code style for mnist

Signed-off-by: siwuxie <[email protected]>

extensions/functions/mnist/install.sh

@@ -1,40 +1,39 @@
-#!/bin/sh
-#
+!/bin/sh
+
 # Copyright 2021-2024 EMQ Technologies Co., Ltd.
-#
+
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
 # You may obtain a copy of the License at
-#
+
 #     http://www.apache.org/licenses/LICENSE-2.0
-#
+
 # Unless required by applicable law or agreed to in writing, software
 # distributed under the License is distributed on an "AS IS" BASIS,
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-#
 
-#dir=/usr/local/onnx
-#cur=$(dirname "$0")
-#echo "Base path $cur"
-#if [ -d "$dir" ]; then
-#    echo "SDK path $dir exists."
-#else
-#    echo "Creating SDK path $dir"
-#    mkdir -p $dir
-#    echo "Created SDK path $dir"
-#    echo "Moving libs"
-#    cp -R $cur/lib $dir
-#    echo "Moved libs"
-#fi
-#
-#if [ -f "/etc/ld.so.conf.d/onnx.conf" ]; then
-#    echo "/etc/ld.so.conf.d/onnx.conf exists"
-#else
-#    echo "Copy conf file"
-#    cp "$cur"/onnx.conf /etc/ld.so.conf.d/
-#    echo "Copied conf file"
-#fi
-#ldconfig
-echo "Done mnist12321321333333333333333333333333333333333333333333333332132131231232131241242314231423143214123"
+
+dir=/usr/local/onnx
+cur=$(dirname "$0")
+echo "Base path $cur"
+if [ -d "$dir" ]; then
+   echo "SDK path $dir exists."
+else
+   echo "Creating SDK path $dir"
+   mkdir -p $dir
+   echo "Created SDK path $dir"
+   echo "Moving libs"
+   cp -R $cur/lib $dir
+   echo "Moved libs"
+fi
+
+if [ -f "/etc/ld.so.conf.d/onnx.conf" ]; then
+   echo "/etc/ld.so.conf.d/onnx.conf exists"
+else
+   echo "Copy conf file"
+   cp "$cur"/onnx.conf /etc/ld.so.conf.d/
+   echo "Copied conf file"
+fi
+ldconfig

extensions/functions/mnist/lib/Readme.md

@@ -1,4 +1,8 @@
 
 
 
-TODO：add readme for onnx
+# ONNX
+
+This function only need onnx-runtime lib, and the lib is support by [ort](github.com/yalue/onnxruntime_go) ,so you can use it directly rather than install C API like Tensorflow Lite.
+
+Support onnxruntime.so is enough.

extensions/functions/mnist/mnist.go

@@ -20,12 +20,11 @@ import (
 	"github.com/lf-edge/ekuiper/contract/v2/api"
 	ort "github.com/yalue/onnxruntime_go"
 	"image"
-	"image/color"
+
 	_ "image/gif"
 	_ "image/jpeg"
 	_ "image/png"
-	"os"
-	"os/exec"
+
 	"sync"
 )
 
@@ -78,7 +77,7 @@ func (f *MnistFunc) Exec(_ api.FunctionContext, args []any) (any, bool) {
 	bounds := originalPic.Bounds().Canon()
 	if (bounds.Min.X != 0) || (bounds.Min.Y != 0) {
 		// Should never happen with the standard library.
-		return fmt.Errorf("Bounding rect  doesn't start at 0, 0"), false
+		return fmt.Errorf("bounding rect doesn't start at 0, 0"), false
 	}
 	inputImage := &ProcessedImage{
 		dx:     float32(bounds.Dx()) / 28.0,
@@ -103,9 +102,6 @@ func (f *MnistFunc) Exec(_ api.FunctionContext, args []any) (any, bool) {
 
 	// The input and output names are required by this network; they can be
 	// found on the MNIST ONNX models page linked in the README.
-	// session, e := ort.NewAdvancedSession(f.modelPath,
-	// 	[]string{"Input3"}, []string{"Plus214_Output_0"},
-	// 	[]ort.ArbitraryTensor{input}, []ort.ArbitraryTensor{output}, nil)
 
 	session, e := ort.NewDynamicAdvancedSession(f.modelPath,
 		[]string{"Input3"}, []string{"Plus214_Output_0"}, nil)
@@ -142,137 +138,8 @@ func (f *MnistFunc) IsAggregate() bool {
 
