gstat = 'GABLAMO'
if not self.bool['Ordered']: gstat = 'GABLAM'
hitdict = {}
for hit in hits[0:]:
hitname = hit['Hit']
hitdict[hitname] = {'Seq':hitseq[hitname],'MapRank':hits.index(hit)+1,'EVal':hit['E-Value'],'Score':hit['BitScore']}
hitgab = self.db('GABLAM').data(self.db('GABLAM').makeKey({'Query':hit['Query'],'Hit':hitname,'QryHit':'Hit'}))
qrygab = self.db('GABLAM').data(self.db('GABLAM').makeKey({'Query':hit['Query'],'Hit':hitname,'QryHit':'Query'}))
for st in ['ID','Len','Sim']:
hitdict[hitname]['Query_%s' % st] = 100 * float(qrygab['%s %s' % (gstat,st)]) / len(seq[1])
hitdict[hitname]['Hit_%s' % st] = 100 * float(hitgab['%s %s' % (gstat,st)]) / hit['Length']
hitdict[hitname]['Both_%s' % st] = (hitdict[hitname]['Query_%s' % st] + hitdict[hitname]['Hit_%s' % st]) / 2.0
if hitdict[hitname]['Query_%s' % st] >= hitdict[hitname]['Hit_%s' % st]:
hitdict[hitname]['Either_%s' % st] = hitdict[hitname]['Query_%s' % st]
else:
hitdict[hitname]['Either_%s' % st] = hitdict[hitname]['Hit_%s' % st]
### ~ [3] Rank Hits ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
focus = self.str['MapFocus'][:1].upper() + self.str['MapFocus'][1:].lower()
scorelist = []
multorder = ['Both_Len','Both_Sim','Both_ID','%s_Len' % focus,'%s_Sim' % focus,'%s_ID' % focus]
for hit in hits:
hitname = hit['Hit']
s = 0
m = 1
for mult in multorder:
s += (hitdict[hitname][mult] * m)
m *= 1000
scorelist.append(s)
ranks = rje.rankList(scorelist,rev=True,absolute=True,lowest=True) ### Returns rank of scores as list
rankdict = {}
for h in range(len(hits)):
hit = hits[h]
hitname = hit['Hit']
r = ranks[h]
hitdict[hitname]['HitRank'] = r
if rankdict.has_key(r): rankdict[r].append(hit)
else: rankdict[r] = [hit]
newlist = []
for r in rje.sortKeys(rankdict): newlist += rankdict[r]
return (newlist,hitdict)
except: self.errorLog('Bad goings on in SeqMapper.orderHits()',quitchoice=True)
def orderHits(self,seq,hits,hitseq): ### Returns orderd hits and fuller hitdict with GABLAM Stats etc.
'''
Returns orderd hits and fuller hitdict with GABLAM Stats etc.
>> seq:Query Sequence Object (name,sequence)
>> hits:List of hit entries
>> hitseq:Dictionary of {Hit:Sequence}
<< Tuple of (hits,hitdict)
'''
try:### ~ [0] Order hits by BLAST Rank ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
sorted = hits[0:]
for hit in hits: sorted[hit['Rank']-1] = hit
### ~ [1] Screen Species ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
hits = sorted
hits.reverse()
mapspec = None
qryspec = rje_sequence.specCodeFromName(seq[0])
if self.getStr('MapSpec').lower() == 'self': mapspec = qryspec
elif self.getStr('MapSpec').lower() not in ['any','none']: mapspec = self.getStr('MapSpec').upper()
for hit in hits[0:]:
#self.debug(hit)
#self.debug('%s vs %s' % (rje_sequence.specCodeFromName(hitseq[hit['Hit']][0]), mapspec))
if mapspec and rje_sequence.specCodeFromName(hitseq[hit['Hit']][0]) != mapspec:
hits.remove(hit)
hitseq.pop(hit['Hit'])
elif not mapspec and rje_sequence.specCodeFromName(hitseq[hit['Hit']][0]) == qryspec:
hits.remove(hit)
hits.append(hit) # => Move hits from same species to top of list
hits.reverse()
### ~ [2] Convert GABLAM Stats to Percentages and move to hitdict ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
self.debug(self.db())
gstat = 'GABLAMO'
if not self.bool['Ordered']: gstat = 'GABLAM'
hitdict = {}
for hit in hits[0:]:
hitname = hit['Hit']
hitdict[hitname] = {'Seq':hitseq[hitname],'MapRank':hits.index(hit)+1,'EVal':hit['E-Value'],'Score':hit['BitScore']}
