opener-kaf-naf-parser 1.0.1 → 1.0.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +67 -8
- data/bin/kaf-naf-parser-daemon +10 -0
- data/core/kaf-naf-parser.py +5 -5
- data/exec/kaf-naf-parser.rb +9 -0
- data/ext/hack/Rakefile +13 -0
- data/lib/opener/kaf_naf_parser/version.rb +1 -1
- data/opener-kaf-naf-parser.gemspec +5 -1
- data/pre_install_requirements.txt +3 -0
- metadata +37 -51
- data/core/packages/KafNafParser-1.2.tar.gz +0 -0
- data/core/packages/VUA_pylib-1.3.tar.gz +0 -0
- data/core/site-packages/pre_build/KafNafParser/KafNafParserMod.py +0 -338
- data/core/site-packages/pre_build/KafNafParser/__init__.py +0 -14
- data/core/site-packages/pre_build/KafNafParser/constituency_data.py +0 -125
- data/core/site-packages/pre_build/KafNafParser/coreference_data.py +0 -52
- data/core/site-packages/pre_build/KafNafParser/dependency_data.py +0 -80
- data/core/site-packages/pre_build/KafNafParser/entity_data.py +0 -59
- data/core/site-packages/pre_build/KafNafParser/external_references_data.py +0 -41
- data/core/site-packages/pre_build/KafNafParser/feature_extractor/__init__.py +0 -2
- data/core/site-packages/pre_build/KafNafParser/feature_extractor/constituency.py +0 -205
- data/core/site-packages/pre_build/KafNafParser/feature_extractor/dependency.py +0 -300
- data/core/site-packages/pre_build/KafNafParser/features_data.py +0 -71
- data/core/site-packages/pre_build/KafNafParser/header_data.py +0 -127
- data/core/site-packages/pre_build/KafNafParser/opinion_data.py +0 -200
- data/core/site-packages/pre_build/KafNafParser/references_data.py +0 -15
- data/core/site-packages/pre_build/KafNafParser/span_data.py +0 -63
- data/core/site-packages/pre_build/KafNafParser/term_data.py +0 -111
- data/core/site-packages/pre_build/KafNafParser/term_sentiment_data.py +0 -42
- data/core/site-packages/pre_build/KafNafParser/text_data.py +0 -90
- data/core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/PKG-INFO +0 -10
- data/core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/SOURCES.txt +0 -22
- data/core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/dependency_links.txt +0 -1
- data/core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/installed-files.txt +0 -47
- data/core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/top_level.txt +0 -1
- data/core/site-packages/pre_build/VUA_pylib/__init__.py +0 -1
- data/core/site-packages/pre_build/VUA_pylib/common/__init__.py +0 -1
- data/core/site-packages/pre_build/VUA_pylib/common/common.py +0 -28
- data/core/site-packages/pre_build/VUA_pylib/corpus_reader/__init__.py +0 -1
- data/core/site-packages/pre_build/VUA_pylib/corpus_reader/google_web_nl.py +0 -156
- data/core/site-packages/pre_build/VUA_pylib/io_utils/__init__.py +0 -1
- data/core/site-packages/pre_build/VUA_pylib/io_utils/feature_file.py +0 -121
- data/core/site-packages/pre_build/VUA_pylib/lexicon/__init__.py +0 -1
- data/core/site-packages/pre_build/VUA_pylib/lexicon/lexicon.py +0 -72
- data/core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/PKG-INFO +0 -10
- data/core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/SOURCES.txt +0 -14
- data/core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/dependency_links.txt +0 -1
- data/core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/installed-files.txt +0 -23
- data/core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/top_level.txt +0 -1
- data/core/site-packages/pre_build/VUKafParserPy/KafDataObjectsMod.py +0 -165
- data/core/site-packages/pre_build/VUKafParserPy/KafParserMod.py +0 -439
- data/core/site-packages/pre_build/VUKafParserPy/__init__.py +0 -7
- data/core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/PKG-INFO +0 -10
- data/core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/SOURCES.txt +0 -7
- data/core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/dependency_links.txt +0 -1
- data/core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/installed-files.txt +0 -11
- data/core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/top_level.txt +0 -1
- data/pre_build_requirements.txt +0 -3
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 4da7b5e3c461558e2f52191daa6dace50c953144
|
4
|
+
data.tar.gz: fa8c4a239654f1b04c42ea735e781aec9589c857
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: e4d0ea293d50e961ea9e2b58bc643c553e77511fdb05f2fc1893d9cbaff6064ac07c08de9d93bb4eb3e6255762455a0daf4f0b92b9c7e1c4a519ff219bbf3c2b
|
7
|
+
data.tar.gz: 248ba180f04b3f4a7ccd3ec9b1411974d1a535f926b58a40b388e6407bd0ec3e78d4eb9a9fca4f548aba385575064fffc03b7ca5503856a6a4e563bcf017b68b
|
data/README.md
CHANGED
@@ -1,5 +1,9 @@
|
|
1
|
+
replace:
|
1
2
|
kaf-naf-parser
|
2
|
-
|
3
|
+
kaf-naf-parser
|
4
|
+
|
5
|
+
Introduction
|
6
|
+
------------
|
3
7
|
|
4
8
|
This library converts KAF to NAF and NAF to KAF. It also contains a webservice
|
5
9
|
for doing exactly this.
