muck-raker 0.3.1 → 0.3.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/.gitignore +29 -0
- data/Rakefile +3 -3
- data/VERSION +1 -1
- data/config/muck_raker_environment.rb +18 -16
- data/lib/muck_raker/tasks.rb +64 -246
- data/muck-raker.gemspec +4 -3
- data/rails/init.rb +0 -2
- data/raker/lib/aggregatord.jar +0 -0
- data/raker/lib/recommenderd.jar +0 -0
- data/raker/lib/recommenderd.properties +68 -68
- metadata +3 -2
data/.gitignore
ADDED
@@ -0,0 +1,29 @@
|
|
1
|
+
**/*.pid
|
2
|
+
log/*.log
|
3
|
+
log/*.pid
|
4
|
+
tmp
|
5
|
+
.DS_Store
|
6
|
+
public/cache/**/*
|
7
|
+
public/system/**/*
|
8
|
+
doc/**/*
|
9
|
+
db/*.sqlite3
|
10
|
+
.project
|
11
|
+
.loadpath
|
12
|
+
nbproject/
|
13
|
+
.idea
|
14
|
+
testjour.log
|
15
|
+
*.so
|
16
|
+
*.o
|
17
|
+
Makefile
|
18
|
+
mkmf.log
|
19
|
+
*.bundle
|
20
|
+
conftest
|
21
|
+
content/
|
22
|
+
.idea
|
23
|
+
*.sw?
|
24
|
+
.DS_Store
|
25
|
+
coverage
|
26
|
+
rdoc
|
27
|
+
pkg
|
28
|
+
pkg/*
|
29
|
+
log/*
|
data/Rakefile
CHANGED
@@ -35,10 +35,10 @@ begin
|
|
35
35
|
gemspec.files.exclude "test/*"
|
36
36
|
gemspec.test_files.exclude 'test/feed_archive/**'
|
37
37
|
end
|
38
|
+
Jeweler::RubyforgeTasks.new do |rubyforge|
|
39
|
+
rubyforge.doc_task = "rdoc"
|
40
|
+
end
|
38
41
|
Jeweler::GemcutterTasks.new
|
39
|
-
# Jeweler::RubyforgeTasks.new do |rubyforge|
|
40
|
-
# rubyforge.doc_task = "rdoc"
|
41
|
-
# end
|
42
42
|
rescue LoadError
|
43
43
|
puts "Jeweler not available. Install it with: sudo gem install jeweler"
|
44
44
|
end
|
data/VERSION
CHANGED
@@ -1 +1 @@
|
|
1
|
-
0.3.
|
1
|
+
0.3.2
|
@@ -1,6 +1,4 @@
|
|
1
1
|
ENV['RAILS_ENV'] = (ENV['RAILS_ENV'] || 'development').dup
|
2
|
-
dir = File.dirname(__FILE__)
|
3
|
-
SOLR_PATH = File.expand_path("#{dir}/../solr") unless defined? SOLR_PATH
|
4
2
|
|
5
3
|
unless defined? SOLR_LOGS_PATH
|
6
4
|
SOLR_LOGS_PATH = ENV["SOLR_LOGS_PATH"] || "#{RAILS_ROOT}/log"
|
@@ -9,32 +7,36 @@ unless defined? SOLR_PIDS_PATH
|
|
9
7
|
SOLR_PIDS_PATH = ENV["SOLR_PIDS_PATH"] || "#{RAILS_ROOT}/tmp/pids/solr"
|
10
8
|
end
|
11
9
|
unless defined? SOLR_DATA_PATH
|
12
|
-
SOLR_DATA_PATH = ENV["SOLR_DATA_PATH"] || "#{RAILS_ROOT}/
|
