def runMatrixJobs(outFname, datasets, wordListFname, posPmidFname, negPmidFname, \
skipMap, outFormat, onlyTest, docIdFname, posPmids=None, negPmids=None, runner=None):
""" run jobs to convert the articles to a bag-of-words matrix """
assert (outFormat in ["svml", "arff", "pmidsvml"])
if isinstance(datasets, basestring):
datasets = [datasets]
if runner == None:
runner = pubGeneric.makeClusterRunner(__file__)
logging.debug("pos and neg pmid fnames are: %s, %s" %
(posPmidFname, negPmidFname))
if posPmidFname != None:
posPmids = parsePmids(posPmidFname)
if negPmidFname != None:
negPmids = parsePmids(negPmidFname)
termList = parseTerms(wordListFname)
paramDict = {"termList" : termList, "posPmids" : posPmids, \
"negPmids" : negPmids, "outFormat" : outFormat }
paramDict["docIdOutFname"] = docIdFname
pubAlg.mapReduce(__file__+":MatrixMaker", datasets, paramDict, \
outFname, skipMap=skipMap, runTest=True, runner=runner, onlyTest=onlyTest)
def runMatrixJobs(outFname, datasets, wordListFname, posPmidFname, negPmidFname, \
skipMap, outFormat, onlyTest, docIdFname, posPmids=None, negPmids=None, runner=None):
""" run jobs to convert the articles to a bag-of-words matrix """
assert (outFormat in ["svml", "arff", "pmidsvml"])
if isinstance(datasets, basestring):
datasets = [datasets]
if runner==None:
runner = pubGeneric.makeClusterRunner(__file__)
logging.debug("pos and neg pmid fnames are: %s, %s" % (posPmidFname, negPmidFname))
if posPmidFname!=None:
posPmids = parsePmids(posPmidFname)
if negPmidFname!=None:
negPmids = parsePmids(negPmidFname)
termList = parseTerms(wordListFname)
paramDict = {"termList" : termList, "posPmids" : posPmids, \
"negPmids" : negPmids, "outFormat" : outFormat }
paramDict["docIdOutFname"] = docIdFname
pubAlg.mapReduce(__file__+":MatrixMaker", datasets, paramDict, \
outFname, skipMap=skipMap, runTest=True, runner=runner, onlyTest=onlyTest)
def getRunner(self, step):
" return a runner object for the current dataset and pipelineStep"
headNode = pubConf.stepHosts.get(step, None)
logging.debug("Headnode for step %s is %s" % (step, headNode))
return pubGeneric.makeClusterRunner("pubMap-" + self.dataset + "-" +
step,
headNode=headNode)
def submitJobs(inSpec, filterSpec, outDir):
inDirs = pubConf.resolveTextDirs(inSpec)
runner = pubGeneric.makeClusterRunner(__file__, maxJob=pubConf.convertMaxJob, algName=inSpec)
outFnames = []
for inDir in inDirs:
inFnames = glob.glob(join(inDir, "*.articles.gz"))
for inFname in inFnames:
outFname = join(outDir, basename(dirname(inFname))+"-"+basename(inFname))
outFnames.append(outFname)
outFnames.append(outFname.replace('.articles.gz','.files.gz'))
#command = "%s %s filterJob {check in exists %s} %s %s" % \
#(sys.executable, __file__, inFname, pmidFname, outFname)
runner.submitPythonFunc(__file__, "filterOneChunk", [inFname, filterSpec, outFname])
runner.finish(wait=True)
return outFnames
def submitJobs(inSpec, filterSpec, outDir):
inDirs = pubConf.resolveTextDirs(inSpec)
runner = pubGeneric.makeClusterRunner(__file__,
maxJob=pubConf.convertMaxJob,
algName=inSpec)
outFnames = []
for inDir in inDirs:
inFnames = glob.glob(join(inDir, "*.articles.gz"))
for inFname in inFnames:
outFname = join(
outDir,
basename(dirname(inFname)) + "-" + basename(inFname))
outFnames.append(outFname)
outFnames.append(outFname.replace('.articles.gz', '.files.gz'))
#command = "%s %s filterJob {check in exists %s} %s %s" % \
#(sys.executable, __file__, inFname, pmidFname, outFname)
runner.submitPythonFunc(__file__, "filterOneChunk",
[inFname, filterSpec, outFname])
runner.finish(wait=True)
return outFnames
def getRunner(self, step):
" return a runner object for the current dataset and pipelineStep"
headNode = pubConf.stepHosts.get(step, None)
logging.debug("Headnode for step %s is %s" % (step, headNode))
return pubGeneric.makeClusterRunner("pubMap-"+self.dataset+"-"+step, headNode=headNode)