greglu-solr-ruby 0.0.7

Sign up to get free protection for your applications and to get access to all the features.
Files changed (103) hide show
  1. data/CHANGES.yml +50 -0
  2. data/LICENSE.txt +201 -0
  3. data/README +56 -0
  4. data/Rakefile +190 -0
  5. data/examples/delicious_library/dl_importer.rb +60 -0
  6. data/examples/delicious_library/sample_export.txt +164 -0
  7. data/examples/marc/marc_importer.rb +106 -0
  8. data/examples/tang/tang_importer.rb +58 -0
  9. data/lib/solr.rb +21 -0
  10. data/lib/solr/connection.rb +179 -0
  11. data/lib/solr/document.rb +73 -0
  12. data/lib/solr/exception.rb +13 -0
  13. data/lib/solr/field.rb +39 -0
  14. data/lib/solr/importer.rb +19 -0
  15. data/lib/solr/importer/array_mapper.rb +26 -0
  16. data/lib/solr/importer/delimited_file_source.rb +38 -0
  17. data/lib/solr/importer/hpricot_mapper.rb +27 -0
  18. data/lib/solr/importer/mapper.rb +51 -0
  19. data/lib/solr/importer/solr_source.rb +43 -0
  20. data/lib/solr/importer/xpath_mapper.rb +35 -0
  21. data/lib/solr/indexer.rb +52 -0
  22. data/lib/solr/request.rb +26 -0
  23. data/lib/solr/request/add_document.rb +63 -0
  24. data/lib/solr/request/base.rb +36 -0
  25. data/lib/solr/request/commit.rb +31 -0
  26. data/lib/solr/request/delete.rb +50 -0
  27. data/lib/solr/request/dismax.rb +46 -0
  28. data/lib/solr/request/index_info.rb +22 -0
  29. data/lib/solr/request/modify_document.rb +51 -0
  30. data/lib/solr/request/optimize.rb +21 -0
  31. data/lib/solr/request/ping.rb +36 -0
  32. data/lib/solr/request/select.rb +56 -0
  33. data/lib/solr/request/spellcheck.rb +30 -0
  34. data/lib/solr/request/standard.rb +374 -0
  35. data/lib/solr/request/update.rb +23 -0
  36. data/lib/solr/response.rb +27 -0
  37. data/lib/solr/response/add_document.rb +17 -0
  38. data/lib/solr/response/base.rb +42 -0
  39. data/lib/solr/response/commit.rb +17 -0
  40. data/lib/solr/response/delete.rb +13 -0
  41. data/lib/solr/response/dismax.rb +20 -0
  42. data/lib/solr/response/index_info.rb +26 -0
  43. data/lib/solr/response/modify_document.rb +17 -0
  44. data/lib/solr/response/optimize.rb +14 -0
  45. data/lib/solr/response/ping.rb +28 -0
  46. data/lib/solr/response/ruby.rb +42 -0
  47. data/lib/solr/response/select.rb +17 -0
  48. data/lib/solr/response/spellcheck.rb +20 -0
  49. data/lib/solr/response/standard.rb +60 -0
  50. data/lib/solr/response/xml.rb +42 -0
  51. data/lib/solr/solrtasks.rb +27 -0
  52. data/lib/solr/util.rb +32 -0
  53. data/lib/solr/xml.rb +47 -0
  54. data/script/setup.rb +14 -0
  55. data/script/solrshell +18 -0
  56. data/solr-ruby.gemspec +26 -0
  57. data/solr/conf/admin-extra.html +31 -0
  58. data/solr/conf/protwords.txt +21 -0
  59. data/solr/conf/schema.xml +221 -0
  60. data/solr/conf/scripts.conf +24 -0
  61. data/solr/conf/solrconfig.xml +394 -0
  62. data/solr/conf/stopwords.txt +58 -0
  63. data/solr/conf/synonyms.txt +31 -0
  64. data/solr/conf/xslt/example.xsl +132 -0
  65. data/test/conf/admin-extra.html +31 -0
  66. data/test/conf/protwords.txt +21 -0
  67. data/test/conf/schema.xml +237 -0
  68. data/test/conf/scripts.conf +24 -0
  69. data/test/conf/solrconfig.xml +376 -0
  70. data/test/conf/stopwords.txt +58 -0
  71. data/test/conf/synonyms.txt +31 -0
  72. data/test/functional/server_test.rb +218 -0
  73. data/test/functional/test_solr_server.rb +104 -0
  74. data/test/unit/add_document_test.rb +40 -0
  75. data/test/unit/array_mapper_test.rb +37 -0
  76. data/test/unit/changes_yaml_test.rb +21 -0
  77. data/test/unit/commit_test.rb +41 -0
  78. data/test/unit/connection_test.rb +55 -0
  79. data/test/unit/data_mapper_test.rb +75 -0
  80. data/test/unit/delete_test.rb +56 -0
  81. data/test/unit/delimited_file_source_test.rb +29 -0
  82. data/test/unit/dismax_request_test.rb +26 -0
  83. data/test/unit/document_test.rb +69 -0
  84. data/test/unit/field_test.rb +48 -0
  85. data/test/unit/hpricot_mapper_test.rb +44 -0
  86. data/test/unit/hpricot_test_file.xml +26 -0
  87. data/test/unit/indexer_test.rb +57 -0
  88. data/test/unit/modify_document_test.rb +24 -0
  89. data/test/unit/ping_test.rb +51 -0
  90. data/test/unit/request_test.rb +61 -0
  91. data/test/unit/response_test.rb +43 -0
  92. data/test/unit/select_test.rb +25 -0
  93. data/test/unit/solr_mock_base.rb +40 -0
  94. data/test/unit/spellcheck_response_test.rb +26 -0
  95. data/test/unit/spellchecker_request_test.rb +27 -0
  96. data/test/unit/standard_request_test.rb +324 -0
  97. data/test/unit/standard_response_test.rb +174 -0