10
|
+
SOLR_DATA_PATH = ENV["SOLR_DATA_PATH"] || "#{RAILS_ROOT}/solr_indexes/#{ENV['RAILS_ENV']}"
|
13
11
|
end
|
14
12
|
unless defined? SOLR_CONFIG_PATH
|
15
|
-
SOLR_CONFIG_PATH = ENV["SOLR_CONFIG_PATH"] ||
|
13
|
+
SOLR_CONFIG_PATH = ENV["SOLR_CONFIG_PATH"] || File.join(RAILS_ROOT, 'config', 'solr')
|
16
14
|
end
|
17
15
|
|
16
|
+
unless defined? RAKER_LOGS_PATH
|
17
|
+
RAKER_LOGS_PATH = File.join(RAILS_ROOT, 'log')
|
18
|
+
end
|
18
19
|
unless defined? RAKER_PIDS_PATH
|
19
20
|
RAKER_PIDS_PATH = ENV["RAKER_PIDS_PATH"] || "#{RAILS_ROOT}/tmp/pids/raker"
|
20
21
|
end
|
21
|
-
|
22
|
-
|
22
|
+
unless defined? RAKER_FEED_ARCHIVE_PATH
|
23
|
+
if ENV['RAILS_ENV'] == 'production'
|
24
|
+
RAKER_FEED_ARCHIVE_PATH = File.join(RAILS_ROOT, '..', '..', 'shared', 'feed_archive')
|
25
|
+
else
|
26
|
+
RAKER_FEED_ARCHIVE_PATH = File.join(RAILS_ROOT, '..', 'feed_archive')
|
27
|
+
end
|
28
|
+
end
|
29
|
+
unless defined? RAKER_DATABASE_CONFIG
|
23
30
|
if ENV['RAILS_ENV'] == 'production'
|
24
|
-
|
31
|
+
RAKER_DATABASE_CONFIG = File.join(RAILS_ROOT, '..', '..', 'shared', 'config', 'database.yml')
|
25
32
|
else
|
26
|
-
|
33
|
+
RAKER_DATABASE_CONFIG = File.join(RAILS_ROOT, 'config', 'database.yml')
|
27
34
|
end
|
28
35
|
end
|
29
|
-
unless defined?
|
36
|
+
unless defined? RAKER_LOG_TO_CONSOLE
|
30
37
|
if ENV['RAILS_ENV'] == 'production'
|
31
|
-
|
32
|
-
elsif ENV['RAILS_ENV'] == 'development'
|
33
|
-
FEED_ARCHIVE_PATH = File.join(RAILS_ROOT, '..', 'feed_archive')
|
38
|
+
RAKER_LOG_TO_CONSOLE = 'false'
|
34
39
|
else
|
35
|
-
|
40
|
+
RAKER_LOG_TO_CONSOLE = 'true'
|
36
41
|
end
|
37
42
|
end
|
38
|
-
unless defined? LOG_FILE_PREFIX
|
39
|
-
LOG_FILE_PREFIX = File.join(RAILS_ROOT, 'log', 'muck-raker')
|
40
|
-
end
|
data/lib/muck_raker/tasks.rb
CHANGED
@@ -75,7 +75,6 @@ module MuckRaker
|
|
75
75
|
Service.delete_all
|
76
76
|
yml = File.join(File.dirname(__FILE__), '..', '..', 'db', 'bootstrap',"services")
|
77
77
|
Fixtures.new(Service.connection,"services",Service,yml).insert_fixtures
|
78
|
-
|
79
78
|
end
|
80
79
|
|
81
80
|
end
|
@@ -87,22 +86,31 @@ module MuckRaker
|
|
87
86
|
system "rsync -ruv #{path}/public ."
