SUMPY: a python automatic text summarization library
We currently have several baseline summarizers implemented:
[x] lede
[x] TextRank
[x] LexRank
[x] Centroid
[x] ROUGE ngram evaluation
and have plans to implement many more.
TODO:
[ ] ROUGE skip-gram and LCS evaluation
[ ] FreqSum/SumBasic
[ ] Submodular optimization based summarizers
[ ] lda/distributes sentence representation based summarizer
[ ] DEMS
[ ] ILP based summarizers
[ ] collect topic signatures/important word lists
SUMPY contains several ready to use summarizers with sensible defaults. Here is a simple example to get you started:
import sumpy
doc1 = "This is the text for document1. It is for explanatory purposes..."
doc2 = "This is another document text..."
doc3 = "And yet another document..."
inputs = [doc1, doc2, doc3]
print "lede summarizer:"
print sumpy.lede(inputs)
print "\ntextrank summarizer:"
print sumpy.textrank(inputs)
print "\ncentroid summarizer:"
print sumpy.centroid(inputs)