Merge branch 'master' into develop

CONTRIBUTORS.md

@@ -4,6 +4,7 @@ This is a list of everyone who has made significant contributions to spaCy, in a
 
 * Adam Bittlingmayer, [@bittlingmayer](https://github.com/bittlingmayer)
 * Andreas Grivas, [@andreasgrv](https://github.com/andreasgrv)
+* Andrew Poliakov, [@pavlin99th](https://github.com/pavlin99th)
 * Aniruddha Adhikary [@aniruddha-adhikary](https://github.com/aniruddha-adhikary)
 * Bhargav Srinivasa, [@bhargavvader](https://github.com/bhargavvader)
 * Chris DuBois, [@chrisdubois](https://github.com/chrisdubois)
@@ -11,12 +12,16 @@ This is a list of everyone who has made significant contributions to spaCy, in a
 * Dafne van Kuppevelt, [@dafnevk](https://github.com/dafnevk)
 * Daniel Rapp, [@rappdw](https://github.com/rappdw)
 * Dmytro Sadovnychyi, [@sadovnychyi](https://github.com/sadovnychyi)
+* Eric Zhao, [@ericzhao28](https://github.com/ericzhao28)
+* Greg Baker, [@solresol](https://github.com/solresol)
 * György Orosz, [@oroszgy](https://github.com/oroszgy)
 * Henning Peters, [@henningpeters](https://github.com/henningpeters)
+* Iddo Berger, [@iddoberger](https://github.com/iddoberger)
 * Ines Montani, [@ines](https://github.com/ines)
 * J Nicolas Schrading, [@NSchrading](https://github.com/NSchrading)
 * Janneke van der Zwaan, [@jvdzwaan](https://github.com/jvdzwaan)
 * Jordan Suchow, [@suchow](https://github.com/suchow)
+* Juan Miguel Cejuela, [@juanmirocks](https://github.com/juanmirocks)
 * Kendrick Tan, [@kendricktan](https://github.com/kendricktan)
 * Kyle P. Johnson, [@kylepjohnson](https://github.com/kylepjohnson)
 * Liling Tan, [@alvations](https://github.com/alvations)
@@ -29,6 +34,7 @@ This is a list of everyone who has made significant contributions to spaCy, in a
 * Pokey Rule, [@pokey](https://github.com/pokey)
 * Raphaël Bournhonesque, [@raphael0202](https://github.com/raphael0202)
 * Rob van Nieuwpoort, [@RvanNieuwpoort](https://github.com/RvanNieuwpoort)
+* Roman Inflianskas, [@rominf](https://github.com/rominf)
 * Sam Bozek, [@sambozek](https://github.com/sambozek)
 * Sasho Savkov, [@savkov](https://github.com/savkov)
 * Shuvanon Razik, [@shuvanon](https://github.com/shuvanon)

README.rst

@@ -5,8 +5,8 @@ spaCy is a library for advanced natural language processing in Python and
 Cython. spaCy is built on  the very latest research, but it isn't researchware.
 It was designed from day one to be used in real products. spaCy currently supports
 English and German,  as well as tokenization for Chinese, Spanish, Italian, French,
-Portuguese, Dutch, Swedish, Finnish, Hungarian and Bengali. It's commercial  open-source
-software, released under the MIT license.
+Portuguese, Dutch, Swedish, Finnish, Hungarian, Bengali and Hebrew. It's commercial 
+open-source software, released under the MIT license.
 
 💫 **Version 1.7 out now!** `Read the release notes here. <https://github.com/explosion/spaCy/releases/>`_
 
@@ -316,6 +316,7 @@ and ``--model`` are optional and enable additional tests:
 =========== ============== ===========
 Version     Date           Description
 =========== ============== ===========
+`v1.7.3`_   ``2017-03-26`` Alpha support for Hebrew, new CLI commands and bug fixes
 `v1.7.2`_   ``2017-03-20`` Small fixes to beam parser and model linking
 `v1.7.1`_   ``2017-03-19`` Fix data download for system installation
 `v1.7.0`_   ``2017-03-18`` New 50 MB model, CLI, better downloads and lots of bug fixes
@@ -344,6 +345,7 @@ Version     Date           Description
 `v0.93`_    ``2015-09-22`` Bug fixes to word vectors
 =========== ============== ===========
 