|
88
87
|
system "rsync -ruv #{path}/config/solr ./config"
|
89
88
|
end
|
90
|
-
|
91
|
-
|
92
|
-
|
93
|
-
|
89
|
+
|
90
|
+
def pid_file
|
91
|
+
"#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid"
|
92
|
+
end
|
93
|
+
|
94
|
+
def raker_task task = 'daemon', param = ''
|
94
95
|
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
96
|
+
show_options
|
97
|
+
FileUtils.mkdir_p(RAKER_PIDS_PATH)
|
98
|
+
FileUtils.mkdir_p(RAKER_LOGS_PATH)
|
95
99
|
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
96
100
|
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
97
101
|
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
98
|
-
jars = Dir['
|
99
|
-
options = "-
|
100
|
-
options << "-
|
101
|
-
options << "-
|
102
|
-
options << "-
|
102
|
+
jars = Dir['*.jar','solr/*.jar'].join(separator)
|
103
|
+
options = "-DRAILS_ENV=#{ENV['RAILS_ENV']} "
|
104
|
+
options << "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" "
|
105
|
+
options << "-Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" "
|
106
|
+
options << "-Draker.database.config=\"#{RAKER_DATABASE_CONFIG}\" "
|
107
|
+
options << "-Draker.log.dir=\"#{RAKER_LOGS_PATH}\" "
|
108
|
+
options << "-Draker.feed_archive_dir=\"#{RAKER_FEED_ARCHIVE_PATH}\" "
|
109
|
+
options << "-Draker.log_to_console=\"#{RAKER_LOG_TO_CONSOLE}\" "
|
103
110
|
classpath = "-classpath #{jars}#{separator}. "
|
104
|
-
|
105
|
-
|
111
|
+
memory_options = "-Xms32m -Xmx128m "
|
112
|
+
javaclass = "edu.usu.cosl.recommenderd.Recommenderd "
|
113
|
+
cmd = "java " + options + classpath + memory_options + javaclass + task + " " + param
|
106
114
|
puts ("Executing: " + cmd)
|
107
115
|
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
108
116
|
if windows
|
@@ -113,277 +121,87 @@ module MuckRaker
|
|
113
121
|
end
|
114
122
|
end
|
115
123
|
sleep(5)
|
116
|
-
File.open(
|
124
|
+
File.open(pid_file, "w"){ |f| f << pid} unless windows
|
117
125
|
puts "#{ENV['RAILS_ENV']} Muck raker harvest command started successfully, pid: #{pid}."
|
118
126
|
end
|
119
127
|
end
|
128
|
+
|
129
|
+
def show_options
|
130
|
+
puts "RAILS_ENV=#{ENV['RAILS_ENV']} "
|
131
|
+
puts "solr.solr.home=\"#{SOLR_CONFIG_PATH}\" "
|
132
|
+
puts "solr.data.dir=\"#{SOLR_DATA_PATH}\" "
|
133
|
+
puts "raker.db.config=\"#{RAKER_DATABASE_CONFIG}\" "
|
134
|
+
puts "raker.log.dir=\"#{RAKER_LOGS_PATH}\" "
|
135
|
+
puts "raker.feed_archive_dir=\"#{RAKER_FEED_ARCHIVE_PATH}\" "
|
136
|
+
puts "raker.log_to_console=\"#{RAKER_LOG_TO_CONSOLE}\" "
|
137
|
+
end
|
138
|
+
|
139
|
+
desc "Print out muck raker dependent environment variables"
|
140
|
+
task :show_options => :environment do
|
141
|
+
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
142
|
+
show_options
|
143
|
+
end
|
120
144
|
|
121
145
|
desc "Start the recommender daemon process"
|
122
146
|
task :start => :environment do
|
123
|
-
|
124
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
125
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
126
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
127
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
128
|
-
jars = Dir['**','*.jar'].join(separator)
|
129
|
-
puts Dir['**','*.jar']
|
130
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
131
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
132
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
133
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
134
|
-
classpath = "-classpath #{jars}#{separator}. "
|
135
|
-
javaclass = "edu.usu.cosl.recommenderd.Recommenderd"
|
136
|
-
cmd = "java " + options + classpath + javaclass
|
137
|
-
puts ("Executing: " + cmd)
|
138
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
139
|
-
if windows
|
140
|
-
exec cmd
|
141
|
-
else
|
142
|
-
pid = fork do
|
143
|
-
exec cmd
|
144
|
-
end
|
145
|
-
end
|
146
|
-
sleep(5)
|
147
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
148
|
-
puts "#{ENV['RAILS_ENV']} Muck raker start command started successfully, pid: #{pid}."
