made spacy model configurable

app/__init__.py

@@ -7,6 +7,8 @@
 
 db = MongoEngine()
 
+spacy_tokenizer = None
+
 def create_app(env = 'Development'):
     app = Flask(__name__)
     CORS(app)
@@ -15,9 +17,14 @@ def create_app(env = 'Development'):
         env = os.environ['APPLICATION_ENV']
     except KeyError as e:
         app.logger.info('Unknown environment key, defaulting to Development')
+
     app.config.from_object('config.%s' % env)
     db.init_app(app)
 
+    import spacy
+    global spacy_tokenizer
+    spacy_tokenizer = spacy.load(app.config["SPACY_LANG_MODEL"])
+
     from app.agents.controllers import bots
     from app.nlu.controllers import nlu
     from app.intents.controllers import intents
@@ -50,6 +57,7 @@ def not_found(error):
     from app.endpoint.controllers import update_model
     with app.app_context():
         update_model()
+
     return app
 
 

app/endpoint/controllers.py

@@ -1,10 +1,8 @@
 # -*- coding: utf-8 -*-
 import copy
 import json
-
 from flask import Blueprint, request, abort
 from jinja2 import Template
-
 from flask import current_app as app
 from app.agents.models import Bot
 from app.commons import build_response
@@ -16,36 +14,64 @@
 from app.nlu.classifiers.sklearn_intent_classifer import \
     SklearnIntentClassifier
 from app.nlu.entity_extractor import EntityExtractor
-
 endpoint = Blueprint('api', __name__, url_prefix='/api')
-
 sentence_classifier = SklearnIntentClassifier()
 synonyms = None
 entity_extraction = None
 
-
-# Request Handler
 @endpoint.route('/v1', methods=['POST'])
 def api():
     """
-    Endpoint to converse with chatbot.
+    Endpoint to Converse with the Chatbot.
     Chat context is maintained by exchanging the payload between client and bot.
 
-    sample input/output payload =>
-
-    {
-      "currentNode": "",
-      "complete": false,
-      "parameters": [],
-      "extractedParameters": {},
-      "missingParameters": [],
-      "intent": {
-      },
-      "context": {},
-      "input": "hello",
-      "speechResponse": [
-      ]
-    }
+    ---
+    definitions:
+      currentNode:
+        type: string
+      complete:
+        type: boolean
+      context:
+        type: object
+        properties:
+      parameters:
+        type: array
+        items:
+          type: object
+          properties:
+            name:
+              type: string
+            type:
+              type: string
+            required:
+              type: boolean
+      extractedParameters:
+        type: object
+        properties:
+          country:
+            type: string
+      speechResponse:
+        type: array
+        items:
+          type: string
+      intent:
+        type: object
+        properties:
+          object_id:
+            type: string
+          confidence:
+            type: number
+          id:
+            type: string
+      input:
+        type: string
+      missingParameters:
+        type: array
+        items:
+      owner:
+        type: string
+      date:
+        type: string
 
     :param json:
     :return json:

app/nlu/__init__.py

@@ -1,2 +0,0 @@
-import spacy
-spacy_tokenizer = spacy.load("en")

app/nlu/classifiers/sklearn_intent_classifer.py

@@ -1,10 +1,7 @@
-import string
 import os
 import cloudpickle
-import spacy
 import numpy as np
-
-spacynlp = spacy.load('en')
+from app import spacy_tokenizer
 
 class SklearnIntentClassifier:
 
@@ -17,7 +14,7 @@ def get_spacy_embedding(self, sentence):
         :param sentence:
         :return list of clean tokens:
         """
-        spacy_obj = spacynlp(sentence)
+        spacy_obj = spacy_tokenizer(sentence)
         return np.array(spacy_obj.vector)
 
     def train(self, X, y, outpath=None, verbose=True):

app/nlu/entity_extractor.py

@@ -2,8 +2,7 @@
 
 import pycrfsuite
 from flask import current_app as app
-
-from app.nlu import spacy_tokenizer
+from app import spacy_tokenizer
 
 
 class EntityExtractor:

app/nlu/tasks.py

@@ -3,7 +3,7 @@
 from flask import current_app as app
 from app.endpoint.controllers import update_model
 from app.intents.models import Intent
-from app.nlu import spacy_tokenizer
+from app import spacy_tokenizer
 from app.nlu.classifiers.sklearn_intent_classifer import \
     SklearnIntentClassifier
 from app.nlu.entity_extractor import EntityExtractor