stellr 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/History.txt +4 -0
- data/Manifest.txt +36 -0
- data/README.txt +109 -0
- data/Rakefile +28 -0
- data/bin/stellr +64 -0
- data/bin/stellr-search +50 -0
- data/config/stellr.yml +8 -0
- data/lib/stellr.rb +37 -0
- data/lib/stellr/client.rb +78 -0
- data/lib/stellr/collections.rb +6 -0
- data/lib/stellr/collections/base.rb +79 -0
- data/lib/stellr/collections/multi_collection.rb +32 -0
- data/lib/stellr/collections/rsync.rb +38 -0
- data/lib/stellr/collections/searchable_collection.rb +166 -0
- data/lib/stellr/collections/static.rb +97 -0
- data/lib/stellr/collections/writeable_collection.rb +119 -0
- data/lib/stellr/config.rb +88 -0
- data/lib/stellr/search.rb +2 -0
- data/lib/stellr/search/search_result.rb +21 -0
- data/lib/stellr/search/search_results.rb +50 -0
- data/lib/stellr/server.rb +166 -0
- data/lib/stellr/strategies.rb +4 -0
- data/lib/stellr/strategies/base.rb +16 -0
- data/lib/stellr/strategies/blocking.rb +13 -0
- data/lib/stellr/strategies/queueing.rb +78 -0
- data/lib/stellr/utils.rb +24 -0
- data/lib/stellr/utils/observable.rb +20 -0
- data/lib/stellr/utils/shutdown.rb +30 -0
- data/test/fixtures/movies.yml +4 -0
- data/test/stellr_test.rb +38 -0
- data/test/test_client.rb +27 -0
- data/test/test_collections_base.rb +25 -0
- data/test/test_helper.rb +1 -0
- data/test/test_rsync_collection.rb +72 -0
- data/test/test_server.rb +94 -0
- data/test/test_static_collection.rb +40 -0
- data/test/test_stellr.rb +11 -0
- metadata +110 -0
@@ -0,0 +1,88 @@
|
|
1
|
+
require 'yaml'
|
2
|
+
require 'erb'
|
3
|
+
|
4
|
+
module Stellr
|
5
|
+
# = Config
|
6
|
+
#
|
7
|
+
# The +stellr+ and +stellr-search+ commands both rely on a YAML file
|
8
|
+
# to provide some basic configuration:
|
9
|
+
#
|
10
|
+
# +:port+ Port the server should listen to
|
11
|
+
# +:host+ Hostname or IP of the server
|
12
|
+
# +:script+ Optional ruby file to load during startup. This is the place to load custom code like self made Analyzers you intend to use.
|
13
|
+
# +:log_level+ Log level, default is +:debug+
|
14
|
+
# +:base_dir+ Base directory where the server will store index data, log
|
15
|
+
# files and configuration data.
|
16
|
+
# +:data_dir+ Index directory, relative to +base_dir+. Defaults to +data+
|
17
|
+
# +:log_dir+ Log file directory, relative to +base_dir+. Defaults to +log+
|
18
|
+
# +:conf_dir+ Configuration directory, relative to +base_dir+. Defaults to +conf+. Here stellr will keep the configuration of registered collections, one YAML file per collection.