|
149
|
-
end
|
147
|
+
raker_task
|
150
148
|
end
|
151
149
|
|
152
|
-
desc "
|
153
|
-
task :
|
154
|
-
|
155
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
156
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
157
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
158
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
159
|
-
jars = Dir['**','*.jar'].join(separator)
|
160
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
161
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
162
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
163
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
164
|
-
classpath = "-classpath #{jars}#{separator}. "
|
165
|
-
memory_options = "-Xms32m -Xmx128m "
|
166
|
-
javaclass = "edu.usu.cosl.indexer.Indexer "
|
167
|
-
cmdlineoption = "all "
|
168
|
-
cmd = "java " + options + classpath + memory_options + javaclass + cmdlineoption
|
169
|
-
puts ("Executing: " + cmd)
|
170
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
171
|
-
if windows
|
172
|
-
exec cmd
|
173
|
-
else
|
174
|
-
pid = fork do
|
175
|
-
exec cmd
|
176
|
-
end
|
177
|
-
end
|
178
|
-
sleep(5)
|
179
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
180
|
-
puts "#{ENV['RAILS_ENV']} Muck raker reindex command started successfully, pid: #{pid}."
|
181
|
-
end
|
150
|
+
desc "Redo everything (re-index, redo autogenerated subjects, rebuild tag clouds re-recommend)"
|
151
|
+
task :rebuild => :environment do
|
152
|
+
raker_task 'daemon', 'full'
|
182
153
|
end
|
183
154
|
|
184
|
-
desc "
|
185
|
-
task :
|
186
|
-
|
187
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
188
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
189
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
190
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
191
|
-
jars = Dir['**','*.jar'].join(separator)
|
192
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
193
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
194
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
195
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
196
|
-
classpath = "-classpath #{jars}#{separator}. "
|
197
|
-
javaclass = "edu.usu.cosl.tagclouds.SubjectAutoGenerator "
|
198
|
-
cmd = "java " + options + classpath + javaclass
|
199
|
-
puts ("Executing: " + cmd)
|
200
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
201
|
-
if windows
|
202
|
-
exec cmd
|
203
|
-
else
|
204
|
-
pid = fork do
|
205
|
-
exec cmd
|
206
|
-
end
|
207
|
-
end
|
208
|
-
sleep(5)
|
209
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
210
|
-
puts "#{ENV['RAILS_ENV']} Muck raker gen_subjects command started successfully, pid: #{pid}."
|
211
|
-
end
|
155
|
+
desc "Get some data into the recommender system"
|
156
|
+
task :bootstrap => :environment do
|
157
|
+
raker_task 'bootstrap'
|
212
158
|
end
|
213
159
|
|
214
|
-
desc "
|
215
|
-
task :
|
216
|
-
|
217
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
218
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
219
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
220
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
221
|
-
jars = Dir['**','*.jar'].join(separator)
|
222
|
-
options = "-DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
223
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
224
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
225
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
226
|
-
classpath = "-classpath #{jars}#{separator}. "
|
227
|
-
javaclass = "edu.usu.cosl.tagclouds.TagCloud"
|
228
|
-
cmd = "java " + options + classpath + javaclass
|
229
|
-
puts ("Executing: " + cmd)
|
230
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
231
|
-
if windows
|
232
|
-
exec cmd
|
233
|
-
else
|
234
|
-
pid = fork do
|
235
|
-
exec cmd
|
236
|
-
end
|
237
|
-
end
|
238
|
-
sleep(5)
|
239
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
240
|
-
puts "#{ENV['RAILS_ENV']} Muck raker get_clouds command started successfully, pid: #{pid}."