|
@@ -8,8 +12,23 @@ It uses slightly altered
|
|
8
12
|
distributions of the VUA_pylib and KafNafParserPy but it keeps the external
|
9
13
|
libraries contained, so that shouldn't influence anything on your system.
|
10
14
|
|
11
|
-
|
12
|
-
|
15
|
+
### Confused by some terminology?
|
16
|
+
|
17
|
+
This software is part of a larger collection of natural language processing
|
18
|
+
tools known as "the OpeNER project". You can find more information about the
|
19
|
+
project at (the OpeNER portal)[http://opener-project.github.io]. There you can
|
20
|
+
also find references to terms like KAF (an XML standard to represent linguistic
|
21
|
+
annotations in texts), component, cores, scenario's and pipelines.
|
22
|
+
|
23
|
+
Quick Use Example
|
24
|
+
-----------------
|
25
|
+
|
26
|
+
Installing the kaf-naf-parser can be done by executing:
|
27
|
+
|
28
|
+
gem install opener-kaf-naf-parser
|
29
|
+
|
30
|
+
Please bare in mind that all components in OpeNER take KAF as an input and
|
31
|
+
output KAF by default.
|
13
32
|
|
14
33
|
It ships with 3 command line tools and a webservice:
|
15
34
|
|
@@ -38,9 +57,49 @@ And can then be accessed at: http://localhost:9292
|
|
38
57
|
|
39
58
|
Easiest way to install the gem:
|
40
59
|
|
41
|
-
|
42
|
-
|
43
|
-
|
44
|
-
|
45
|
-
|
60
|
+
### Daemon
|
61
|
+
|
62
|
+
Last but not least the kaf-naf-parser comes shipped with a daemon that
|
63
|
+
can read jobs (and write) jobs to and from Amazon SQS queues. For more
|
64
|
+
information type:
|
65
|
+
|
66
|
+
kaf-naf-parser-daemon -h
|
67
|
+
|
68
|
+
|
69
|
+
Description of dependencies
|
70
|
+
---------------------------
|
71
|
+
|
72
|
+
This component runs best if you run it in an environment suited for OpeNER
|
73
|
+
components. You can find an installation guide and helper tools in the (OpeNER
|
74
|
+
installer)[https://github.com/opener-project/opener-installer] and (an
|
75
|
+
installation guide on the Opener
|
76
|
+
Website)[http://opener-project.github.io/getting-started/how-to/local-installation.html]
|
77
|
+
|
78
|
+
At least you need the following system setup:
|
79
|
+
|
80
|
+
### Dependencies for normal use:
|
81
|
+
|
82
|
+
* Ruby 1.9.3 or newer
|
83
|
+
* Python 2.6
|
84
|
+
|
85
|
+
Where to go from here
|
86
|
+
---------------------
|
87
|
+
|
88
|
+
* Check (the project website)[http://opener-project.github.io]
|
89
|
+
* (Checkout the webservice)[http://opener.olery.com/kaf-naf-parser]
|
90
|
+
|
91
|
+
Report problem/Get help
|
92
|
+
-----------------------
|
93
|
+
|
94
|
+
If you encounter problems, please email support@opener-project.eu or leave an
|
95
|
+
issue in the (issue tracker)[https://github.com/opener-project/kaf-naf-parser/issues].