|
19
|
+
# +:tmp_dir+ Temp directory, relative to +base_dir+. Defaults to +tmp+
|
20
|
+
#
|
21
|
+
#
|
22
|
+
class Config
|
23
|
+
DEFAULTS = { :port => 9010,
|
24
|
+
:host => 'localhost',
|
25
|
+
:base_dir => '/var/stellr',
|
26
|
+
:data_dir => 'data',
|
27
|
+
:log_dir => 'log',
|
28
|
+
:tmp_dir => 'tmp',
|
29
|
+
:conf_dir => 'conf',
|
30
|
+
:log_level => :warn }
|
31
|
+
|
32
|
+
|
33
|
+
# Configfile search order:
|
34
|
+
# - argument
|
35
|
+
# - /etc/stellr.yml
|
36
|
+
# - +gem_directory+/config/stellr.yml
|
37
|
+
#
|
38
|
+
def initialize( config_file = nil, extra_options = {} )
|
39
|
+
load_config config_file
|
40
|
+
@config.update extra_options
|
41
|
+
@config.each { |k,v| v.untaint } # we trust our config file
|
42
|
+
end
|
43
|
+
|
44
|
+
def data_dir
|
45
|
+
resolve_directory_name( :data_dir )
|
46
|
+
end
|
47
|
+
|
48
|
+
def log_dir
|
49
|
+
resolve_directory_name( :log_dir )
|
50
|
+
end
|
51
|
+
|
52
|
+
def tmp_dir
|
53
|
+
resolve_directory_name( :tmp_dir )
|
54
|
+
end
|
55
|
+
|
56
|
+
def conf_dir
|
57
|
+
resolve_directory_name( :conf_dir )
|
58
|
+
end
|
59
|
+
|
60
|
+
def collection_dir
|
61
|
+
end
|
62
|
+
|
63
|
+
def drb_uri
|
64
|
+
"druby://#{host}:#{port}"
|
65
|
+
end
|
66
|
+
|
67
|
+
protected
|
68
|
+
|
69
|
+
def method_missing( method_name, *args )
|
70
|
+
return @config[method_name] if @config.has_key?( method_name )
|
71
|
+
raise NameError.new( "unknown configuration key: #{method_name}" )
|
72
|
+
end
|
73
|
+
|
74
|
+
def resolve_directory_name( sub_dir )
|
75
|
+
raise NameError.new unless @config.has_key?( sub_dir )
|
76
|
+
File.join( base_dir, @config[sub_dir] )
|
77
|
+
end
|
78
|
+
|
79
|
+
def load_config( config_file )
|
80
|
+
config_file ||= "/etc/stellr.yml"
|
81
|
+
config_file = File.join( File.dirname(__FILE__), "../../config/stellr.yml" ) unless File.exists?( config_file )
|
82
|
+
|
83
|
+
@config = DEFAULTS.merge(
|
84
|
+
YAML.load( ERB.new( IO.read(config_file) ).result )
|
85
|
+
)
|
86
|
+
end
|
87
|
+
end
|
88
|
+
end
|
@@ -0,0 +1,21 @@
|
|
1
|
+
module Stellr
|
2
|
+
module Search
|
3
|
+
|
4
|
+
# A single search result
|
5
|
+
class SearchResult
|
6
|
+
attr_reader :score, :doc_id
|
7
|
+
|
8
|
+
def initialize(id, score, field_data)
|
9
|
+
@doc_id = id
|
10
|
+
@score = score
|
11
|
+
@field_data = field_data
|
12
|
+
end
|
13
|
+
|
14
|
+
# retrieve contents of the field +name+
|
15
|
+
def field(name)
|
16
|
+
@field_data[name]
|
17
|
+
end
|
18
|
+
alias [] field
|
19
|
+
end
|
20
|
+
end
|
21
|
+
end
|
@@ -0,0 +1,50 @@
|
|
1
|
+
module Stellr
|
2
|
+
module Search
|
3
|
+
|
4
|
+
# Thin wrapper around an array of search results
|
5
|
+
class SearchResults
|
6
|
+
attr_accessor :total_hits, :current_page, :per_page
|
7
|
+
alias total_entries total_hits
|
8
|
+
|
9
|
+
def initialize
|
10
|
+
@results = []
|
11
|
+
end
|
12
|
+
|
13
|
+
def method_missing(symbol, *args, &block)
|
14
|
+
@results.send(symbol, *args, &block)
|
15
|
+
end
|
16
|
+
|
17
|
+
def respond_to?(name)
|
18
|
+
self.methods.include?(name) || @results.respond_to?(name)
