the origin form of phrase already included in itertools.product()

also add lemmas with '-' replaced by ' '
xxyzz · Oct 23, 2021 · e9d0d0e · e9d0d0e
1 parent f6b6d50
commit e9d0d0e
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/data/dump_lemmas.py b/data/dump_lemmas.py
@@ -15,9 +15,6 @@
     if '(' in lemma:  # 'take (something) into account'
         continue
 
-    if '/' not in lemma:
-        keyword_processor.add_keyword(lemma, values)
-
     if ' ' in lemma:  # phrase, for example: 'slick back/down'
         list_of_inflections_list = []
         for word in lemma.split(' '):
@@ -32,9 +29,13 @@
                           product(*list_of_inflections_list)):
             keyword_processor.add_keyword(phrase, values)
     else:
+        keyword_processor.add_keyword(lemma, values)
         for inflection in filter(lambda x: x != lemma and x not in lemma,
                                  chain(*getAllInflections(lemma).values())):
             keyword_processor.add_keyword(inflection, values)
 
+    if '-' in lemma:
+        keyword_processor.add_keyword(lemma.replace('-', ' '), values)
+
 with open('lemmas_dump', 'wb') as f:
     pickle.dump(keyword_processor, f)
diff --git a/tests/LanguageLayer.en.json b/tests/LanguageLayer.en.json
@@ -1,7 +1,7 @@
 {
   "count": [
     [
-      9758
+      9779
     ]
   ],
   "glosses": [