  98. data/test/unit/suite.rb +16 -0
  99. data/test/unit/tab_delimited.txt +2 -0
  100. data/test/unit/util_test.rb +24 -0
  101. data/test/unit/xpath_mapper_test.rb +38 -0
  102. data/test/unit/xpath_test_file.xml +25 -0
  103. metadata +173 -0
@@ -0,0 +1,17 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ require 'rexml/xpath'
14
+
15
+ class Solr::Response::Commit < Solr::Response::Xml
16
+ end
17
+
@@ -0,0 +1,13 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Delete < Solr::Response::Xml; end
@@ -0,0 +1,20 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Dismax < Solr::Response::Standard
14
+ # no need for special processing
15
+
16
+ # FIXME: 2007-02-07 <coda.hale@gmail.com> -- The existence of this class indicates that
17
+ # the Request/Response pair architecture is a little hinky. Perhaps we could refactor
18
+ # out some of the most common functionality -- Common Query Parameters, Highlighting Parameters,
19
+ # Simple Facet Parameters, etc. -- into modules?
20
+ end
@@ -0,0 +1,26 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::IndexInfo < Solr::Response::Ruby
14
+ def initialize(ruby_code)
15
+ super
16
+ end
17
+
18
+ def num_docs
19
+ return @data['index']['numDocs']
20
+ end
21
+
22
+ def field_names
23
+ return @data['fields'].keys
24
+ end
25
+
26
+ end
@@ -0,0 +1,17 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::ModifyDocument < Solr::Response::Xml
14
+ def initialize(xml)
15
+ super
16
+ end
17
+ end
@@ -0,0 +1,14 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Optimize < Solr::Response::Commit
14
+ end
@@ -0,0 +1,28 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ require 'rexml/xpath'
14
+
15
+ class Solr::Response::Ping < Solr::Response::Xml
16
+
17
+ def initialize(xml)
18
+ super
19
+ @ok = REXML::XPath.first(@doc, './solr/ping') ? true : false
20
+ end
21
+
22
+ # returns true or false depending on whether the ping
23
+ # was successful or not
24
+ def ok?
25
+ @ok
26
+ end
27
+
28
+ end
@@ -0,0 +1,42 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Ruby < Solr::Response::Base
14
+ attr_reader :data, :header
15
+
16
+ def initialize(ruby_code)
17
+ super
18
+ begin
19
+ #TODO: what about pulling up data/header/response to ResponseBase,
20
+ # or maybe a new middle class like SelectResponseBase since
21
+ # all Select queries return this same sort of stuff??
22
+ # XML (&wt=xml) and Ruby (&wt=ruby) responses contain exactly the same structure.
23
+ # a goal of solrb is to make it irrelevant which gets used under the hood,
24
+ # but favor Ruby responses.
25
+ @data = eval(ruby_code)
26
+ @header = @data['responseHeader']
27
+ raise "response should be a hash" unless @data.kind_of? Hash
28
+ raise "response header missing" unless @header.kind_of? Hash
29
+ rescue SyntaxError => e
30
+ raise Solr::Exception.new("invalid ruby code: #{e}")
31
+ end
32
+ end
33
+
34
+ def ok?
35
+ @header['status'] == 0
36
+ end
37
+
38
+ def query_time
39
+ @header['QTime']
40
+ end
41
+
42
+ end
@@ -0,0 +1,17 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Select < Solr::Response::Ruby
14
+ def initialize(ruby_code)
15
+ super
16
+ end
17
+ end
@@ -0,0 +1,20 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Spellcheck < Solr::Response::Ruby
14
+ attr_reader :suggestions
15
+
16
+ def initialize(ruby_code)
17
+ super
18
+ @suggestions = @data['suggestions']
19
+ end
20
+ end
@@ -0,0 +1,60 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Response::Standard < Solr::Response::Ruby
14
+ FacetValue = Struct.new(:name, :value)
15
+ include Enumerable
16
+
17
+ def initialize(ruby_code)
18
+ super
19
+ @response = @data['response']
20
+ raise "response section missing" unless @response.kind_of? Hash
21
+ end
22
+
23
+ def total_hits
24
+ @response['numFound']
25
+ end
26
+
27
+ def start
28
+ @response['start']
29
+ end
30
+
31
+ def hits
32
+ @response['docs']
33
+ end
34
+
35
+ def max_score
36
+ @response['maxScore']
37
+ end
38
+
39
+ # TODO: consider the use of json.nl parameter
40
+ def field_facets(field)
41
+ facets = []
42
+ values = @data['facet_counts']['facet_fields'][field]
43
+ Solr::Util.paired_array_each(values) do |key, value|
44
+ facets << FacetValue.new(key, value)
45
+ end
46
+
47
+ facets
48
+ end
49
+
50
+ def highlighted(id, field)
51
+ @data['highlighting'][id.to_s][field.to_s] rescue nil
52
+ end
53
+
54
+ # supports enumeration of hits
55
+ # TODO revisit - should this iterate through *all* hits by re-requesting more?