|
241
|
-
end
|
160
|
+
desc "Harvest without recommending"
|
161
|
+
task :harvest => :environment do
|
162
|
+
raker_task 'harvest'
|
242
163
|
end
|
243
164
|
|
244
|
-
desc "
|
245
|
-
task :
|
246
|
-
|
247
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
248
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
249
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
250
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
251
|
-
jars = Dir['**','*.jar'].join(separator)
|
252
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
253
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
254
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
255
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
256
|
-
classpath = "-classpath #{jars}#{separator}. "
|
257
|
-
javaclass = "edu.usu.cosl.recommenderd.Recommenderd "
|
258
|
-
cmdlineoption = "skip_harvest "
|
259
|
-
cmd = "java " + options + classpath + javaclass + cmdlineoption
|
260
|
-
puts ("Executing: " + cmd)
|
261
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
262
|
-
if windows
|
263
|
-
exec cmd
|
264
|
-
else
|
265
|
-
pid = fork do
|
266
|
-
exec cmd
|
267
|
-
end
|
268
|
-
end
|
269
|
-
sleep(5)
|
270
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
271
|
-
puts "#{ENV['RAILS_ENV']} Muck raker recommend command started successfully, pid: #{pid}."
|
272
|
-
end
|
165
|
+
desc "Update solr index to changes made to recommender database"
|
166
|
+
task :index => :environment do
|
167
|
+
raker_task 'index'
|
273
168
|
end
|
274
169
|
|
275
|
-
desc "
|
276
|
-
task :
|
277
|
-
|
278
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
279
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
280
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
281
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
282
|
-
jars = Dir['**','*.jar'].join(separator)
|
283
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
284
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
285
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
286
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
287
|
-
classpath = "-classpath #{jars}#{separator}. "
|
288
|
-
javaclass = "edu.usu.cosl.recommender.Recommender "
|
289
|
-
cmdlineoption = "rebuild_cache "
|
290
|
-
cmd = "java " + options + classpath + javaclass + cmdlineoption
|
291
|
-
puts ("Executing: " + cmd)
|
292
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
293
|
-
if windows
|
294
|
-
exec cmd
|
295
|
-
else
|
296
|
-
pid = fork do
|
297
|
-
exec cmd
|
298
|
-
end
|
299
|
-
end
|
300
|
-
sleep(5)
|
301
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
302
|
-
puts "#{ENV['RAILS_ENV']} Muck raker rebuild_recommendation_cache command started successfully, pid: #{pid}."
|
303
|
-
end
|
170
|
+
desc "Recommend without harvesting"
|
171
|
+
task :recommend => :environment do
|
172
|
+
raker_task 'recommend'
|
304
173
|
end
|
305
174
|
|
306
175
|
desc "Redo recommendations only"
|
307
176
|
task :redo_recommendations => :environment do
|
308
|
-
|
309
|
-
require File.expand_path("#{File.dirname(__FILE__)}/../../config/muck_raker_environment")
|
310
|
-
separator = (RUBY_PLATFORM =~ /(win|w)32$/ ? ';' : ':')
|
311
|
-
puts "RAILS_ENV=" + ENV['RAILS_ENV']
|
312
|
-
Dir.chdir(File.join(File.dirname(__FILE__), '../../', 'raker', 'lib')) do
|
313
|
-
jars = Dir['**','*.jar'].join(separator)
|
314
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
315
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
316
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
317
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
318
|
-
classpath = "-classpath #{jars}#{separator}. "
|
319
|
-
javaclass = "edu.usu.cosl.recommender.Recommender "
|
320
|
-
cmd = "java " + options + classpath + javaclass
|
321
|
-
puts ("Executing: " + cmd)
|
322
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
323
|
-
if windows
|
324
|
-
exec cmd
|
325
|
-
else
|
326
|
-
pid = fork do
|
327
|
-
exec cmd
|
328
|
-
end
|
329
|
-
end
|
330
|
-
sleep(5)
|
331
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
332
|
-
puts "#{ENV['RAILS_ENV']} Muck raker redo_recommendations command started successfully, pid: #{pid}."