|
96
|
+
|
97
|
+
|
98
|
+
Contributing
|
99
|
+
------------
|
46
100
|
|
101
|
+
1. Fork it ( http://github.com/opener-project/kaf-naf-parser/fork )
|
102
|
+
2. Create your feature branch (`git checkout -b my-new-feature`)
|
103
|
+
3. Commit your changes (`git commit -am 'Add some feature'`)
|
104
|
+
4. Push to the branch (`git push origin my-new-feature`)
|
105
|
+
5. Create new Pull Request
|
data/core/kaf-naf-parser.py
CHANGED
@@ -1,12 +1,12 @@
|
|
1
1
|
#!/usr/bin/env python
|
2
2
|
|
3
|
-
# This updates the load path to ensure that the local site-packages directory
|
4
|
-
# can be used to load packages (e.g. a locally installed copy of lxml).
|
5
|
-
|
6
3
|
import sys, getopt, os
|
7
4
|
|
8
|
-
|
9
|
-
|
5
|
+
this_folder = os.path.dirname(os.path.realpath(__file__))
|
6
|
+
|
7
|
+
# This updates the load path to ensure that the local site-packages directory
|
8
|
+
# can be used to load packages (e.g. a locally installed copy of lxml).
|
9
|
+
sys.path.append(os.path.join(this_folder, 'site-packages/pre_install'))
|
10
10
|
|
11
11
|
from KafNafParser import KafNafParser
|
12
12
|
|
data/ext/hack/Rakefile
ADDED
@@ -0,0 +1,13 @@
|
|
1
|
+
require 'rake'
|
2
|
+
require_relative 'support'
|
3
|
+
|
4
|
+
desc 'Verifies the requirements'
|
5
|
+
task :requirements do
|
6
|
+
verify_requirements
|
7
|
+
end
|
8
|
+
|
9
|
+
task :default => :requirements do
|
10
|
+
path = File.join(PYTHON_SITE_PACKAGES, 'pre_install')
|
11
|
+
|
12
|
+
pip_install(PRE_INSTALL_REQUIREMENTS, path)
|
13
|
+
end
|
@@ -8,6 +8,7 @@ Gem::Specification.new do |gem|
|
|
8
8
|
gem.description = gem.summary
|
9
9
|
gem.homepage = "http://github.com/cltl/KafNafParserPy"
|
10
10
|
gem.required_ruby_version = '>= 1.9.2'
|
11
|
+
gem.extensions = ['ext/hack/Rakefile']
|
11
12
|
|
12
13
|
gem.files = Dir.glob([
|
13
14
|
'core/packages/**/*',
|
@@ -19,7 +20,8 @@ Gem::Specification.new do |gem|
|
|
19
20
|
'*.gemspec',
|
20
21
|
'LICENSE',
|
21
22
|
'*_requirements.txt',
|
22
|
-
'README.md'
|
23
|
+
'README.md',
|
24
|
+
'exec/**/*'
|
23
25
|
]).select { |file| File.file?(file) }
|
24
26
|
|
25
27
|
gem.executables = Dir.glob('bin/*').map { |file| File.basename(file) }
|
@@ -31,6 +33,8 @@ Gem::Specification.new do |gem|
|
|
31
33
|
gem.add_dependency 'uuidtools'
|
32
34
|
gem.add_dependency 'opener-webservice'
|
33
35
|
gem.add_dependency 'puma'
|
36
|
+
gem.add_dependency 'opener-daemons'
|
37
|
+
gem.add_dependency 'opener-core', ['>= 0.1.2']
|
34
38
|
|
35
39
|
gem.add_development_dependency 'rspec'
|
36
40
|
gem.add_development_dependency 'cucumber'
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: opener-kaf-naf-parser
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 1.0.
|
4
|
+
version: 1.0.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- ruben.izquierdobevia@vu.nl
|
@@ -9,7 +9,7 @@ authors:
|
|
9
9
|
autorequire:
|
10
10
|
bindir: bin
|
11
11
|
cert_chain: []
|
12
|
-
date: 2014-05-
|
12
|
+
date: 2014-05-23 00:00:00.000000000 Z
|
13
13
|
dependencies:
|
14
14
|
- !ruby/object:Gem::Dependency
|
15
15
|
name: builder
|
@@ -109,6 +109,34 @@ dependencies:
|
|
109
109
|
- - ">="
|
110
110
|
- !ruby/object:Gem::Version
|
111
111
|
version: '0'
|
112
|
+
- !ruby/object:Gem::Dependency
|
113
|
+
name: opener-daemons
|
114
|
+
requirement: !ruby/object:Gem::Requirement
|
115
|
+
requirements:
|
116
|
+
- - ">="
|
117
|
+
- !ruby/object:Gem::Version
|
118
|
+
version: '0'
|
119
|
+
type: :runtime
|
120
|
+
prerelease: false
|
121
|
+
version_requirements: !ruby/object:Gem::Requirement
|
122
|
+
requirements:
|
123
|
+
- - ">="
|
124
|
+
- !ruby/object:Gem::Version
|
125
|
+
version: '0'
|
126
|
+
- !ruby/object:Gem::Dependency
|
127
|
+
name: opener-core
|
128
|
+
requirement: !ruby/object:Gem::Requirement
|
129
|
+
requirements:
|
130
|
+
- - ">="
|
131
|
+
- !ruby/object:Gem::Version
|
132
|
+
version: 0.1.2
|
133
|
+
type: :runtime
|
134
|
+
prerelease: false
|
135
|
+
version_requirements: !ruby/object:Gem::Requirement
|
136
|
+
requirements:
|
137
|
+
- - ">="
|
138
|
+
- !ruby/object:Gem::Version
|
139
|
+
version: 0.1.2
|
112
140
|
- !ruby/object:Gem::Dependency
|
113
141
|
name: rspec
|
114
142
|
requirement: !ruby/object:Gem::Requirement
|
@@ -168,68 +196,26 @@ dependencies:
|
|
168
196
|
description: Parser to translate KAF to NAF and back.
