Merge pull request #2162 from github/instrumentation

Add instrumentation to detection and classification

lib/linguist.rb

@@ -6,3 +6,15 @@
 require 'linguist/samples'
 require 'linguist/shebang'
 require 'linguist/version'
+
+class << Linguist
+  attr_accessor :instrumenter
+
+  def instrument(*args, &bk)
+    if instrumenter
+      instrumenter.instrument(*args, &bk)
+    else
+      yield if block_given?
+    end
+  end
+end

lib/linguist/language.rb

@@ -105,19 +105,31 @@ def self.detect(blob)
       # Bail early if the blob is binary or empty.
       return nil if blob.likely_binary? || blob.binary? || blob.empty?
 
-      # Call each strategy until one candidate is returned.
-      STRATEGIES.reduce([]) do |languages, strategy|
-        candidates = strategy.call(blob, languages)
-        if candidates.size == 1
-          return candidates.first
-        elsif candidates.size > 1
-          # More than one candidate was found, pass them to the next strategy.
-          candidates
-        else
-          # No candiates were found, pass on languages from the previous strategy.
-          languages
+      Linguist.instrument("linguist.detection", :blob => blob) do
+        # Call each strategy until one candidate is returned.
+        languages = []
+        returning_strategy = nil
+
+        STRATEGIES.each do |strategy|
+          returning_strategy = strategy
+          candidates = Linguist.instrument("linguist.strategy", :blob => blob, :strategy => strategy, :candidates => languages) do
+            strategy.call(blob, languages)
+          end
+          if candidates.size == 1
+            languages = candidates
+            break
+          elsif candidates.size > 1
+            # More than one candidate was found, pass them to the next strategy.
+            languages = candidates
+          else
+            # No candidates, try the next strategy
+          end
         end
-      end.first
+
+        Linguist.instrument("linguist.detected", :blob => blob, :strategy => returning_strategy, :language => languages.first)
+
+        languages.first
+      end
     end
 
     # Public: Get all Languages

test/test_instrumentation.rb

@@ -0,0 +1,50 @@
+require_relative "./helper"
+
+class TestInstrumentation < Minitest::Test
+  include Linguist
+
+  class LocalInstrumenter
+    Event = Struct.new(:name, :args)
+
+    attr_reader :events
+
+    def initialize
+      @events = []
+    end
+
+    def instrument(name, *args)
+      @events << Event.new(name, args)
+      yield if block_given?
+    end
+  end
+
+  def setup
+    Linguist.instrumenter = LocalInstrumenter.new
+  end
+
+  def teardown
+    Linguist.instrumenter = nil
+  end
+
+  def test_detection_instrumentation_with_binary_blob
+    binary_blob = fixture_blob("Binary/octocat.ai")
+    Language.detect(binary_blob)
+
+    # Shouldn't instrument this (as it's binary)
+    assert_equal 0, Linguist.instrumenter.events.size
+  end
+
+  def test_modeline_instrumentation
+    blob = fixture_blob("Data/Modelines/ruby")
+    Language.detect(blob)
+
+    detect_event = Linguist.instrumenter.events.last
+    detect_event_payload = detect_event[:args].first
+
+    assert_equal 3, Linguist.instrumenter.events.size
+    assert_equal "linguist.detected", detect_event.name
+    assert_equal Language['Ruby'], detect_event_payload[:language]
+    assert_equal blob, detect_event_payload[:blob]
+    assert_equal Linguist::Strategy::Modeline, detect_event_payload[:strategy]
+  end
+end