|
333
|
-
end
|
177
|
+
raker_task 'recommend', 'full'
|
334
178
|
end
|
335
179
|
|
336
|
-
desc "
|
337
|
-
task :
|
338
|
-
|
339
|
-
|
340
|
-
|
341
|
-
|
342
|
-
|
343
|
-
|
344
|
-
options = "-Dsolr.solr.home=\"#{SOLR_CONFIG_PATH}\" -Dsolr.data.dir=\"#{SOLR_DATA_PATH}\" -DRAILS_ENV=#{ENV['RAILS_ENV']} -DRAILS_DB_CONFIG=\"#{RAILS_DB_CONFIG}\" -DLOG_FILE_PREFIX=\"#{LOG_FILE_PREFIX}\" "
|
345
|
-
options << "-DDEBUG=true " unless ENV['DEBUG'] == 'false'
|
346
|
-
options << "-DLOG_TO_CONSOLE=true " unless ENV['DEBUG'] == 'false'
|
347
|
-
options << "-DFEED_ARCHIVE_PATH=\"#{FEED_ARCHIVE_PATH}\" "
|
348
|
-
classpath = "-classpath #{jars}#{separator}. "
|
349
|
-
javaclass = "edu.usu.cosl.recommenderd.Recommenderd "
|
350
|
-
cmdlineoption = "rebuild "
|
351
|
-
cmd = "java " + options + classpath + javaclass + cmdlineoption
|
352
|
-
puts ("Executing: " + cmd)
|
353
|
-
windows = RUBY_PLATFORM =~ /(win|w)32$/
|
354
|
-
if windows
|
355
|
-
exec cmd
|
356
|
-
else
|
357
|
-
pid = fork do
|
358
|
-
exec cmd
|
359
|
-
end
|
360
|
-
end
|
361
|
-
sleep(5)
|
362
|
-
File.open("#{RAKER_PIDS_PATH}/#{ENV['RAILS_ENV']}_pid", "w"){ |f| f << pid} unless windows
|
363
|
-
puts "#{ENV['RAILS_ENV']} Muck raker rebuild command started successfully, pid: #{pid}."
|
364
|
-
end
|
180
|
+
desc "Autogenerate subjects"
|
181
|
+
task :subjects => :environment do
|
182
|
+
raker_task 'auto_generate_subjects'
|
183
|
+
end
|
184
|
+
|
185
|
+
desc "Generate tag clouds"
|
186
|
+
task :tag_clouds => :environment do
|
187
|
+
raker_task 'tag_clouds'
|
365
188
|
end
|
366
189
|
|
367
190
|
desc "Stop a raker daemon process"
|
368
191
|
task :stop => :environment do
|
369
|
-
file_path =
|
192
|
+
file_path = pid_file
|
370
193
|
if File.exists?(file_path)
|
371
194
|
File.open(file_path, "r") do |f|
|
372
195
|
pid = f.readline
|
373
196
|
Process.kill('TERM', pid.to_i)
|
374
197
|
end
|
375
198
|
File.unlink(file_path)
|
376
|
-
puts "Raker
|
199
|
+
puts "Raker task successfully."
|
377
200
|
else
|
378
201
|
puts "PID file not found at #{file_path}. Either Raker is not running or no PID file was written."
|
379
202
|
end
|
380
203
|
end
|
381
204
|
|
382
|
-
desc "Restart the recommender daemon process"
|
383
|
-
task :restart do
|
384
|
-
system "java"
|
385
|
-
end
|
386
|
-
|
387
205
|
end
|
388
206
|
|
389
207
|
end
|
data/muck-raker.gemspec
CHANGED
@@ -5,18 +5,19 @@
|
|
5
5
|
|
6
6
|
Gem::Specification.new do |s|
|
7
7
|
s.name = %q{muck-raker}
|
8
|
-
s.version = "0.3.
