documentrix 0.3.2 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 332c67275b90bcc797cdd8e8df75a751bcfa7a52c759c9bcbe4a55970e905401
-  data.tar.gz: c546f89c61613b11d18c2cd724cadae8c303a988f925ad1624c5121f511e88fd
+  metadata.gz: e32a72c0a1f93a96f7c3cecd185f13a0f17a2629b2dc509ec3a29fd2d7b51a41
+  data.tar.gz: 3f2c21125adf7061dcba94f1843456064245222839de0df6ddbcd8743a1aea13
 SHA512:
-  metadata.gz: c80a4c253b4fe367fae1dbb7e6aea50e274a495a0026a7b676e1e962b80c33806fe89976338c12931f11a5059e384fdfddcfe4665f68bde406b118edd8c04d15
-  data.tar.gz: '09292fd5d6aed6c939c0244467c743d515e68e5e253d7f443ef96a9c295a61689f1be5b9aec53e402287b6b6c7d305757eb26f79f5ade0161f7e82d534a7b3bf'
+  metadata.gz: 91dbf3ddfdeb124661ff78a4cbc1635dbb8bdf6606078aeeef77b4f9d14688d073261ad08b59f0333895c7391d83eff8a6c76467f19af4f04ca55172c1d934a5
+  data.tar.gz: d40e5a53ceeda71c7a2be37d3bfb718ffc2ddb0fddc0c6eb2346c295d06713c895493854fab5f768af075bec2595e2f23c40e893d7f7266c6f61cd95c534d733

data/CHANGES.md

@@ -1,5 +1,23 @@
 # Changes
 
+## 2026-05-22 v0.4.0
+
+### Added
+
+- Added introspection methods to `Documents`:
+    - Implemented `sources` to return an array of unique source identifiers.
+    - Implemented `each_record` to iterate over records, returning an
+      `Enumerator` if no block is provided.
+
+### Changed
+
+- Updated CI images from Alpine to Debian:
+    - Switched Ruby **4.0**, **3.4**, **3.3**, and **3.2** images to the
+      `trixie` distribution.
+    - Switched Ruby **3.1** image to the `bookworm` distribution.
+    - Updated the `dockerfile` in `.all_images.yml` to use `apt-get` instead of
+      `apk` for installing build dependencies.
+
 ## 2026-05-20 v0.3.2
 
 ### Performance Improvements

data/documentrix.gemspec

@@ -1,9 +1,9 @@
 # -*- encoding: utf-8 -*-
-# stub: documentrix 0.3.2 ruby lib
+# stub: documentrix 0.4.0 ruby lib
 
 Gem::Specification.new do |s|
   s.name = "documentrix".freeze
-  s.version = "0.3.2".freeze
+  s.version = "0.4.0".freeze
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0".freeze) if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib".freeze]

data/lib/documentrix/documents.rb

@@ -409,6 +409,22 @@ class Documentrix::Documents
     @cache.tags
   end
 
+  # Returns an array of all unique sources stored in the cache.
+  #
+  # @return [Array<String>] An array of unique source identifiers.
+  def sources
+    @cache.each_source.to_a
+  end
+
+ # The each_record method iterates over all records stored in the cache.
+ #
+ # @yield [record] The record being iterated over.
+ # @return [ Enumerator ] an enumerator if no block is provided.
+  def each_record(&block)
+    block or return enum_for(__method__)
+    @cache.each { |_key, record| block.(record) }
+  end
+
   private
 
   # Resets the memoized list of collections.

data/lib/documentrix/version.rb

@@ -1,6 +1,6 @@
 module Documentrix
   # Documentrix version
-  VERSION         = '0.3.2'
+  VERSION         = '0.4.0'
   VERSION_ARRAY   = VERSION.split('.').map(&:to_i) # :nodoc:
   VERSION_MAJOR   = VERSION_ARRAY[0] # :nodoc:
   VERSION_MINOR   = VERSION_ARRAY[1] # :nodoc:

data/spec/documents_spec.rb

@@ -226,6 +226,32 @@ describe Documentrix::Documents do
         from(:default).
         to(:new_collection)
     end
+
+    it 'returns unique sources' do
+      allow(ollama).to receive(:embed).and_return(double(embeddings: [ [ 0.1 ] ]))
+      documents.add('foo', source: 's1')
+      documents.add('bar', source: 's1')
+      documents.add('baz', source: 's2')
+      expect(documents.sources).to match_array %w[ s1 s2 ]
+    end
+
+    it 'can iterate over records' do
+      allow(ollama).to receive(:embed).and_return(double(embeddings: [ [ 0.1 ] ]))
+      documents.add('foo')
+      documents.add('bar')
+      records = []
+      documents.each_record { |r| records << r }
+      expect(records.size).to eq 2
+      expect(records.map(&:text)).to match_array %w[ foo bar ]
+    end
+
+    it 'returns an enumerator for each_record' do
+      allow(ollama).to receive(:embed).and_return(double(embeddings: [ [ 0.1 ] ]))
+      documents.add('foo')
+      documents.add('bar')
+      expect(documents.each_record).to be_a Enumerator
+      expect(documents.each_record.map(&:text)).to match_array %w[ foo bar ]
+    end
   end
 
   context 'source management' do
@@ -233,7 +259,7 @@ describe Documentrix::Documents do
       allow(documents).to receive(:compute_file_digest).and_return('d1')
       allow(documents.cache).to receive(:compute_file_digest).and_return('d1')
 
-      allow(ollama).to receive(:embed).and_return(double(embeddings: [[0.1]]))
+      allow(ollama).to receive(:embed).and_return(double(embeddings: [ [ 0.1 ] ]))
     end
 
     it 'can check if a source exists' do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: documentrix
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.4.0
 platform: ruby
 authors:
 - Florian Frank