Filter out indexed entries before fuzzy searching

We're doing extra work fuzzy matching across the whole index, even though only a subset of entries are considered valid results. We can filter out dependency/private entries *before* fuzzy matching, which should speed things up, assuming that fuzzy matching is more expensive than filtering.

Author: janko

lib/ruby_indexer/lib/ruby_indexer/index.rb

@@ -196,11 +196,12 @@ def prefix_search(query, nesting = nil)
 
     # Fuzzy searches index entries based on Jaro-Winkler similarity. If no query is provided, all entries are returned
     #: (String? query) -> Array[Entry]
-    def fuzzy_search(query)
+    def fuzzy_search(query, &condition)
       unless query
         entries = @entries.filter_map do |_name, entries|
           next if entries.first.is_a?(Entry::SingletonClass)
 
+          entries = entries.select(&condition) if condition
           entries
         end
 
@@ -212,6 +213,9 @@ def fuzzy_search(query)
       results = @entries.filter_map do |name, entries|
         next if entries.first.is_a?(Entry::SingletonClass)
 
+        entries = entries.select(&condition) if condition
+        next if entries.empty?
+
         similarity = DidYouMean::JaroWinkler.distance(name.gsub("::", "").downcase, normalized_query)
         [entries, -similarity] if similarity > ENTRY_SIMILARITY_THRESHOLD
       end

lib/ruby_lsp/requests/workspace_symbol.rb

@@ -20,17 +20,7 @@ def initialize(global_state, query)
       # @override
       #: -> Array[Interface::WorkspaceSymbol]
       def perform
-        @index.fuzzy_search(@query).filter_map do |entry|
-          uri = entry.uri
-          file_path = uri.full_path
-
-          # We only show symbols declared in the workspace
-          in_dependencies = file_path && !not_in_dependencies?(file_path)
-          next if in_dependencies
-
-          # We should never show private symbols when searching the entire workspace
-          next if entry.private?
-
+        fuzzy_search.filter_map do |entry|
           kind = kind_for_entry(entry)
           loc = entry.location
 
@@ -44,7 +34,7 @@ def perform
             container_name: container.join("::"),
             kind: kind,
             location: Interface::Location.new(
-              uri: uri.to_s,
+              uri: entry.uri.to_s,
               range:  Interface::Range.new(
                 start: Interface::Position.new(line: loc.start_line - 1, character: loc.start_column),
                 end: Interface::Position.new(line: loc.end_line - 1, character: loc.end_column),
@@ -53,6 +43,23 @@ def perform
           )
         end
       end
+
+      private
+
+      def fuzzy_search
+        @index.fuzzy_search(@query) do |entry|
+          file_path = entry.uri.full_path
+
+          # We only show symbols declared in the workspace
+          in_dependencies = file_path && !not_in_dependencies?(file_path)
+          next if in_dependencies
+
+          # We should never show private symbols when searching the entire workspace
+          next if entry.private?
+
+          true
+        end
+      end
     end
   end
 end