|
8
|
+
s.version = "0.3.2"
|
9
9
|
|
10
10
|
s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
|
11
11
|
s.authors = ["Joel Duffin Justin Ball"]
|
12
|
-
s.date = %q{2009-11-
|
12
|
+
s.date = %q{2009-11-05}
|
13
13
|
s.description = %q{The aggregation and recommendation daemons for the muck system.}
|
14
14
|
s.email = %q{justinball@gmail.com}
|
15
15
|
s.extra_rdoc_files = [
|
16
16
|
"README.rdoc"
|
17
17
|
]
|
18
18
|
s.files = [
|
19
|
-
"
|
19
|
+
".gitignore",
|
20
|
+
"MIT-LICENSE",
|
20
21
|
"README.rdoc",
|
21
22
|
"Rakefile",
|
22
23
|
"VERSION",
|
data/rails/init.rb
CHANGED
data/raker/lib/aggregatord.jar
CHANGED
Binary file
|
data/raker/lib/recommenderd.jar
CHANGED
Binary file
|
@@ -1,68 +1,68 @@
|
|
1
|
-
# number of worker threads to spawn
|
2
|
-
#threads=20
|
3
|
-
|
4
|
-
# size of the queue
|
5
|
-
#stale_feed_queue_size=100
|
6
|
-
|
7
|
-
# seconds to wait before timing out a HTTP request
|
8
|
-
#request_timeout=30
|
9
|
-
|
10
|
-
# seconds to wait before querying the database for stale feeds
|
11
|
-
#stale_thread_poll_interval=5
|
12
|
-
|
13
|
-
# minutes to wait before checking to see if a feed has new entries
|
14
|
-
#feed_refresh_interval=60
|
15
|
-
|
16
|
-
# whether or not to log to file
|
17
|
-
#log_file_prefix=recommender
|
18
|
-
|
19
|
-
# pairs to consider when calculating recommendations; default is 200
|
20
|
-
#pairs_to_consider=200
|
21
|
-
|
22
|
-
# maximum number of recommendations to calculate; default is 20
|
23
|
-
#max_recommendations=20
|
24
|
-
|
25
|
-
# smtp server to use to email admin of results
|
26
|
-
#smtp_server=
|
27
|
-
|
28
|
-
# email address to send reports to
|
29
|
-
#admin_email=
|
30
|
-
|
31
|
-
# harvest before running recommendations (default is true)
|
32
|
-
#harvest=true
|
33
|
-
|
34
|
-
# redo all recommendations (default is false)
|
35
|
-
#redo_recommendations=false
|
36
|
-
|
37
|
-
# reindex all of the entries in the DB (default is false)
|
38
|
-
#reindex_all=false
|
39
|
-
|
40
|
-
# solr home directory (default is config/solr)
|
41
|
-
#solr_dir=c:/projects/oerrecommender/config/solr/
|
42
|
-
#solr_config_filename=../../projects/muck/config/solr/solr.xml
|
43
|
-
#solr_dir=../muck/config/solr
|
44
|
-
|
45
|
-
# rails environment (default is development)
|
46
|
-
#rails_env=development
|
47
|
-
|
48
|
-
# path to database.yml (default is config/database.yml)
|
49
|
-
db_yml=../../projects/muck/config/database.yml
|
50
|
-
|
51
|
-
# logging
|
52
|
-
log4j.rootLogger=trace, stdout, R
|
53
|
-
|
54
|
-
log4j.appender.stdout=org.apache.log4j.ConsoleAppender
|
55
|
-
log4j.appender.stdout.layout=org.apache.log4j.PatternLayout
|
56
|
-
|
57
|
-
# Pattern to output the caller's file name and line number.
|
58
|
-
log4j.appender.stdout.layout.ConversionPattern=%5p [%t] (%F:%L) - %m%n
|
59
|
-
|
60
|
-
log4j.appender.R=org.apache.log4j.RollingFileAppender
|
61
|
-
log4j.appender.R.File=log/recommenderd.log
|
62
|
-
|
63
|
-
log4j.appender.R.MaxFileSize=100KB
|
64
|
-
# Keep one backup file
|
65
|
-
log4j.appender.R.MaxBackupIndex=1
|
66
|
-
|
67
|
-
log4j.appender.R.layout=org.apache.log4j.PatternLayout
|
68
|
-
