Merge pull request #680 from savvopoulos/train-ner-update

actually commit load_ner.py

examples/training/load_ner.py

@@ -0,0 +1,22 @@
+# Load NER
+from __future__ import unicode_literals
+import spacy
+import pathlib
+from spacy.pipeline import EntityRecognizer
+from spacy.vocab import Vocab
+
+def load_model(model_dir):
+    model_dir = pathlib.Path(model_dir)
+    nlp = spacy.load('en', parser=False, entity=False, add_vectors=False)
+    with (model_dir / 'vocab' / 'strings.json').open('r', encoding='utf8') as file_:
+        nlp.vocab.strings.load(file_)
+    nlp.vocab.load_lexemes(model_dir / 'vocab' / 'lexemes.bin')
+    ner = EntityRecognizer.load(model_dir, nlp.vocab, require=True)
+    return (nlp, ner)
+
+(nlp, ner) = load_model('ner')
+doc = nlp.make_doc('Who is Shaka Khan?')
+nlp.tagger(doc)
+ner(doc)
+for word in doc:
+    print(word.text, word.orth, word.lower, word.tag_, word.ent_type_, word.ent_iob)