Merge pull request #17 from mazrean/doc/add-next-part-benchmark

add NextPart benchmark
mazrean · Mar 30, 2024 · 5335f06 · 5335f06
2 parents 0b5ca3b + ac143d5
commit 5335f06
Show file tree

Hide file tree

Showing 3 changed files with 95 additions and 6 deletions.
diff --git a/formstream_test.go b/formstream_test.go
@@ -2,6 +2,7 @@ package formstream_test
 
 import (
 	"bytes"
+	"errors"
 	"fmt"
 	"io"
 	"log"
@@ -42,7 +43,7 @@ large file contents
 		}
 
 		return nil
-	})
+	}, formstream.WithRequiredPart("field"))
 	if err != nil {
 		log.Fatal(err)
 	}
@@ -209,7 +210,7 @@ func benchmarkFormStream(b *testing.B, fileSize formstream.DataSize, reverse boo
 
 		parser := formstream.NewParser(boundary)
 
-		err = parser.Register("stream", func(r io.Reader, header formstream.Header) error {
+		err = parser.Register("stream", func(r io.Reader, _ formstream.Header) error {
 			// get field value
 			_, _, _ = parser.Value("field")
 
@@ -231,6 +232,85 @@ func benchmarkFormStream(b *testing.B, fileSize formstream.DataSize, reverse boo
 	}
 }
 
+func BenchmarkStdMultipartNextPart(b *testing.B) {
+	b.Run("1MB", func(b *testing.B) {
+		benchmarkStdMultipartNextPart(b, 1*formstream.MB)
+	})
+	b.Run("10MB", func(b *testing.B) {
+		benchmarkStdMultipartNextPart(b, 10*formstream.MB)
+	})
+	b.Run("100MB", func(b *testing.B) {
+		benchmarkStdMultipartNextPart(b, 100*formstream.MB)
+	})
+	b.Run("1GB", func(b *testing.B) {
+		benchmarkStdMultipartNextPart(b, 1*formstream.GB)
+	})
+	b.Run("5GB", func(b *testing.B) {
+		if testing.Short() {
+			b.Skip("skipping test in short mode.")
+		}
+		benchmarkStdMultipartNextPart(b, 5*formstream.GB)
+	})
+	b.Run("10GB", func(b *testing.B) {
+		if testing.Short() {
+			b.Skip("skipping test in short mode.")
+		}
+		benchmarkStdMultipartNextPart(b, 10*formstream.GB)
+	})
+}
+
+func benchmarkStdMultipartNextPart(b *testing.B, fileSize formstream.DataSize) {
+	r, err := sampleForm(fileSize, boundary, false)
+	if err != nil {
+		b.Fatal(err)
+	}
+	defer r.Close()
+
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		b.StopTimer()
+		_, err := r.Seek(0, io.SeekStart)
+		if err != nil {
+			b.Fatal(err)
+		}
+		b.StartTimer()
+
+		func() {
+			mr := multipart.NewReader(r, boundary)
+
+			for {
+				p, err := mr.NextPart()
+				if errors.Is(err, io.EOF) {
+					break
+				}
+				if err != nil {
+					b.Fatal(err)
+				}
+
+				if p.FormName() == "field" {
+					sb := &strings.Builder{}
+					_, err := io.Copy(sb, p)
+					if err != nil {
+						b.Fatal(err)
+					}
+
+					_ = sb.String()
+				} else {
+					_, err := io.Copy(io.Discard, p)
+					if err != nil {
+						b.Fatal(err)
+					}
+				}
+
+				_, err = io.Copy(io.Discard, p)
+				if err != nil {
+					b.Fatal(err)
+				}
+			}
+		}()
+	}
+}
+
 func BenchmarkStdMultipartReadForm(b *testing.B) {
 	b.Run("1MB", func(b *testing.B) {
 		benchmarkStdMultipartReadForm(b, 1*formstream.MB)

diff --git a/http/parser_test.go b/http/parser_test.go
@@ -65,7 +65,7 @@ func createUserHandler(res http.ResponseWriter, req *http.Request) {
 		return
 	}
 
-	err = parser.Register("icon", func(r io.Reader, header formstream.Header) error {
+	err = parser.Register("icon", func(r io.Reader, _ formstream.Header) error {
 		name, _, _ := parser.Value("name")
 		password, _, _ := parser.Value("password")
 
@@ -317,7 +317,7 @@ func benchmarkFormStream(b *testing.B, fileSize formstream.DataSize, reverse boo
 			return
 		}
 
-		err = parser.Register("stream", func(r io.Reader, header formstream.Header) error {
+		err = parser.Register("stream", func(r io.Reader, _ formstream.Header) error {
 			// get field value
 			_, _, _ = parser.Value("field")
 

diff --git a/scripts/graph.py b/scripts/graph.py
@@ -26,6 +26,13 @@ def parse_file_size(mem_size: str):
         raise Exception("Unknown memory size: " + mem_size)
 
 
+group_colors = {
+    "FormStream(FastPath)": "#ADFF2F",
+    "FormStream(SlowPath)": "#00FF00",
+    "std(with NextPart)": "#6495ED",
+    "std(with ReadForm)": "#D3D3D3",
+}
+
 file_sizes = []
 time_per_ops = {}
 mem_per_ops = {}
@@ -51,15 +58,16 @@ def parse_file_size(mem_size: str):
 # remove duplicates
 file_sizes = sorted(set(file_sizes), key=file_sizes.index)
 
-bar_width = 0.25
+bar_width = 0.2
 
-index = np.arange(len(file_sizes))
+index = np.arange(len(file_sizes)) - 0.1
 fig, ax_time = plt.subplots(figsize=(12, 7))
 for i, (group, group_time_dict) in enumerate(time_per_ops.items()):
     ax_time.bar(
         index + i * bar_width,
         [group_time_dict[fs] / 1e6 for fs in file_sizes],
         bar_width,
+        color=group_colors[group],
         label=group,
     )
 
@@ -79,6 +87,7 @@ def parse_file_size(mem_size: str):
         index + i * bar_width,
         [group_mem_dict[fs] for fs in file_sizes],
         bar_width,
+        color=group_colors[group],
         label=group,
     )