beta.2

.dockerignore

@@ -0,0 +1,3 @@
+examples/
+.gitignore
+README.md

Dockerfile

@@ -1,15 +1,22 @@
-FROM python:3.4.5
+FROM python:3.5.2
 MAINTAINER Open Semantics Group <[email protected]>
 
-WORKDIR /root
-ADD . ./
+# update old version
+RUN pip list --local --outdated --format=freeze | cut -d= -f 1 | xargs -n1 pip install -U
+
+WORKDIR /app
+ADD requirements.txt /app/
 
 # Project requirements
 RUN pip install -r requirements.txt
 # .. plus some NLTK corpora
 RUN python -m nltk.downloader names punkt
+
 # Deployment requirements
 RUN pip install 'uwsgi==2.0.14'
 
+# finally add application
+ADD src/ /app/
+
 # Go!
 CMD [ "uwsgi", "--ini", "uwsgi.ini" ]

README.md

@@ -8,57 +8,64 @@ Useful hints that can save you some time.
 
 ## Python Interpreter
 
-__Python 3.4.x__. Probably will run on any Python 3x for the development evironment but the deployment will use version __3.4.5__.
+__Python 3.5.x__
 
+Python 3.6 should also work but is not tested.
 
 ## Running locally
 
 Install dependencies.
 
 ```
-$ pip install -r requirements
+$ pip install -r requirements.txt
+$ pip install uwsgi
 ```
 
 This will start the service locally bound to `0.0.0.0:5128`.
 
 ```
-$ python service.py
+$ cd src
+$ uwsgi --ini uwsgi.ini
 ```
 
+You must have a CoreNLP server running at `http://corenlp:9000`. If not,
+change it in `Resolver.py` to your instance.
+
 ## Running within a Docker Container
 
 Build the image.
 
 ```
-$ docker build -t coref-resolver:dev .
+$ docker build -t coreferenece:v1.1.0-beta.2 .
 ```
 
 Run in background.
 
 ```
-$ docker run -d --name coref-resolver -p 5128:5128 coref-resolver:dev
+$ docker run -d --name coreference -p 5128:5128 coreference:v1.1.0-beta.2
 ```
 
 Look at the container stdout.
 
 ```
-$ docker logs -f coref-resolver
+$ docker logs -f coreference
 ```
 
+
 ## Testing
 
 
 ```
 curl -X POST \
  -H "Content-Type: application/json" \
- -d '{"text": "Bernhard is working on two projects. He is employed in Mario and PACE."}' "http://localhost:5128/resolve/text"
+ -H "Accept: application/json" \
+ -d '{"text": "Donald Trump is the president of USA. He is a business man."}' "http://localhost:5128/resolve/text"
 ```
 
 Expected output:
 
 ```json
 {
-  "text": "Donald Trump is the president of USA. He is a business man.",
+  "text": "Donald Trump is the president of USA. Donal Trump is a business man.",
 }
-
 ```

requirements.txt

@@ -1,5 +1,4 @@
-joblib==0.10.3
+joblib==0.11
 pycorenlp==0.3.0
-nltk==3.2.1
-flask==0.11.1
-typing==3.5.2.2
+nltk==3.2.2
+Flask==0.12

run.py

@@ -12,6 +12,9 @@
 BATCH_SIZE = 200
 
 
+# This file is for mass extractions and not recommended for single use.
+
+
 def run(input_file: TextIOWrapper, output_file: TextIOWrapper, jobs: int, max_articles: Optional[int]) -> None:
     log.info("Import done, now starting work")
 

src/Resolver.py → src/coreference/Resolver.py

@@ -1,19 +1,19 @@
 # -*- coding: utf-8 -*-
 import logging
+from typing import List, Tuple, Dict, Union, Set
+
 import os
 import re
 from collections import Counter
-
 from nltk.tokenize import sent_tokenize, word_tokenize
 from pycorenlp import StanfordCoreNLP
-from typing import List, Tuple, Dict, Union, Set
 
-from classifier import GenderClassifier
-from constant_types import PronounType, QuantityType, GenderType, EntityType
-from references import NameReference, NominalReference, PronominalReference
-from references import ResolvedPassage
-from references import Substitution
-from references.TypeLookup import TypeLookup
+from .classifier import GenderClassifier
+from .constant_types import PronounType, QuantityType, GenderType, EntityType
+from .references import NameReference, NominalReference, PronominalReference
+from .references import ResolvedPassage
+from .references import Substitution
+from .references.TypeLookup import TypeLookup
 
 logging.getLogger('requests').setLevel(logging.WARNING)
 
@@ -31,8 +31,8 @@ class Resolver(object):
         )
     )
     _separator = '---- <> ----'
-    _nlp = StanfordCoreNLP('http://localhost:9000')
     # _nlp = StanfordCoreNLP('http://localhost:9000')
+    _nlp = StanfordCoreNLP('http://corenlp:9000')
 
     _gender_classifier = GenderClassifier()
 

src/classifier/GenderClassifier.py → src/coreference/classifier/GenderClassifier.py

@@ -1,9 +1,8 @@
-import random
-
 import nltk
+import random
 from nltk.corpus import names
 
-from constant_types import GenderType
+from ..constant_types import GenderType
 
 
 class GenderClassifier(object):

src/references/NameReference.py → src/coreference/references/NameReference.py

@@ -1,6 +1,6 @@
-from constant_types import EntityType
-from constant_types import GenderType
-from constant_types import QuantityType
+from ..constant_types import EntityType
+from ..constant_types import GenderType
+from ..constant_types import QuantityType
 
 
 class NameReference(object):

src/references/PronominalReference.py → src/coreference/references/PronominalReference.py