|
19
|
+
end
|
20
|
+
|
21
|
+
|
22
|
+
# code from here on derived from will_paginate's collection.rb
|
23
|
+
|
24
|
+
# The total number of pages.
|
25
|
+
def total_pages
|
26
|
+
@total_pages ||= per_page > 0 ? (total_hits / per_page.to_f).ceil : 0
|
27
|
+
end
|
28
|
+
alias page_count total_pages
|
29
|
+
|
30
|
+
# Current offset of the paginated collection. If we're on the first page,
|
31
|
+
# it is always 0. If we're on the 2nd page and there are 30 entries per page,
|
32
|
+
# the offset is 30. This property is useful if you want to render ordinals
|
33
|
+
# besides your records: simply start with offset + 1.
|
34
|
+
#
|
35
|
+
def offset
|
36
|
+
(current_page - 1) * per_page
|
37
|
+
end
|
38
|
+
|
39
|
+
# current_page - 1 or nil if there is no previous page
|
40
|
+
def previous_page
|
41
|
+
current_page > 1 ? (current_page - 1) : nil
|
42
|
+
end
|
43
|
+
|
44
|
+
# current_page + 1 or nil if there is no next page
|
45
|
+
def next_page
|
46
|
+
current_page < page_count ? (current_page + 1) : nil
|
47
|
+
end
|
48
|
+
end
|
49
|
+
end
|
50
|
+
end
|
@@ -0,0 +1,166 @@
|
|
1
|
+
require 'logger'
|
2
|
+
|
3
|
+
module Stellr
|
4
|
+
class Server
|
5
|
+
include Utils::Shutdown
|
6
|
+
protected :shutdown
|
7
|
+
|
8
|
+
attr_accessor :config
|
9
|
+
attr_reader :collections
|
10
|
+
|
11
|
+
def initialize( config )
|
12
|
+
@config = config
|
13
|
+
create_directories
|
14
|
+
@logger = Logger.new File.join(config.log_dir, 'stellr.log')
|
15
|
+
@logger.level = Logger.const_get config.log_level.to_s.upcase
|
16
|
+
@collections = {}
|
17
|
+
@collections.extend MonitorMixin
|
18
|
+
end
|
19
|
+
|
20
|
+
def add_record( collection_name, record, boost = nil )
|
21
|
+
collection( collection_name ).add_record record, boost
|
22
|
+
end
|
23
|
+
alias :<< :add_record
|
24
|
+
|
25
|
+
def delete_record( collection_name, record )
|
26
|
+
collection( collection_name ).delete_record record
|
27
|
+
end
|
28
|
+
|
29
|
+
def size( collection_name )
|
30
|
+
collection( collection_name ).size
|
31
|
+
end
|
32
|
+
|
33
|
+
# Initialize a collection.
|
34
|
+
#
|
35
|
+
# Before anything can be done with a collection, it has to be registered
|
36
|
+
# through this method. When called for a non-existing collection, this will
|
37
|
+
# also create the empty physical index. The given options are saved to a yml
|
38
|
+
# file so they can be loaded again later on.
|
39
|
+
#
|
40
|
+
# Calling register for an existing collection will update the saved index
|
41
|
+
# configuration from the options given, unless the options hash is nil.
|
42
|
+
# An already existing physical index won't be touched and used as is.
|
43
|
+
# Remember that changing Ferret options like analyzers or field
|
44
|
+
# configuration requires an index rebuild to be effective for existing
|
45
|
+
# content. Stellr doesn't take care of this.
|
46
|
+
#
|
47
|
+
# If you access the server through the Stellr::Client class you don't need
|
48
|
+
# to call +register+ explicitly as the client will do it when connecting.
|
49
|
+
#
|
50
|
+
# Name is the name of the collection to create.
|
51
|
+
#
|
52
|
+
# Options is a hash, consisting of:
|
53
|
+
# [+collection+] The collection implementation to use, may be one of :static or :rsync (default)
|
54
|
+
# [+strategy+] The strategy implementation to use (optional, atm there is only :queueing)
|
55
|
+
# [+fields+] +hash+ (see http://ferret.davebalmain.com/api/classes/Ferret/Index/FieldInfo.html)
|
56
|
+
# [+recreate+] Recreate the index (defaults to +false+). A true value will lead to the deletion of any already indexed data.
|
57
|
+
# [+analyzer+] The class name (String) of the Analyzer to use for this collection. By default, Ferret's StandardAnalyzer will be used.
