[TO TEST] lang_detection

gargantext/util/languages.py

 from gargantext.constants import *
-from langdetect import detect
-from langdetect import DetectorFactory
-
+from langdetect import detect, DetectorFactory
 
 class Language:
     def __init__(self, iso2=None, iso3=None,full_name=None, name=None):
@@ -18,9 +16,6 @@ class Language:
         return result
     __repr__ = __str__
 
-def detect_lang(self, text):
-    DetectorFactory.seed = 0
-    return Languages[detect(text)].iso2
 
 class Languages(dict):
     def __missing__(self, key):
@@ -30,6 +25,9 @@ class Languages(dict):
         raise KeyError
 
 languages = Languages()
+def detect_lang(self, text):
+    DetectorFactory.seed = 0
+    return languages[detect(text)].iso2
 
 import pycountry
 pycountry_keys = (

gargantext/util/toolchain/parsing.py

@@ -47,7 +47,7 @@ def parse(corpus):
                     indexed = False
                     # a simple census to raise language info at corpus level
                     for l in ["iso2", "iso3", "full_name"]:
-                        if hyperdata["indexed"] is True:
+                        if indexed is True:
                             break
                         lang_field = "language_"+l
                         if lang_field in hyperdata.keys():

install/gargamelle/requirements.txt

@@ -14,6 +14,7 @@ html5lib==0.9999999
 python-igraph>=0.7.1
 jdatetime==1.7.2
 kombu==3.0.33                  # messaging
+langdetect==1.0.6              #detectinglanguage
 nltk==3.1
 numpy==1.10.4
 psycopg2==2.6.1