|
169
197
|
email:
|
170
198
|
executables:
|
199
|
+
- kaf-naf-parser-daemon
|
171
200
|
- naf-to-kaf
|
172
201
|
- kaf-naf-parser
|
173
202
|
- kaf-to-naf
|
174
203
|
- kaf-naf-parser-server
|
175
|
-
extensions:
|
204
|
+
extensions:
|
205
|
+
- ext/hack/Rakefile
|
176
206
|
extra_rdoc_files: []
|
177
207
|
files:
|
178
208
|
- LICENSE
|
179
209
|
- README.md
|
180
210
|
- bin/kaf-naf-parser
|
211
|
+
- bin/kaf-naf-parser-daemon
|
181
212
|
- bin/kaf-naf-parser-server
|
182
213
|
- bin/kaf-to-naf
|
183
214
|
- bin/naf-to-kaf
|
184
215
|
- config.ru
|
185
216
|
- core/kaf-naf-parser.py
|
186
|
-
-
|
187
|
-
-
|
188
|
-
- core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/PKG-INFO
|
189
|
-
- core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/SOURCES.txt
|
190
|
-
- core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/dependency_links.txt
|
191
|
-
- core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/installed-files.txt
|
192
|
-
- core/site-packages/pre_build/KafNafParser-1.2-py2.7.egg-info/top_level.txt
|
193
|
-
- core/site-packages/pre_build/KafNafParser/KafNafParserMod.py
|
194
|
-
- core/site-packages/pre_build/KafNafParser/__init__.py
|
195
|
-
- core/site-packages/pre_build/KafNafParser/constituency_data.py
|
196
|
-
- core/site-packages/pre_build/KafNafParser/coreference_data.py
|
197
|
-
- core/site-packages/pre_build/KafNafParser/dependency_data.py
|
198
|
-
- core/site-packages/pre_build/KafNafParser/entity_data.py
|
199
|
-
- core/site-packages/pre_build/KafNafParser/external_references_data.py
|
200
|
-
- core/site-packages/pre_build/KafNafParser/feature_extractor/__init__.py
|
201
|
-
- core/site-packages/pre_build/KafNafParser/feature_extractor/constituency.py
|
202
|
-
- core/site-packages/pre_build/KafNafParser/feature_extractor/dependency.py
|
203
|
-
- core/site-packages/pre_build/KafNafParser/features_data.py
|
204
|
-
- core/site-packages/pre_build/KafNafParser/header_data.py
|
205
|
-
- core/site-packages/pre_build/KafNafParser/opinion_data.py
|
206
|
-
- core/site-packages/pre_build/KafNafParser/references_data.py
|
207
|
-
- core/site-packages/pre_build/KafNafParser/span_data.py
|
208
|
-
- core/site-packages/pre_build/KafNafParser/term_data.py
|
209
|
-
- core/site-packages/pre_build/KafNafParser/term_sentiment_data.py
|
210
|
-
- core/site-packages/pre_build/KafNafParser/text_data.py
|
211
|
-
- core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/PKG-INFO
|
212
|
-
- core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/SOURCES.txt
|
213
|
-
- core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/dependency_links.txt
|
214
|
-
- core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/installed-files.txt
|
215
|
-
- core/site-packages/pre_build/VUA_pylib-1.3-py2.7.egg-info/top_level.txt
|
216
|
-
- core/site-packages/pre_build/VUA_pylib/__init__.py
|
217
|
-
- core/site-packages/pre_build/VUA_pylib/common/__init__.py
|
218
|
-
- core/site-packages/pre_build/VUA_pylib/common/common.py
|
219
|
-
- core/site-packages/pre_build/VUA_pylib/corpus_reader/__init__.py
|
220
|
-
- core/site-packages/pre_build/VUA_pylib/corpus_reader/google_web_nl.py
|
221
|
-
- core/site-packages/pre_build/VUA_pylib/io_utils/__init__.py
|
222
|
-
- core/site-packages/pre_build/VUA_pylib/io_utils/feature_file.py
|
223
|
-
