using cld2-cffi instead of pycld2

.gitignore

@@ -99,3 +99,4 @@ ENV/
 
 # mypy
 .mypy_cache/
+.idea/

README.md

@@ -1,10 +1,17 @@
 # spaCy-CLD: Bringing simple language detection to spaCy
 
-This package is a [spaCy 2.0 extension](https://spacy.io/usage/processing-pipelines#section-extensions) that adds language detection to spaCy's text processing pipeline. Inspired from a discussion [here](https://github.com/explosion/spaCy/issues/1172).
+This package is a 
+[spaCy 2.0 extension](https://spacy.io/usage/processing-pipelines#section-extensions)
+that adds language detection to spaCy's text processing pipeline.
+Inspired from a discussion [here](https://github.com/explosion/spaCy/issues/1172).
 
 ## Installation
 
-`pip install spacy_cld`
+`python setup.py install`
+
+If you can't compile it, retry after
+
+`export CFLAGS="-Wno-narrowing"`
 
 ## Usage
 
@@ -23,10 +30,20 @@ doc._.languages  # ['en']
 doc._.language_scores['en']  # 0.96
 ```
 
-spaCy-CLD operates on `Doc` and `Span` spaCy objects. When called on a `Doc` or `Span`, the object is given two attributes: `languages` (a list of up to 3 language codes) and `language_scores` (a dictionary mapping language codes to confidence scores between 0 and 1).
+spaCy-CLD operates on `Doc` and `Span` spaCy objects. When called on a `Doc` or `Span`,
+the object is given two attributes: `languages` (a list of up to 3 language codes) 
+and `language_scores` (a dictionary mapping language codes to confidence scores between 
+0 and 1).
 
 ## Under the hood
 
-spacy-cld is a little extension that wraps the [PYCLD2](https://github.com/aboSamoor/pycld2) Python library, which in turn wraps the [Compact Language Detector 2](https://github.com/CLD2Owners/cld2) C library originally built at Google for the Chromium project. CLD2 uses character n-grams as features and a Naive Bayes classifier to identify 80+ languages from Unicode text strings (or XML/HTML). It can detect up to 3 different languages in a given document, and reports a confidence score (reported in with each language.
+spacy-cld is a little extension that wraps the 
+[CLD2-CFFI](https://github.com/GregBowyer/cld2-cffi) Python library, which in turn 
+wraps the [Compact Language Detector 2](https://github.com/CLD2Owners/cld2) 
+C++ library originally built at Google for the Chromium project.
+CLD2 uses character n-grams as features and a Naive Bayes classifier to identify 
+80+ languages from Unicode text strings (or XML/HTML). 
+It can detect up to 3 different languages in a given document, and reports 
+a confidence score (reported in with each language.
 
-For additional details, see the linked project pages for PYCLD2 and CLD2.
+For additional details, see the linked project pages for CLD2-CFFI and CLD2.

requirements.txt

@@ -1,2 +1,2 @@
+cld2-cffi>=0.1.4
 spacy>=2.0.0
-pycld2==0.31

setup.py

@@ -29,7 +29,7 @@ def setup_package():
         packages=find_packages(),
         install_requires=[
             'spacy>=2.0.0,<3.0.0',
-            'pycld2>=0.31'],
+            'cld2-cffi>=0.1.4'],
         zip_safe=False,
     )
 

spacy_cld/spacy_cld.py

@@ -1,4 +1,4 @@
-from pycld2 import detect, error as pycld_error
+from cld2 import detect
 from spacy.tokens import Doc, Span
 
 
@@ -19,7 +19,7 @@ def get_scores(text, cld_results=None):
 def detect_languages(text):
     try:
         _, _, results = detect(text.text)
-    except pycld_error as err:
+    except (ValueError, MemoryError):
         results = [[None, "error", 0.0, None]]
     return results