|
58
|
+
# [+field_defaults+] Default setting for unconfigured fields
|
59
|
+
#
|
60
|
+
# Example
|
61
|
+
# register 'standard_index', { :recreate => false,
|
62
|
+
# :fields => { :author => { :index => :untokenized,
|
63
|
+
# :store => :no,
|
64
|
+
# :term_vector => :with_offsets,
|
65
|
+
# :boost => 2.0 },
|
66
|
+
# :content => { :index => :tokenized } }
|
67
|
+
# }
|
68
|
+
#
|
69
|
+
#
|
70
|
+
def register( name, options = {} )
|
71
|
+
@collections.synchronize do
|
72
|
+
@collections[name] ||= create_collection( name, options )
|
73
|
+
end
|
74
|
+
end
|
75
|
+
|
76
|
+
# Initializes a read-only virtual collection that may be used to search
|
77
|
+
# across multiple physical collections.
|
78
|
+
#
|
79
|
+
# Returns the name of the collection to be used with further calls.
|
80
|
+
def register_multi_collection( names, options = {} )
|
81
|
+
key = ".multi_#{names.join '_'}" # '.' is not allowed for regular collection names, so we are safe from name collisions
|
82
|
+
@collections.synchronize do
|
83
|
+
@collections[key] ||= create_multi_collection( key, names.map{ |name| collection(name) }, options )
|
84
|
+
end
|
85
|
+
return key
|
86
|
+
end
|
87
|
+
|
88
|
+
def collection( name )
|
89
|
+
@collections.synchronize do
|
90
|
+
if @collections.has_key?( name )
|
91
|
+
return @collections[name]
|
92
|
+
else
|
93
|
+
@logger.info "trying to initialize collection #{name} from stored configuration..."
|
94
|
+
return @collections[name] = create_collection( name, nil )
|
95
|
+
end
|
96
|
+
end
|
97
|
+
raise "UnknownCollection #{name}"
|
98
|
+
end
|
99
|
+
|
100
|
+
protected
|
101
|
+
|
102
|
+
# pass through commands to collection
|
103
|
+
def method_missing(method, *args)
|
104
|
+
if args.size >= 1
|
105
|
+
collection_name = args.shift
|
106
|
+
return collection( collection_name ).send( method, *args )
|
107
|
+
end
|
108
|
+
super
|
109
|
+
end
|
110
|
+
|
111
|
+
def create_multi_collection( name, collections, options = {} )
|
112
|
+
Stellr::Collections::MultiCollection.new name, collections, { :logger => @logger }.merge(options)
|
113
|
+
end
|
114
|
+
|
115
|
+
# initializes a new collection object
|
116
|
+
#
|
117
|
+
# if nil is given for options, the method tries to locate a previously
|
118
|
+
# saved collection configuration and restore from it.
|
119
|
+
def create_collection( name, options )
|
120
|
+
raise "invalid collection name >#{name}<, may only contain a-zA-Z0-9_-" unless name =~ /^([a-zA-Z0-9_-]+)$/
|
121
|
+
name.untaint
|
122
|
+
save_config = true
|
123
|
+
if options.nil?
|
124
|
+
options = load_collection_config name
|
125
|
+
save_config = false
|
126
|
+
end
|
127
|
+
raise "No options given for collection #{name} and no stored configuration found." if options.nil?
|
128
|
+
|
129
|
+
options[:path] = File.join( @config.data_dir, name )
|
130
|
+
save_collection_config name, options if save_config
|
131
|
+
return Collections::Base.create( name, {:logger => @logger}.merge(options) )
|
132
|
+
end
|
133
|
+
|
134
|
+
def save_collection_config( name, options )
|
135
|
+
path = collection_config_path name
|
136
|
+
( File.open(path, 'w') << YAML.dump(options) ).close
|
137
|
+
@logger.info "wrote collection config to #{path}"
|
138
|
+
end
|
139
|
+
|
140
|
+
def load_collection_config( name )
|
141
|
+
path = collection_config_path name
|
142
|
+
conf = YAML.load( File.read(path) ) if File.readable?(path) rescue nil
|
143
|
+
@logger.info "loaded collection config from #{path}" unless conf.nil?
