Add a new filter: regex_extract

CHANGES.md

@@ -2,6 +2,7 @@
 
 | DateOfChange   | Changes                                                                                                      |
 |----------------|--------------------------------------------------------------------------------------------------------------|
+| Feb 19, 2023   | Add a new Liquid filter named `regex_extract`. [3220](https://github.com/huginn/huginn/pull/3220) |
 | Feb 18, 2023   | Fix permissions on /app in the Docker images. [3218](https://github.com/huginn/huginn/pull/3218) |
 | Feb 06, 2023   | BoxcarAgent is removed; the service is long gone. |
 | Dec 15, 2022   | DataOutputAgent Add dc namespace option. [3189](https://github.com/huginn/huginn/pull/3189) |

app/concerns/liquid_interpolatable.rb

@@ -231,6 +231,12 @@ def to_xpath(string)
       end
     end
 
+    def regex_extract(input, regex, index = 0)
+      input.to_s[Regexp.new(regex), index]
+    rescue Index
+      nil
+    end
+
     def regex_replace(input, regex, replacement = nil)
       input.to_s.gsub(Regexp.new(regex), unescape_replacement(replacement.to_s))
     end

spec/concerns/liquid_interpolatable_spec.rb

@@ -228,6 +228,22 @@ def @filter.to_xpath_roundtrip(string)
     end
   end
 
+  describe 'regex_extract' do
+    let(:agent) { Agents::InterpolatableAgent.new(name: "test") }
+
+    it 'should extract the matched part' do
+      agent.interpolation_context['something'] = "foo BAR BAZ"
+      agent.options['test'] = "{{ something | regex_extract: '[A-Z]+' }} / {{ something | regex_extract: '[A-Z]([A-Z]+)', 1 }} / {{ something | regex_extract: '(?<x>.)AZ', 'x' }}"
+      expect(agent.interpolated['test']).to eq("BAR / AR / B")
+    end
+
+    it 'should return nil if not matched' do
+      agent.interpolation_context['something'] = "foo BAR BAZ"
+      agent.options['test'] = "{% assign var = something | regex_extract: '[A-Z][a-z]+' %}{% if var == nil %}nil{% else %}non-nil{% endif %}"
+      expect(agent.interpolated['test']).to eq("nil")
+    end
+  end
+
   describe 'regex_replace' do
     let(:agent) { Agents::InterpolatableAgent.new(name: "test") }