- core/site-packages/pre_build/VUA_pylib/lexicon/__init__.py
|
224
|
-
- core/site-packages/pre_build/VUA_pylib/lexicon/lexicon.py
|
225
|
-
- core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/PKG-INFO
|
226
|
-
- core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/SOURCES.txt
|
227
|
-
- core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/dependency_links.txt
|
228
|
-
- core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/installed-files.txt
|
229
|
-
- core/site-packages/pre_build/VUKafParserPy-1.0-py2.7.egg-info/top_level.txt
|
230
|
-
- core/site-packages/pre_build/VUKafParserPy/KafDataObjectsMod.py
|
231
|
-
- core/site-packages/pre_build/VUKafParserPy/KafParserMod.py
|
232
|
-
- core/site-packages/pre_build/VUKafParserPy/__init__.py
|
217
|
+
- exec/kaf-naf-parser.rb
|
218
|
+
- ext/hack/Rakefile
|
233
219
|
- ext/hack/support.rb
|
234
220
|
- lib/opener/kaf_naf_parser.rb
|
235
221
|
- lib/opener/kaf_naf_parser/cli.rb
|
@@ -239,7 +225,7 @@ files:
|
|
239
225
|
- lib/opener/kaf_naf_parser/views/index.erb
|
240
226
|
- lib/opener/kaf_naf_parser/views/result.erb
|
241
227
|
- opener-kaf-naf-parser.gemspec
|
242
|
-
-
|
228
|
+
- pre_install_requirements.txt
|
243
229
|
homepage: http://github.com/cltl/KafNafParserPy
|
244
230
|
licenses: []
|
245
231
|
metadata: {}
|
Binary file
|
Binary file
|
@@ -1,338 +0,0 @@
|
|
1
|
-
## LIST OF CHANGES
|
2
|
-
# Ruben 8-nov-2013
|
3
|
-
# + included layers for entities, properties, opinions
|
4
|
-
# + renamed all classes to Cnameoftheclass
|
5
|
-
# Ruben 15-nov-2013
|
6
|
-
# + included constituency layer
|
7
|
-
#
|
8
|
-
# Ruben 19-nov-2013
|
9
|
-
# + included dependency layer
|
10
|
-
# Ruben 17-dec-2013
|
11
|
-
# + modified all to red/write NAF and KAF
|
12
|
-
#
|
13
|
-
# Ruben 21-Feb-2014
|
14
|
-
# + Included coreference layer
|
15
|
-
|
16
|
-
|
17
|
-
__last_modified = '17dec2013'
|
18
|
-
|
19
|
-
from lxml import etree
|
20
|
-
from header_data import *
|
21
|
-
from text_data import *
|
22
|
-
from term_data import *
|
23
|
-
from entity_data import *
|
24
|
-
from features_data import *
|
25
|
-
from opinion_data import *
|
26
|
-
from constituency_data import *
|
27
|
-
from dependency_data import *
|
28
|
-
from feature_extractor import Cdependency_extractor, Cconstituency_extractor
|
29
|
-
from coreference_data import *
|
30
|
-
from references_data import Creferences
|
31
|
-
|
32
|
-
import sys
|
33
|
-
|
34
|
-
|
35
|
-
|
36
|
-
class KafNafParser:
|
37
|
-
def __init__(self,filename):
|
38
|
-
self.tree = None
|
39
|
-
self.filename = filename
|
40
|
-
self.tree = etree.parse(filename,etree.XMLParser(remove_blank_text=True))
|
41
|
-
self.root = self.tree.getroot()
|
42
|
-
self.type = self.root.tag # KAF NAF
|
43
|
-
|
44
|
-
self.header = None
|
45
|
-
self.text_layer = None
|
46
|
-
self.term_layer = None
|
47
|
-
self.entity_layer = None
|
48
|
-
self.features_layer = None
|
49
|
-
self.opinion_layer = None
|
50
|
-
self.constituency_layer = None
|
51
|
-
self.dependency_layer = None
|
52
|
-
self.coreference_layer = None
|
53
|
-
|
54
|
-
## Specific feature extractor for complicated layers
|
55
|
-
self.my_dependency_extractor = None
|
56
|
-
self.my_constituency_extractor = None
|
57
|
-
##################################################
|
58
|
-
|
59
|
-
#######
|
60
|
-
self.dict_tokens_for_tid = None