hitgab = self.db('GABLAM').data(self.db('GABLAM').makeKey({'Query':hit['Query'],'Hit':hitname,'QryHit':'Hit'}))
qrygab = self.db('GABLAM').data(self.db('GABLAM').makeKey({'Query':hit['Query'],'Hit':hitname,'QryHit':'Query'}))
for st in ['ID','Len','Sim']:
hitdict[hitname]['Query_%s' % st] = 100 * float(qrygab['%s %s' % (gstat,st)]) / len(seq[1])
hitdict[hitname]['Hit_%s' % st] = 100 * float(hitgab['%s %s' % (gstat,st)]) / hit['Length']
hitdict[hitname]['Both_%s' % st] = (hitdict[hitname]['Query_%s' % st] + hitdict[hitname]['Hit_%s' % st]) / 2.0
if hitdict[hitname]['Query_%s' % st] >= hitdict[hitname]['Hit_%s' % st]:
hitdict[hitname]['Either_%s' % st] = hitdict[hitname]['Query_%s' % st]
else:
hitdict[hitname]['Either_%s' % st] = hitdict[hitname]['Hit_%s' % st]
### ~ [3] Rank Hits ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ###
focus = self.str['MapFocus'][:1].upper() + self.str['MapFocus'][1:].lower()
scorelist = []
multorder = ['Both_Len','Both_Sim','Both_ID','%s_Len' % focus,'%s_Sim' % focus,'%s_ID' % focus]
for hit in hits:
hitname = hit['Hit']
s = 0
m = 1
for mult in multorder:
s += (hitdict[hitname][mult] * m)
m *= 1000
scorelist.append(s)
ranks = rje.rankList(scorelist,rev=True,absolute=True,lowest=True) ### Returns rank of scores as list
rankdict = {}
for h in range(len(hits)):
hit = hits[h]
hitname = hit['Hit']
r = ranks[h]
hitdict[hitname]['HitRank'] = r
if rankdict.has_key(r): rankdict[r].append(hit)
else: rankdict[r] = [hit]
newlist = []
for r in rje.sortKeys(rankdict): newlist += rankdict[r]
return (newlist,hitdict)
except: self.errorLog('Bad goings on in SeqMapper.orderHits()',quitchoice=True)
def rankObj(
callobj,
objlist,
dkey,
dlist=["stat"],
case=False,
default=None,
rev=True,
absolute=True,
lowest=True,
addstat="Rank",
cutoff=0,
convert=True,
): ### Ranks objects using numerical data
"""
Ranks objects using data in object and rje.getData()
>> objlist:list of objects to be ranked, ordered and returned
>> dkey:str = Key for dictionaries
>> dlist:list [self.stat] = list of dictionaries to try after dict['Data']
>> case:bool [False] = whether to match case for dkey
>> default [None] = what value to give entry if no dictionary has key (if None, will not be returned in ranked list)
>> rev:bool [True] = whether high values should be ranked number 1
>> absolute:boolean [True] = return 1 to n, rather than 0 to 1
>> lowest:boolean [True] = returns lowest rank rather mean rank in case of ties
>> addstat:str ['Rank'] = key for callobj.stat to add rank to (will not add if None)
>> cutoff:int [0] = only returns top X ranked motifs (if given) (Can be float if absolute=False)
>> convert:bool [True] = convert returned data into numeric
<< returns list of ranked objects
"""
try:
### NewRanks ###
scores = []
objlist = objlist[0:]
for obj in objlist[0:]:
score = obj.getData(dkey, dlist, case, str=False, default=default, dp=-1)
if convert:
try:
score = float(score)
except:
score = default
if score == None:
objlist.remove(obj)
else:
scores.append(score)
newranks = rje.rankList(scores, rev, absolute, lowest)
### Rerank ###
rankdict = {}
for i in range(len(objlist)):
obj = objlist[i]
r = newranks[i]
if rankdict.has_key(r):
rankdict[r].append(obj)
else:
rankdict[r] = [obj]
if addstat:
obj.stat[addstat] = r
obj.dict["Data"][addstat] = r
newlist = []
for r in rje.sortKeys(rankdict):
if cutoff <= 0 or r <= cutoff:
newlist += rankdict[r]
### Finish ###
return newlist
except:
callobj.log.errorLog("Problem during rje_scoring.rankObj()", quitchoice=True)
return objlist[0:]