Fixup naming, turn off firstmatch, add benchmark filtering by regex

Author: rctcwyvrn

Sources/RegexBenchmark/Benchmark.swift

@@ -69,11 +69,12 @@ struct CrossBenchmark {
   /// TODO: Probably better ot have a whole-line vs search anywhere, maybe
   /// accomodate multi-line matching, etc.
   var isWhole: Bool = false
+  
+  /// Whether or not to do firstMatch as well or just allMatches
+  var includeFirst: Bool = false
 
   func register(_ runner: inout BenchmarkRunner) {
     let swiftRegex = try! Regex(regex)
-
-    let nsPattern = isWhole ? "^" + regex + "$" : regex
     let nsRegex: NSRegularExpression
     if isWhole {
       nsRegex = try! NSRegularExpression(pattern: "^" + regex + "$")
@@ -95,30 +96,32 @@ struct CrossBenchmark {
           type: .first,
           target: input))
     } else {
-      runner.register(
-        Benchmark(
-          name: baseName + "First",
-          regex: swiftRegex,
-          type: .first,
-          target: input))
       runner.register(
         Benchmark(
           name: baseName + "All",
           regex: swiftRegex,
           type: .allMatches,
           target: input))
-      runner.register(
-        NSBenchmark(
-          name: baseName + "First_NS",
-          regex: nsRegex,
-          type: .first,
-          target: input))
       runner.register(
         NSBenchmark(
           name: baseName + "All_NS",
           regex: nsRegex,
           type: .allMatches,
           target: input))
+      if includeFirst {
+        runner.register(
+          Benchmark(
+            name: baseName + "First",
+            regex: swiftRegex,
+            type: .first,
+            target: input))
+        runner.register(
+          NSBenchmark(
+            name: baseName + "First_NS",
+            regex: nsRegex,
+            type: .first,
+            target: input))
+      }
     }
   }
 }

Sources/RegexBenchmark/CLI.swift

@@ -2,7 +2,7 @@ import ArgumentParser
 
 @main
 struct Runner: ParsableCommand {
-  @Argument(help: "Names of benchmarks to run")
+  @Argument(help: "Patterns for benchmarks to run")
   var specificBenchmarks: [String] = []
 
   @Flag(help: "Run only once for profiling purposes")
@@ -20,19 +20,30 @@ struct Runner: ParsableCommand {
   @Flag(help: "Should the results be saved")
   var save = false
 
-  @Flag(help: "Compare this result with the latest saved result")
+  @Flag(help: "Compare this result with a saved result")
   var compare = false
 
   @Option(help: "The result file to compare against, if this flag is not set it will compare against the most recent result file")
   var compareFile: String?
 
+  @Flag(help: "Exclude the comparisons to NSRegex")
+  var excludeNs = false
+  
   mutating func run() throws {
     var runner = BenchmarkRunner.makeRunner(samples, outputPath)
 
-    // todo: regex based filter 
+    if excludeNs {
+      self.specificBenchmarks.append("((?!NS).)*")
+    }
+    
     if !self.specificBenchmarks.isEmpty {
-      runner.suite = runner.suite.filter { b in specificBenchmarks.contains(b.name) }
+      runner.suite = runner.suite.filter { b in
+        specificBenchmarks.contains { pattern in
+          try! Regex(pattern).wholeMatch(in: b.name) != nil
+        }
+      }
     }
+    
     switch (profile, debug) {
     case (true, true): print("Cannot run both profile and debug")
     case (true, false): runner.profile()

Sources/RegexBenchmark/Suite/CssRegex.swift

@@ -6,7 +6,7 @@ extension BenchmarkRunner {
     let r = #"--([a-zA-Z0-9_-]+)\s*:\s*(.*?);"#
 
     let css = CrossBenchmark(
-      baseName: "css", regex: r, input: Inputs.swiftOrgCSS)
+      baseName: "Css", regex: r, input: Inputs.swiftOrgCSS)
     css.register(&self)
   }
 }

