From c692b0d6433cd298d93360c04236355b3f107edc Mon Sep 17 00:00:00 2001
From: Robby Horvath <horvathrobby@gmail.com>
Date: Thu, 14 Oct 2021 06:35:14 -0400
Subject: [PATCH] Fix Bleu Score smoothing function from taking log(0) (#2839)

* Add conditional to prevent log(0)

* Add name to AUTHORS.md

* Add smoothin 4 test

Co-authored-by: Steven Bird <stevenbird1@gmail.com>
---
 AUTHORS.md                            |  1 +
 nltk/test/unit/translate/test_bleu.py | 10 ++++++++++
 nltk/translate/bleu_score.py          |  2 +-
 3 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/AUTHORS.md b/AUTHORS.md
index 49e99d05d1..2b4a88f98f 100644
--- a/AUTHORS.md
+++ b/AUTHORS.md
@@ -280,6 +280,7 @@
 - Hiroki Teranishi <https://github.com/chantera>
 - Ruben Cartuyvels <https://github.com/rubencart>
 - Dalton Pearson <https://github.com/daltonpearson>
+- Robby Horvath <https://github.com/robbyhorvath>
 - Gavish Poddar <https://github.com/gavishpoddar>
 - Saibo Geng <https://github.com/Saibo-creator>
 - Ahmet Yildirim <https://github.com/RnDevelover>
diff --git a/nltk/test/unit/translate/test_bleu.py b/nltk/test/unit/translate/test_bleu.py
index fa26df5823..e6afbd7dc0 100644
--- a/nltk/test/unit/translate/test_bleu.py
+++ b/nltk/test/unit/translate/test_bleu.py
@@ -181,6 +181,16 @@ def test_empty_hypothesis(self):
         hypothesis = []
         assert sentence_bleu(references, hypothesis) == 0
 
+    def test_length_one_hypothesis(self):
+        # Test case where there's hypothesis is of length 1 in Smoothing method 4.
+        references = ["The candidate has no alignment to any of the references".split()]
+        hypothesis = ["Foo"]
+        method4 = SmoothingFunction().method4
+        try:
+            sentence_bleu(references, hypothesis, smoothing_function=method4)
+        except ValueError:
+            pass  # unittest.TestCase.assertWarns is only supported in Python >= 3.2.
+
     def test_empty_references(self):
         # Test case where there's reference is empty.
         references = [[]]
diff --git a/nltk/translate/bleu_score.py b/nltk/translate/bleu_score.py
index 29e93a731e..b0d50e57d9 100644
--- a/nltk/translate/bleu_score.py
+++ b/nltk/translate/bleu_score.py
@@ -215,7 +215,7 @@ def corpus_bleu(
     p_n = smoothing_function(
         p_n, references=references, hypothesis=hypothesis, hyp_len=hyp_lengths
     )
-    s = (w_i * math.log(p_i) for w_i, p_i in zip(weights, p_n))
+    s = (w_i * math.log(p_i) for w_i, p_i in zip(weights, p_n) if p_i > 0)
     s = bp * math.exp(math.fsum(s))
     return s