|
61
|
-
##
|
62
|
-
|
63
|
-
self.lang = self.root.get('{http://www.w3.org/XML/1998/namespace}lang')
|
64
|
-
self.version = self.root.get('version')
|
65
|
-
|
66
|
-
if self.type == 'NAF':
|
67
|
-
node_header = self.root.find('nafHeader')
|
68
|
-
elif self.type == 'KAF':
|
69
|
-
node_header = self.root.find('kafHeader')
|
70
|
-
|
71
|
-
if node_header is not None:
|
72
|
-
self.header = CHeader(node_header,self.type)
|
73
|
-
|
74
|
-
# Text layer adapted to naf/kaf
|
75
|
-
node_text = self.root.find('text')
|
76
|
-
if node_text is not None:
|
77
|
-
self.text_layer = Ctext(node=node_text,type=self.type)
|
78
|
-
|
79
|
-
node_term = self.root.find('terms')
|
80
|
-
if node_term is not None:
|
81
|
-
self.term_layer = Cterms(node=node_term,type=self.type)
|
82
|
-
|
83
|
-
node_entity = self.root.find('entities')
|
84
|
-
if node_entity is not None:
|
85
|
-
self.entity_layer = Centities(node_entity,type=self.type)
|
86
|
-
|
87
|
-
node_features = self.root.find('features')
|
88
|
-
if node_features is not None:
|
89
|
-
self.features_layer = Cfeatures(node_features,type=self.type)
|
90
|
-
|
91
|
-
node_opinions = self.root.find('opinions')
|
92
|
-
if node_opinions is not None:
|
93
|
-
self.opinion_layer = Copinions(node_opinions,type=self.type)
|
94
|
-
|
95
|
-
# Definition KAF/NAF is the same
|
96
|
-
node_constituency = self.root.find('constituency')
|
97
|
-
if node_constituency is not None:
|
98
|
-
self.constituency_layer = Cconstituency(node_constituency)
|
99
|
-
|
100
|
-
# Definition KAF/NAF is the same
|
101
|
-
node_dependency = self.root.find('deps')
|
102
|
-
if node_dependency is not None:
|
103
|
-
self.dependency_layer = Cdependencies(node_dependency)
|
104
|
-
|
105
|
-
node_coreferences = self.root.find('coreferences')
|
106
|
-
if node_coreferences is not None:
|
107
|
-
self.coreference_layer = Ccoreferences(node_coreferences,type=self.type)
|
108
|
-
|
109
|
-
def get_type(self):
|
110
|
-
return self.type
|
111
|
-
|
112
|
-
def get_filename(self):
|
113
|
-
return self.filename
|
114
|
-
|
115
|
-
def to_kaf(self):
|
116
|
-
#Convert the root
|
117
|
-
if self.type == 'NAF':
|
118
|
-
self.root.tag = 'KAF'
|
119
|
-
self.type = 'KAF'
|
120
|
-
|
121
|
-
## Convert the header
|
122
|
-
if self.header is not None:
|
123
|
-
self.header.to_kaf()
|
124
|
-
|
125
|
-
## Convert the token layer
|
126
|
-
if self.text_layer is not None:
|
127
|
-
self.text_layer.to_kaf()
|
128
|
-
|
129
|
-
## Convert the term layer
|
130
|
-
if self.term_layer is not None:
|
131
|
-
self.term_layer.to_kaf()
|
132
|
-
|
133
|
-
## Convert the entity layer
|
134
|
-
if self.entity_layer is not None:
|
135
|
-
self.entity_layer.to_kaf()
|
136
|
-
|
137
|
-
## Convert the features layer
|
138
|
-
## There is no feature layer defined in NAF, but we assumed
|
139
|
-
## that is defined will be followin the same rules
|
140
|
-
if self.features_layer is not None:
|
141
|
-
self.features_layer.to_kaf()
|
142
|
-
|
143
|
-
|
144
|
-
##Convert the opinion layer
|
145
|
-
if self.opinion_layer is not None:
|
146
|
-
self.opinion_layer.to_kaf()
|
147
|
-
|
148
|
-
## Convert the constituency layer
|
149
|
-
## This layer is exactly the same in KAF/NAF
|
150
|
-
if self.constituency_layer is not None:
|
151
|
-
self.constituency_layer.to_kaf() #Does nothing...
