ec60b6d78f0fc423a8158d9f42210b878a8bcf0e,src/gensim/corpora/svmlightcorpus.py,SvmLightCorpus,__iter__,#SvmLightCorpus#,58

Before Change

        
        Iterate over the corpus, returning one sparse vector at a time.
        
        for lineNo, line in enumerate(open(self.fname)):
            line = line[: line.find("//")].strip()
            if not line:
                continue // ignore comments and empty lines
            parts = line.split()
            if not parts:
                raise ValueError("invalid format at line no. %i in %s" %
                                 (lineNo, self.fname))
            target, fields = parts[0], [part.rsplit(":", 1) for part in parts[1:]]
            doc = [(int(p1) - 1, float(p2)) for p1, p2 in fields if p1 != "qid"] // ignore "qid" features, convert 1-based feature ids to 0-based
            yield doc
    

    @staticmethod
    def saveCorpus(fname, corpus, id2word = None):
        
        Save a corpus in the SVMlight format. 

After Change

        with open(self.fname) as fin:
            for lineNo, line in enumerate(fin):
                doc = self.line2doc(line)
                if doc is not None:
                    yield doc
                self.streamposition += len(line)
    
    
    def line2doc(self, line):

Instances