Sources/RegexBenchmark/Suite/CustomCharacterClasses.swift

@@ -13,37 +13,37 @@ extension BenchmarkRunner {
     let input = Inputs.graphemeBreakData
 
     register(Benchmark(
-      name: "basicCCC",
+      name: "BasicCCC",
       regex: try! Regex(basic),
       type: .allMatches,
       target: input))
 
     register(Benchmark(
-      name: "basicRangeCCC",
+      name: "BasicRangeCCC",
       regex: try! Regex(basicRange),
       type: .allMatches,
       target: input))
 
     register(Benchmark(
-      name: "caseInsensitiveCCC",
+      name: "CaseInsensitiveCCC",
       regex: try! Regex(caseInsensitive),
       type: .allMatches,
       target: input))
 
     register(Benchmark(
-      name: "invertedCCC",
+      name: "InvertedCCC",
       regex: try! Regex(inverted),
       type: .allMatches,
       target: input))
 
     register(Benchmark(
-      name: "subtractionCCC",
+      name: "SubtractionCCC",
       regex: try! Regex(subtraction),
       type: .allMatches,
       target: input))
 
     register(Benchmark(
-      name: "intersectionCCC",
+      name: "IntersectionCCC",
       regex: try! Regex(intersection),
       type: .allMatches,
       target: input))

Sources/RegexBenchmark/Suite/EmailRegex.swift

@@ -13,22 +13,22 @@ extension BenchmarkRunner {
     let emailWithLookaheads = #"(?=[A-z0-9][A-z0-9@._%+-]{5,253})[A-z0-9._%+-]{1,64}@(?:(?=[A-z0-9-]{1,63}\.)[A-z0-9]+(?:-[A-z0-9]+)*\.){1,8}[A-z]{2,63}"#
 
     let emailRFCValid = CrossBenchmark(
-      baseName: "emailRFC", regex: emailRFC, input: Inputs.validEmails)
+      baseName: "EmailRFC", regex: emailRFC, input: Inputs.validEmails)
 
     let emailRFCInvalid = CrossBenchmark(
-      baseName: "emailRFCNoMatches",
+      baseName: "EmailRFCNoMatches",
       regex: emailRFC,
       input: Inputs.graphemeBreakData
     )
 
     let emailValid = CrossBenchmark(
-      baseName: "emailLookahead",
+      baseName: "EmailLookahead",
       regex: emailWithLookaheads,
       input: Inputs.validEmails
     )
 
     let emailInvalid = CrossBenchmark(
-      baseName: "emailLookaheadNoMatches",
+      baseName: "EmailLookaheadNoMatches",
       regex: emailWithLookaheads,
       input: Inputs.graphemeBreakData
     )

Sources/RegexBenchmark/Suite/GraphemeBreak.swift

@@ -18,7 +18,7 @@ extension BenchmarkRunner {
     let regex = #"HANGUL SYLLABLE [A-Z]+(?:\.\.HANGUL SYLLABLE [A-Z]+)?"#
 
     let benchmark = CrossBenchmark(
-      baseName: "HangulSyllable", regex: regex, input: input)
+      baseName: "HangulSyllable", regex: regex, input: input, includeFirst: true)
     benchmark.register(&self)
   }
 }

Sources/RegexBenchmark/Suite/HtmlRegex.swift

@@ -6,7 +6,7 @@ extension BenchmarkRunner {
     let r = #"<(\w*)\b[^>]*>(.*?)<\/\1>"#
 
     let html = CrossBenchmark(
-      baseName: "html", regex: r, input: Inputs.swiftOrgHTML)
+      baseName: "Html", regex: r, input: Inputs.swiftOrgHTML)
     html.register(&self)
   }
 }

Sources/RegexBenchmark/Suite/NotFound.swift

@@ -6,11 +6,11 @@ extension BenchmarkRunner {
     let input = String(repeating: " ", count: 100_000)
 
     let notFound = CrossBenchmark(
-      baseName: "notFound", regex: "a", input: input)
+      baseName: "NotFound", regex: "a", input: input)
     notFound.register(&self)
 
     let anchoredNotFound = CrossBenchmark(
-      baseName: "notFound", regex: "^ +a", input: input)
+      baseName: "AnchoredNotFound", regex: "^ +a", input: input)
     anchoredNotFound.register(&self)
   }
 }