+.. _v1.7.3: https://github.com/explosion/spaCy/releases/tag/v1.7.3
 .. _v1.7.2: https://github.com/explosion/spaCy/releases/tag/v1.7.2
 .. _v1.7.1: https://github.com/explosion/spaCy/releases/tag/v1.7.1
 .. _v1.7.0: https://github.com/explosion/spaCy/releases/tag/v1.7.0

spacy/__main__.py

@@ -14,7 +14,7 @@ from spacy.cli import train as cli_train
 class CLI(object):
     """Command-line interface for spaCy"""
 
-    commands = ('download', 'link', 'info', 'package', 'train', 'train_config')
+    commands = ('download', 'link', 'info', 'package', 'train')
 
     @plac.annotations(
         model=("model to download (shortcut or model name)", "positional", None, str),

spacy/about.py

@@ -3,7 +3,7 @@
 # https://github.com/pypa/warehouse/blob/master/warehouse/__about__.py
 
 __title__ = 'spacy'
-__version__ = '1.7.2'
+__version__ = '1.7.3'
 __summary__ = 'Industrial-strength Natural Language Processing (NLP) with Python and Cython'
 __uri__ = 'https://spacy.io'
 __author__ = 'Matthew Honnibal'

spacy/cli/train.py

@@ -93,7 +93,7 @@ def evaluate(Language, gold_tuples, output_path):
 def check_dirs(output_path, train_path, dev_path):
     if not output_path.exists():
         util.sys_exit(output_path.as_posix(), title="Output directory not found")
-    if not train_path.exists() or not train_path.is_file():
+    if not train_path.exists():
         util.sys_exit(train_path.as_posix(), title="Training data not found")
     if dev_path and not dev_path.exists():
         util.sys_exit(dev_path.as_posix(), title="Development data not found")

spacy/language.py

@@ -278,7 +278,8 @@ class Language(object):
             path = pathlib.Path(path)
         if path is True:
             path = util.get_data_path() / self.lang
-
+            if not path.exists() and 'path' not in overrides:
+                path = None
         self.meta = overrides.get('meta', {})
         self.path = path
 

spacy/syntax/parser.pyx

@@ -52,7 +52,7 @@ from ._parse_features cimport fill_context
 from .stateclass cimport StateClass
 from ._state cimport StateC
 
-USE_FTRL = False
+USE_FTRL = True
 DEBUG = False
 def set_debug(val):
     global DEBUG
@@ -82,14 +82,19 @@ cdef class ParserModel(AveragedPerceptron):
     def update(self, Example eg, itn=0):
         '''Does regression on negative cost. Sort of cute?'''
         self.time += 1
-        best = arg_max_if_gold(eg.c.scores, eg.c.costs, eg.c.nr_class)
-        guess = eg.guess
+        cdef int best = arg_max_if_gold(eg.c.scores, eg.c.costs, eg.c.nr_class)
+        cdef int guess = eg.guess
         if guess == best or best == -1:
             return 0.0
+        cdef FeatureC feat
+        cdef int clas
+        cdef weight_t gradient
         if USE_FTRL:
             for feat in eg.c.features[:eg.c.nr_feat]:
-                self.update_weight_ftrl(feat.key, guess, feat.value * eg.c.costs[guess])
-                self.update_weight_ftrl(feat.key, best, -feat.value * eg.c.costs[guess])
+                for clas in range(eg.c.nr_class):
+                    if eg.c.is_valid[clas] and eg.c.scores[clas] >= eg.c.scores[best]:
+                        gradient = eg.c.scores[clas] + eg.c.costs[clas]
+                        self.update_weight_ftrl(feat.key, clas, feat.value * gradient)
         else:
             for feat in eg.c.features[:eg.c.nr_feat]:
                 self.update_weight(feat.key, guess, feat.value * eg.c.costs[guess])