56
+ def each
57
+ @response['docs'].each {|hit| yield hit}
58
+ end
59
+
60
+ end
@@ -0,0 +1,42 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ require 'rexml/document'
14
+ require 'solr/exception'
15
+
16
+ class Solr::Response::Xml < Solr::Response::Base
17
+ attr_reader :doc, :status_code, :status_message
18
+
19
+ def initialize(xml)
20
+ super
21
+ # parse the xml
22
+ @doc = REXML::Document.new(xml)
23
+
24
+ # look for the result code and string
25
+ # <?xml version="1.0" encoding="UTF-8"?>
26
+ # <response>
27
+ # <lst name="responseHeader"><int name="status">0</int><int name="QTime">2</int></lst>
28
+ # </response>
29
+ result = REXML::XPath.first(@doc, './response/lst[@name="responseHeader"]/int[@name="status"]')
30
+ if result
31
+ @status_code = result.text
32
+ @status_message = result.text # TODO: any need for a message?
33
+ end
34
+ rescue REXML::ParseException => e
35
+ raise Solr::Exception.new("invalid response xml: #{e}")
36
+ end
37
+
38
+ def ok?
39
+ return @status_code == '0'
40
+ end
41
+
42
+ end
@@ -0,0 +1,27 @@
1
+ #!/usr/bin/env ruby
2
+ # The ASF licenses this file to You under the Apache License, Version 2.0
3
+ # (the "License"); you may not use this file except in compliance with
4
+ # the License. You may obtain a copy of the License at
5
+ #
6
+ # http://www.apache.org/licenses/LICENSE-2.0
7
+ #
8
+ # Unless required by applicable law or agreed to in writing, software
9
+ # distributed under the License is distributed on an "AS IS" BASIS,
10
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
11
+ # See the License for the specific language governing permissions and
12
+ # limitations under the License.
13
+
14
+ # TODO: fill out Solr tasks: start, stop, ping, optimize, etc.
15
+
16
+ require 'rake'
17
+ require 'rake/tasklib'
18
+
19
+ module Solr
20
+ namespace :solr do
21
+ desc "Start Solr"
22
+ task :start do
23
+ # TODO: actually start it up!
24
+ puts "Starting..."
25
+ end
26
+ end
27
+ end
data/lib/solr/util.rb ADDED
@@ -0,0 +1,32 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ class Solr::Util
14
+ # paired_array_each([key1,value1,key2,value2]) yields twice:
15
+ # |key1,value1| and |key2,value2|
16
+ def self.paired_array_each(a, &block)
17
+ 0.upto(a.size / 2 - 1) do |i|
18
+ n = i * 2
19
+ yield(a[n], a[n+1])
20
+ end
21
+ end
22
+
23
+ # paired_array_to_hash([key1,value1,key2,value2]) => {key1 => value1, key2, value2}
24
+ def self.paired_array_to_hash(a)
25
+ Hash[*a]
26
+ end
27
+
28
+ def self.query_parser_escape(string)
29
+ # backslash prefix everything that isn't a word character
30
+ string.gsub(/(\W)/,'\\\\\1')
31
+ end
32
+ end
data/lib/solr/xml.rb ADDED
@@ -0,0 +1,47 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ module Solr::XML
14
+ end
15
+
16
+ begin
17
+
18
+ # If we can load rubygems and libxml-ruby...
19
+ require 'rubygems'
20
+ require 'xml/libxml'
21
+
22
+ # then make a few modifications to XML::Node so it can stand in for REXML::Element
23
+ class XML::Node
24
+ # element.add_element(another_element) should work
25
+ alias_method :add_element, :<<
26
+
27
+ # element.attributes['blah'] should work
28
+ def attributes
29
+ self
30
+ end
31
+
32
+ # element.text = "blah" should work
33
+ def text=(x)
34
+ self << x.to_s
35
+ end
36
+ end
37
+
38
+ # And use XML::Node for our XML generation
39
+ Solr::XML::Element = XML::Node
40
+
41
+ rescue LoadError => e # If we can't load either rubygems or libxml-ruby
42
+
43
+ # Just use REXML.
44
+ require 'rexml/document'
45
+ Solr::XML::Element = REXML::Element
46
+
47
+ end