|
152
|
-
|
153
|
-
|
154
|
-
## Convert the dedepency layer
|
155
|
-
## It is not defined on KAF so we assme both will be similar
|
156
|
-
if self.dependency_layer is not None:
|
157
|
-
self.dependency_layer.to_kaf()
|
158
|
-
|
159
|
-
if self.coreference_layer is not None:
|
160
|
-
self.coreference_layer.to_kaf()
|
161
|
-
|
162
|
-
|
163
|
-
def to_naf(self):
|
164
|
-
#Convert the root
|
165
|
-
if self.type == 'KAF':
|
166
|
-
self.root.tag = self.type = 'NAF'
|
167
|
-
|
168
|
-
## Convert the header
|
169
|
-
if self.header is not None:
|
170
|
-
self.header.to_naf()
|
171
|
-
|
172
|
-
## Convert the token layer
|
173
|
-
if self.text_layer is not None:
|
174
|
-
self.text_layer.to_naf()
|
175
|
-
|
176
|
-
|
177
|
-
## Convert the term layer
|
178
|
-
if self.term_layer is not None:
|
179
|
-
self.term_layer.to_naf()
|
180
|
-
|
181
|
-
|
182
|
-
## Convert the entity layer
|
183
|
-
if self.entity_layer is not None:
|
184
|
-
self.entity_layer.to_naf()
|
185
|
-
|
186
|
-
## Convert the features layer
|
187
|
-
## There is no feature layer defined in NAF, but we assumed
|
188
|
-
## that is defined will be followin the same rules
|
189
|
-
if self.features_layer is not None:
|
190
|
-
self.features_layer.to_naf()
|
191
|
-
|
192
|
-
|
193
|
-
##Convert the opinion layer
|
194
|
-
if self.opinion_layer is not None:
|
195
|
-
self.opinion_layer.to_naf()
|
196
|
-
|
197
|
-
|
198
|
-
## Convert the constituency layer
|
199
|
-
## This layer is exactly the same in KAF/NAF
|
200
|
-
if self.constituency_layer is not None:
|
201
|
-
self.constituency_layer.to_naf() #Does nothing...
|
202
|
-
|
203
|
-
|
204
|
-
## Convert the dedepency layer
|
205
|
-
## It is not defined on KAF so we assume both will be similar
|
206
|
-
if self.dependency_layer is not None:
|
207
|
-
self.dependency_layer.to_naf() #Does nothing...
|
208
|
-
|
209
|
-
if self.coreference_layer is not None:
|
210
|
-
self.coreference_layer.to_naf()
|
211
|
-
|
212
|
-
|
213
|
-
|
214
|
-
def print_constituency(self):
|
215
|
-
print self.constituency_layer
|
216
|
-
|
217
|
-
def get_trees(self):
|
218
|
-
if self.constituency_layer is not None:
|
219
|
-
for tree in self.constituency_layer.get_trees():
|
220
|
-
yield tree
|
221
|
-
|
222
|
-
|
223
|
-
def get_dependencies(self):
|
224
|
-
if self.dependency_layer is not None:
|
225
|
-
for dep in self.dependency_layer.get_dependencies():
|
226
|
-
yield dep
|
227
|
-
|
228
|
-
def get_language(self):
|
229
|
-
return self.lang
|
230
|
-
|
231
|
-
def get_tokens(self):
|
232
|
-
for token in self.text_layer:
|
233
|
-
yield token
|
234
|
-
|
235
|
-
def get_terms(self):
|
236
|
-
for term in self.term_layer:
|
237
|
-
yield term
|
238
|
-
|
239
|
-
def get_token(self,token_id):
|
240
|
-
if self.text_layer is not None:
|
241
|
-
return self.text_layer.get_wf(token_id)
|
242
|
-
else:
|
243
|
-
return None
|
244
|
-
|
245
|
-
def get_term(self,term_id):
|
246
|
-
if self.term_layer is not None:
|
247
|
-
return self.term_layer.get_term(term_id)
|
248
|
-
else:
|
249
|
-
return None
|
250
|
-
|
251
|
-
def get_properties(self):
|
252
|
-
if self.features_layer is not None:
|
253
|
-
for property in self.features_layer.get_properties():
|
254
|
-
yield property
|
255
|
-
|
256
|
-
def get_entities(self):
|
257
|
-
if self.entity_layer is not None:
|
258
|
-
for entity in self.entity_layer:
|
259
|
-
yield entity
|
260
|
-
|
261
|
-
def get_opinions(self):
|
262
|
-