|
144
|
+
return conf
|
145
|
+
end
|
146
|
+
|
147
|
+
def collection_config_path( name )
|
148
|
+
File.join @config.conf_dir, "#{name}.yml"
|
149
|
+
end
|
150
|
+
|
151
|
+
# called by shutdown
|
152
|
+
def on_shutdown( mode )
|
153
|
+
@collections.synchronize do
|
154
|
+
@collections.values.each { |coll| coll.shutdown mode }
|
155
|
+
end
|
156
|
+
end
|
157
|
+
|
158
|
+
def create_directories
|
159
|
+
FileUtils.mkdir_p config.log_dir
|
160
|
+
FileUtils.mkdir_p config.tmp_dir
|
161
|
+
FileUtils.mkdir_p config.conf_dir
|
162
|
+
FileUtils.mkdir_p config.data_dir
|
163
|
+
end
|
164
|
+
|
165
|
+
end
|
166
|
+
end
|
@@ -0,0 +1,16 @@
|
|
1
|
+
module Stellr
|
2
|
+
module Strategies
|
3
|
+
class Base
|
4
|
+
include Stellr::Utils::Shutdown
|
5
|
+
|
6
|
+
def initialize( collection, options )
|
7
|
+
@collection = collection
|
8
|
+
@options = options.dup
|
9
|
+
end
|
10
|
+
|
11
|
+
def method_missing(name, *args)
|
12
|
+
@collection.send name, *args
|
13
|
+
end
|
14
|
+
end
|
15
|
+
end
|
16
|
+
end
|
@@ -0,0 +1,78 @@
|
|
1
|
+
module Stellr
|
2
|
+
|
3
|
+
module Strategies
|
4
|
+
|
5
|
+
# Queueing strategy. Any index modifying methods return immediately, actions
|
6
|
+
# are queued and executed asynchronously in order of arrival.
|
7
|
+
#
|
8
|
+
class Queueing < Base
|
9
|
+
|
10
|
+
def initialize( collection, options )
|
11
|
+
super collection, options
|
12
|
+
@options[:max_batch_size] ||= 200
|
13
|
+
@queue = Queue.new
|
14
|
+
@thread = spawn_indexing_thread
|
15
|
+
end
|
16
|
+
|
17
|
+
def add_record( record, boost = nil )
|
18
|
+
enqueue :add, [record, boost]
|
19
|
+
end
|
20
|
+
|
21
|
+
def delete_record( record )
|
22
|
+
enqueue :delete, record
|
23
|
+
end
|
24
|
+
|
25
|
+
protected
|
26
|
+
|
27
|
+
def on_shutdown( mode )
|
28
|
+
@queue << 'shutting down' # letztes queue item damit process_queue nicht haengt
|
29
|
+
@thread.join
|
30
|
+
# save_queue :TODO:
|
31
|
+
@collection.shutdown mode
|
32
|
+
end
|
33
|
+
|
34
|
+
# called by the indexer thread as long as the server runs
|
35
|
+
def process_queue
|
36
|
+
counter = 0
|
37
|
+
max_batch_size = @options[:max_batch_size]
|
38
|
+
begin
|
39
|
+
while record = @queue.deq and not shutting_down?( :abort )
|
40
|
+
process_record( *record )
|
41
|
+
break if ((counter += 1) > max_batch_size) or @queue.empty?
|
42
|
+
end
|
43
|
+
@collection.batch_finished
|
44
|
+
rescue Exception => e
|
45
|
+
puts "OH NO! #{e}\n#{e.backtrace.join "\n"}"
|
46
|
+
end
|
47
|
+
end
|
48
|
+
|
49
|
+
# process a single task from the queue
|
50
|
+
def process_record( action, record )
|
51
|
+
case action
|
52
|
+
when :add
|
53
|
+
@collection.add_record( *record )
|
54
|
+
when :delete
|
55
|
+
@collection.delete_record record
|
56
|
+
else
|
57
|
+
raise "UnknownAction"
|
58
|
+
end
|
59
|
+
end
|
60
|
+
|
61
|
+
# Spawns the thread executing the main loop
|
62
|
+
def spawn_indexing_thread
|
63
|
+
Thread.new do
|
64
|
+
process_queue while !shutting_down?
|
65
|
+
end
|
66
|
+
end
|
67
|
+
|
68
|
+
# add a task to the queue
|
69
|
+
def enqueue( action, record )
|
70
|
+
return false if shutting_down?
|
71
|
+
@queue << [ action, record ]
|
72
|
+
true
|
73
|
+
end
|
74
|
+
|
75
|
+
end
|
76
|
+
|
77
|
+
end
|
78
|
+
end
|