From 24ed3fc25c9350b8e3a4342a40fcc9f40ce8ba59 Mon Sep 17 00:00:00 2001
From: Matthew Honnibal <honnibal@gmail.com>
Date: Sun, 13 Sep 2015 10:45:21 +1000
Subject: [PATCH] * Check file existance before opening in lemmatizer

---
 spacy/lemmatizer.py | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/spacy/lemmatizer.py b/spacy/lemmatizer.py
index e2626b4d3..ed04e2d77 100644
--- a/spacy/lemmatizer.py
+++ b/spacy/lemmatizer.py
@@ -18,7 +18,10 @@ class Lemmatizer(object):
         for pos in ['adj', 'noun', 'verb']:
             index[pos] = read_index(path.join(data_dir, 'wordnet', 'index.%s' % pos))
             exc[pos] = read_exc(path.join(data_dir, 'wordnet', '%s.exc' % pos))
-        rules = json.load(open(path.join(data_dir, 'vocab', 'lemma_rules.json')))
+        if path.exists(path.join(data_dir, 'vocab', 'lemma_rules.json')):
+            rules = json.load(open(path.join(data_dir, 'vocab', 'lemma_rules.json')))
+        else:
+            rules = {}
         return cls(index, exc, rules)
 
     def __init__(self, index, exceptions, rules):
@@ -64,6 +67,8 @@ def lemmatize(string, index, exceptions, rules):
 
 def read_index(loc):
     index = set()
+    if not path.exists(loc):
+        return index
     for line in codecs.open(loc, 'r', 'utf8'):
         if line.startswith(' '):
             continue
@@ -76,6 +81,8 @@ def read_index(loc):
 
 def read_exc(loc):
     exceptions = {}
+    if not path.exists(loc):
+        return exceptions
     for line in codecs.open(loc, 'r', 'utf8'):
         if line.startswith(' '):
             continue