if self.opinion_layer is not None:
|
263
|
-
for opinion in self.opinion_layer.get_opinions():
|
264
|
-
yield opinion
|
265
|
-
|
266
|
-
|
267
|
-
def dump(self,filename=sys.stdout):
|
268
|
-
self.tree.write(filename,encoding='UTF-8',pretty_print=True,xml_declaration=True)
|
269
|
-
|
270
|
-
|
271
|
-
def remove_opinion_layer(self):
|
272
|
-
if self.opinion_layer is not None:
|
273
|
-
this_node = self.opinion_layer.get_node()
|
274
|
-
self.root.remove(this_node)
|
275
|
-
self.opinion_layer = None
|
276
|
-
|
277
|
-
def remove_term_layer(self):
|
278
|
-
if self.term_layer is not None:
|
279
|
-
this_node = self.term_layer.get_node()
|
280
|
-
self.root.remove(this_node)
|
281
|
-
self.term_layer = None
|
282
|
-
|
283
|
-
if self.header is not None:
|
284
|
-
self.header.remove_lp('terms')
|
285
|
-
|
286
|
-
def get_constituency_extractor(self):
|
287
|
-
if self.constituency_layer is not None: ##Otherwise there are no constituens
|
288
|
-
if self.my_constituency_extractor is None:
|
289
|
-
self.my_constituency_extractor = Cconstituency_extractor(self)
|
290
|
-
return self.my_constituency_extractor
|
291
|
-
else:
|
292
|
-
return None
|
293
|
-
|
294
|
-
def get_dependency_extractor(self):
|
295
|
-
if self.dependency_layer is not None: #otherwise there are no dependencies
|
296
|
-
if self.my_dependency_extractor is None:
|
297
|
-
self.my_dependency_extractor = Cdependency_extractor(self)
|
298
|
-
return self.my_dependency_extractor
|
299
|
-
else:
|
300
|
-
return None
|
301
|
-
|
302
|
-
## ADDING METHODS
|
303
|
-
def add_wf(self,wf_obj):
|
304
|
-
if self.text_layer is None:
|
305
|
-
self.text_layer = Ctext(type=self.type)
|
306
|
-
self.root.append(self.text_layer.get_node())
|
307
|
-
self.text_layer.add_wf(wf_obj)
|
308
|
-
|
309
|
-
def add_opinion(self,opinion_obj):
|
310
|
-
if self.opinion_layer is None:
|
311
|
-
self.opinion_layer = Copinions()
|
312
|
-
self.root.append(self.opinion_layer.get_node())
|
313
|
-
self.opinion_layer.add_opinion(opinion_obj)
|
314
|
-
|
315
|
-
def add_linguistic_processor(self, layer ,my_lp):
|
316
|
-
self.header.add_linguistic_processor(layer,my_lp)
|
317
|
-
|
318
|
-
|
319
|
-
def add_dependency(self,my_dep):
|
320
|
-
if self.dependency_layer is None:
|
321
|
-
self.dependency_layer = Cdependencies()
|
322
|
-
self.root.append(self.dependency_layer.get_node())
|
323
|
-
self.dependency_layer.add_dependency(my_dep)
|
324
|
-
|
325
|
-
|
326
|
-
## EXTRA FUNCTIONS
|
327
|
-
## Gets the token identifiers in the span of a term id
|
328
|
-
def get_dict_tokens_for_termid(self, term_id):
|
329
|
-
if self.dict_tokens_for_tid is None:
|
330
|
-
self.dict_tokens_for_tid = {}
|
331
|
-
for term in self.get_terms():
|
332
|
-
self.dict_tokens_for_tid[term.get_id()] = term.get_span().get_span_ids()
|
333
|
-
|
334
|
-
return self.dict_tokens_for_tid.get(term_id,[])
|
335
|
-
|
336
|
-
|
337
|
-
|
338
|
-
|
@@ -1,14 +0,0 @@
|
|
1
|
-
from KafNafParserMod import *
|
2
|
-
from header_data import *
|
3
|
-
from external_references_data import *
|
4
|
-
from span_data import *
|
5
|
-
from term_data import *
|
6
|
-
from term_sentiment_data import *
|
7
|
-
from text_data import *
|
8
|
-
from entity_data import *
|
9
|
-
from features_data import *
|
10
|
-
from opinion_data import *
|
11
|
-
from dependency_data import *
|
12
|
-
from constituency_data import *
|
13
|
-
from references_data import *
|
14
|
-
from coreference_data import *
|