Lazy imports language

spacy/__init__.py

@@ -3,18 +3,13 @@ from __future__ import unicode_literals
 
 from . import util
 from .deprecated import resolve_model_name
+import importlib
 from .cli.info import info
 
-from . import en, de, zh, es, it, hu, fr, pt, nl, sv, fi, bn, he, nb
 
-
+_languages_name = set(["en", "de", "es", "pt", "fr",
-_languages = (en.English, de.German, es.Spanish, pt.Portuguese, fr.French,
+             "it", "hu", "zh", "nl", "sv",
-             it.Italian, hu.Hungarian, zh.Chinese, nl.Dutch, sv.Swedish,
+             "fi", "bn", "he", "nb"])
-             fi.Finnish, bn.Bengali, he.Hebrew, nb.Norwegian)
-
-
-for _lang in _languages:
-    util.set_lang_class(_lang.lang, _lang)
 
 
 def load(name, **overrides):
@@ -34,7 +29,7 @@ def load(name, **overrides):
         model_name = ''
     meta = util.parse_package_meta(data_path, model_name, require=False)
     lang = meta['lang'] if meta and 'lang' in meta else name
-    cls = util.get_lang_class(lang)
+    cls = importlib.import_module("."+lang, "spacy")
     overrides['meta'] = meta
     overrides['path'] = model_path
-    return cls(**overrides)
+    return cls.EXPORT(**overrides)

spacy/bn/__init__.py

@@ -22,3 +22,5 @@ class Bengali(Language):
         prefixes = tuple(TOKENIZER_PREFIXES)
         suffixes = tuple(TOKENIZER_SUFFIXES)
         infixes = tuple(TOKENIZER_INFIXES)
+
+EXPORT = Bengali

spacy/de/__init__.py

@@ -20,3 +20,6 @@ class German(Language):
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         tag_map = TAG_MAP
         stop_words = STOP_WORDS
+
+
+EXPORT = German

spacy/en/__init__.py

@@ -32,3 +32,6 @@ class English(Language):
         # Special-case hack for loading the GloVe vectors, to support <1.0
         overrides = fix_glove_vectors_loading(overrides)
         Language.__init__(self, **overrides)
+
+
+EXPORT = English

spacy/es/__init__.py

@@ -19,3 +19,6 @@ class Spanish(Language):
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         tag_map = TAG_MAP
         stop_words = STOP_WORDS
+
+
+EXPORT = Spanish

spacy/fi/__init__.py

@@ -15,3 +15,6 @@ class Finnish(Language):
 
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         stop_words = STOP_WORDS
+
+
+EXPORT = Finnish

spacy/fr/__init__.py

@@ -31,4 +31,7 @@ class FrenchDefaults(BaseDefaults):
 class French(Language):
     lang = 'fr'
 
-    Defaults = FrenchDefaults
+    Defaults = FrenchDefaults
+
+
+EXPORT = French

spacy/he/__init__.py

@@ -16,3 +16,6 @@ class Hebrew(Language):
 
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         stop_words = STOP_WORDS
+
+
+EXPORT = Hebrew

spacy/hu/__init__.py

@@ -24,3 +24,6 @@ class Hungarian(Language):
         stop_words = set(STOP_WORDS)
 
         token_match = TOKEN_MATCH
+
+
+EXPORT = Hungarian

spacy/it/__init__.py

@@ -16,3 +16,6 @@ class Italian(Language):
 
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         stop_words = STOP_WORDS
+
+
+EXPORT = Italian

spacy/language_data/tag_map.py

@@ -22,5 +22,6 @@ TAG_MAP = {
     "CCONJ":    {POS: CCONJ}, # U20
     "ADJ":      {POS: ADJ},
     "VERB":     {POS: VERB},
-    "PART":     {POS: PART}
+    "PART":     {POS: PART},
+	"SP":     	{POS: SPACE}
 }

spacy/nb/__init__.py

@@ -23,3 +23,6 @@ class Norwegian(Language):
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         #tag_map = TAG_MAP
         stop_words = STOP_WORDS
+
+
+EXPORT = Norwegian

spacy/nl/__init__.py

@@ -15,3 +15,6 @@ class Dutch(Language):
 
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         stop_words = STOP_WORDS
+
+
+EXPORT = Dutch

spacy/pt/__init__.py

@@ -16,3 +16,6 @@ class Portuguese(Language):
 
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         stop_words = STOP_WORDS
+
+
+EXPORT = Portuguese

spacy/sv/__init__.py

@@ -15,3 +15,6 @@ class Swedish(Language):
 
         tokenizer_exceptions = TOKENIZER_EXCEPTIONS
         stop_words = STOP_WORDS
+
+
+EXPORT = Swedish

spacy/zh/__init__.py

@@ -9,3 +9,6 @@ class Chinese(Language):
         import jieba
         words = list(jieba.cut(text, cut_all=True))
         return Doc(self.vocab, words=words, spaces=[False]*len(words))
+
+
+EXPORT = Chinese