@@ -1,11 +1,9 @@
 # -*- coding: utf-8 -*-
 
-
-# noinspection PyPep8Naming
-from constant_types import EntityType
-from constant_types import GenderType
-from constant_types import PronounType
-from constant_types import QuantityType
+from ..constant_types import EntityType
+from ..constant_types import GenderType
+from ..constant_types import PronounType
+from ..constant_types import QuantityType
 
 
 class PronominalReference(object):

src/coreference/references/ResolvedPassage.py

@@ -0,0 +1,14 @@
+# -*- coding: utf-8 -*-
+from typing import Dict
+
+
+class ResolvedPassage(object):
+    def __init__(self, resolved_passage: int, linked_entities: Dict[str, str]):
+        self.resolved_passage = resolved_passage
+        self.linked_entities = linked_entities
+
+    def __str__(self):
+        return "{self.resolved_passage} {self.linked_entities}".format(self=self)
+
+    def __repr__(self):
+        return self.__str__()

src/references/Substitution.py → src/coreference/references/Substitution.py

@@ -1,7 +1,6 @@
 # -*- coding: utf-8 -*-
 
 
-# noinspection PyPep8Naming
 class Substitution(object):
     def __init__(self, sentence_index: int, original: str, reference: str):
         self.sentence_index = sentence_index

src/references/TypeLookup.py → src/coreference/references/TypeLookup.py

@@ -5,12 +5,12 @@ class TypeLookup(object):
     __separator = ' -----+++----- '
 
     def __init__(self, filename: str):
-        self.__instanceTypes = {}
+        self.__instance_types = {}
 
         with bz2.open(filename, 'rt') as f:
             for line in f:
                 fields = line.split(self.__separator)
-                self.__instanceTypes[fields[0].replace('<', '').replace('>', '')] = fields[1].replace('\n', '')
+                self.__instance_types[fields[0].replace('<', '').replace('>', '')] = fields[1].replace('\n', '')
 
     def type(self, uri: str) -> str:
-        return self.__instanceTypes[uri] if uri in self.__instanceTypes else ''
+        return self.__instance_types[uri] if uri in self.__instance_types else ''

src/tests/Tests.py → src/coreference/tests/Tests.py

@@ -4,6 +4,8 @@
 from Resolver import Resolver
 
 
+# UNUSED, UNSUPPORTED, UNMAINTAINED
+
 def resolve(filename, uri):
     with open(filename, 'r') as f:
         text = f.read()

src/tests/test_Resolver.py → src/coreference/tests/test_Resolver.py

@@ -1,21 +1,11 @@
 import logging
-import os
-import sys
 import unittest
 
 from nltk import sent_tokenize
 
-logging.getLogger('requests').setLevel(logging.WARNING)
-
-sys.path.append(
-    os.path.dirname(
-        os.path.dirname(
-            os.path.abspath(__file__)
-        )
-    )
-)
+from ..Resolver import Resolver
 
-from Resolver import Resolver
+logging.getLogger('requests').setLevel(logging.WARNING)
 
 
 # noinspection PyProtectedMember

src/tests/test_WikiArticle.py → src/coreference/tests/test_WikiArticle.py

@@ -1,24 +1,14 @@
 import logging
-import os
-import sys
 import unittest
 
+import os
 from nltk import sent_tokenize
 
-logging.getLogger('requests').setLevel(logging.WARNING)
+from ..Resolver import Resolver
 
-sys.path.append(
-    os.path.dirname(
-        os.path.dirname(
-            os.path.abspath(__file__)
-        )
-    )
-)
-
-from Resolver import Resolver
+logging.getLogger('requests').setLevel(logging.WARNING)
 
 
-# noinspection PyProtectedMember
 class TestResolverForBarackObama(unittest.TestCase):
     def setUp(self):
         self.maxDiff = None

service.py → src/service.py

@@ -4,18 +4,18 @@
 
 from flask import Flask, jsonify, make_response, request
 
-from Resolver import Resolver
+from coreference.Resolver import Resolver
 
-application = Flask(__name__)
+app = Flask(__name__)
 
 
-@application.route('/', methods=['GET'])
+@app.route('/', methods=['GET'])
 def hello():
     return jsonify({'name': 'Text Coreference Resolver',
                     'server_time': datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")})
 
 
-@application.route('/resolve/wiki', methods=['POST'])
+@app.route('/resolve/wiki', methods=['POST'])
 def resolve_wiki():
     data = request.get_json()
     if data is None:
@@ -31,7 +31,7 @@ def resolve_wiki():
     return jsonify({'text': substituted})
 
 
-@application.route('/resolve/text', methods=['POST'])
+@app.route('/resolve/text', methods=['POST'])
 def resolve_text():
     data = request.get_json()
     if data is None:
@@ -47,15 +47,15 @@ def resolve_text():
     return jsonify({'text': substituted})
 
 
-@application.route('/resolve/link', methods=['POST'])
+@app.route('/resolve/link', methods=['POST'])
 def resolve_link():
     return make_response(jsonify({'message': 'Not implemented.'}), 501)
 
 
-@application.errorhandler(404)
+@app.errorhandler(404)
 def page_not_found(_):
     return make_response(jsonify({'message': 'No interface defined for URL'}), 404)
 
 
 if __name__ == "__main__":
-    application.run(port=5128, debug=False, use_reloader=False, host='0.0.0.0')
+    app.run(port=5128, debug=False, use_reloader=False, host='0.0.0.0')

uwsgi.ini → src/uwsgi.ini

@@ -1,12 +1,12 @@
 [uwsgi]
 protocol = http
 socket = 0.0.0.0:5128
-mount = /corefapp=service.py
-callable = application
+mount = /app=service.py
+callable = app
 processes = 4
 master = true
 harakiri = 600
 uid = root
 gid = root
 no-orphans = true
-plugins = python3
+enable-threads = "*"