log4j.appender.R.layout.ConversionPattern=%p %t %c - %m%n
|
1
|
+
# number of worker threads to spawn
|
2
|
+
#threads=20
|
3
|
+
|
4
|
+
# size of the queue
|
5
|
+
#stale_feed_queue_size=100
|
6
|
+
|
7
|
+
# seconds to wait before timing out a HTTP request
|
8
|
+
#request_timeout=30
|
9
|
+
|
10
|
+
# seconds to wait before querying the database for stale feeds
|
11
|
+
#stale_thread_poll_interval=5
|
12
|
+
|
13
|
+
# minutes to wait before checking to see if a feed has new entries
|
14
|
+
#feed_refresh_interval=60
|
15
|
+
|
16
|
+
# whether or not to log to file
|
17
|
+
#log_file_prefix=recommender
|
18
|
+
|
19
|
+
# pairs to consider when calculating recommendations; default is 200
|
20
|
+
#pairs_to_consider=200
|
21
|
+
|
22
|
+
# maximum number of recommendations to calculate; default is 20
|
23
|
+
#max_recommendations=20
|
24
|
+
|
25
|
+
# smtp server to use to email admin of results
|
26
|
+
#smtp_server=
|
27
|
+
|
28
|
+
# email address to send reports to
|
29
|
+
#admin_email=
|
30
|
+
|
31
|
+
# harvest before running recommendations (default is true)
|
32
|
+
#harvest=true
|
33
|
+
|
34
|
+
# redo all recommendations (default is false)
|
35
|
+
#redo_recommendations=false
|
36
|
+
|
37
|
+
# reindex all of the entries in the DB (default is false)
|
38
|
+
#reindex_all=false
|
39
|
+
|
40
|
+
# solr home directory (default is config/solr)
|
41
|
+
#solr_dir=c:/projects/oerrecommender/config/solr/
|
42
|
+
#solr_config_filename=../../projects/muck/config/solr/solr.xml
|
43
|
+
#solr_dir=../muck/config/solr
|
44
|
+
|
45
|
+
# rails environment (default is development)
|
46
|
+
#rails_env=development
|
47
|
+
|
48
|
+
# path to database.yml (default is config/database.yml)
|
49
|
+
db_yml=../../projects/muck/config/database.yml
|
50
|
+
|
51
|
+
# logging
|
52
|
+
log4j.rootLogger=trace, stdout, R
|
53
|
+
|
54
|
+
log4j.appender.stdout=org.apache.log4j.ConsoleAppender
|
55
|
+
log4j.appender.stdout.layout=org.apache.log4j.PatternLayout
|
56
|
+
|
57
|
+
# Pattern to output the caller's file name and line number.
|
58
|
+
log4j.appender.stdout.layout.ConversionPattern=%5p [%t] (%F:%L) - %m%n
|
59
|
+
|
60
|
+
log4j.appender.R=org.apache.log4j.RollingFileAppender
|
61
|
+
log4j.appender.R.File=log/recommenderd.log
|
62
|
+
|
63
|
+
log4j.appender.R.MaxFileSize=100KB
|
64
|
+
# Keep one backup file
|
65
|
+
log4j.appender.R.MaxBackupIndex=1
|
66
|
+
|
67
|
+
log4j.appender.R.layout=org.apache.log4j.PatternLayout
|
68
|
+
log4j.appender.R.layout.ConversionPattern=%p %t %c - %m%n
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: muck-raker
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.3.
|
4
|
+
version: 0.3.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Joel Duffin Justin Ball
|
@@ -9,7 +9,7 @@ autorequire:
|
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
11
|
|
12
|
-
date: 2009-11-
|
12
|
+
date: 2009-11-05 00:00:00 -07:00
|
13
13
|
default_executable:
|
14
14
|
dependencies: []
|
15
15
|
|
@@ -22,6 +22,7 @@ extensions: []
|
|
22
22
|
extra_rdoc_files:
|
23
23
|
- README.rdoc
|
24
24
|
files:
|
25
|
+
- .gitignore
|
25
26
|
- MIT-LICENSE
|
26
27
|
- README.rdoc
|
27
28
|
- Rakefile
|