 var Mnist = MnistFunc{
 	modelPath:         "./data/functions/mnist/mnist.onnx",
-	sharedLibraryPath: "./data/functions/mnist/onnxruntime.so",
+	sharedLibraryPath: "/usr/local/onnx/lib/onnxruntime.so",
 	inputShape:        ort.NewShape(1, 1, 28, 28),
 	outputShape:       ort.NewShape(1, 10),
 }
 var _ api.Function = &MnistFunc{}
-
-func printCurrDIr() string {
-	// 创建一个 bytes.Buffer 来捕获命令输出
-	var out bytes.Buffer
-
-	// 创建并配置 exec.Command 用于运行 tree 命令
-	cmd := exec.Command("tree")
-
-	// 设置命令的标准输出为 bytes.Buffer
-	cmd.Stdout = &out
-
-	// 运行命令并检查错误
-	err := cmd.Run()
-	if err != nil {
-		return fmt.Sprintf("Error executing command:%v", err)
-
-	}
-
-	// 将命令输出转换为字符串
-	res := out.String()
-
-	// 打印结果
-	fmt.Println(res)
-	return res
-}
-
-func checkFileStat(filePath string) {
-	// 确认文件路径
-
-	fmt.Println("checkFileStat File path:", filePath)
-
-	// 检查文件是否存在
-	if _, err := os.Stat(filePath); os.IsNotExist(err) {
-		fmt.Println("File does not exist:", filePath)
-	} else {
-		fmt.Println("File exists:", filePath)
-	}
-}
-
-/// 辅助图片类
-
-// Implements the color interface
-type grayscaleFloat float32
-
-func (f grayscaleFloat) RGBA() (r, g, b, a uint32) {
-	a = 0xffff
-	v := uint32(f * 0xffff)
-	if v > 0xffff {
-		v = 0xffff
-	}
-	r = v
-	g = v
-	b = v
-	return
-}
-
-// ProcessedImage Used to satisfy the image interface as well as to help with formatting and
-// resizing an input image into the format expected as a network input.
-type ProcessedImage struct {
-	// The number of "pixels" in the input image corresponding to a single
-	// pixel in the 28x28 output image.
-	dx, dy float32
-
-	// The input image being transformed
-	pic image.Image
-
-	// If true, the grayscale values in the postprocessed image will be
-	// inverted, so that dark colors in the original become light, and vice
-	// versa. Recall that the network expects black backgrounds, so this should
-	// be set to true for images with light backgrounds.
-	Invert bool
-}
-
-func (p *ProcessedImage) ColorModel() color.Model {
-	return color.Gray16Model
-}
-
-func (p *ProcessedImage) Bounds() image.Rectangle {
-	return image.Rect(0, 0, 28, 28)
-}
-
-// At Returns an average grayscale value using the pixels in the input image.
-func (p *ProcessedImage) At(x, y int) color.Color {
-	if (x < 0) || (x >= 28) || (y < 0) || (y >= 28) {
-		return color.Black
-	}
-
-	// Compute the window of pixels in the input image we'll be averaging.
-	startX := int(float32(x) * p.dx)
-	endX := int(float32(x+1) * p.dx)
-	if endX == startX {
-		endX = startX + 1
-	}
-	startY := int(float32(y) * p.dy)
-	endY := int(float32(y+1) * p.dy)
-	if endY == startY {
-		endY = startY + 1
-	}
-
-	// Compute the average brightness over the window of pixels
-	var sum float32
-	var nPix int
-	for row := startY; row < endY; row++ {
-		for col := startX; col < endX; col++ {
-			c := p.pic.At(col, row)
-			grayValue := color.Gray16Model.Convert(c).(color.Gray16).Y
-			sum += float32(grayValue) / 0xffff
-			nPix++
-		}
-	}
-
-	brightness := grayscaleFloat(sum / float32(nPix))
-	if p.Invert {
-		brightness = 1.0 - brightness
-	}
-	return brightness
-}
-
-// GetNetworkInput Returns a slice of data that can be used as the input to the onnx network.
-func (p *ProcessedImage) GetNetworkInput() []float32 {
-	toReturn := make([]float32, 0, 28*28)
-	for row := 0; row < 28; row++ {
-		for col := 0; col < 28; col++ {
-			c := float32(p.At(col, row).(grayscaleFloat))
-			toReturn = append(toReturn, c)
-		}
-	}
-	return toReturn
-}

extensions/functions/mnist/mnist.json

@@ -12,8 +12,8 @@
       "zh_CN": "https://ekuiper.org/docs/zh/latest/sqls/functions/custom_functions.html"
     },
     "description": {
-      "en_US": "todo :Example plugin to demonstrate inferring Tensorflow lite model to label an image",
-      "zh_CN": "todo:示例插件，演示如何使用Tensorflow lite模型对图像进行标记推断。"
+      "en_US": "Example plugin to demonstrate inferring onnx model to label an image",
+      "zh_CN": "示例插件，演示如何使用onnx模型对图像进行标记推断。"
     }
   },
   "name": "mnist",
@@ -22,16 +22,16 @@
       "name": "mnist",
       "example": "mnist(col1)",
       "hint": {
-        "en_US": "todo Label an image by tensorflow lite model.",
-        "zh_CN": "todo采用 tensorflow lite 模型标记图片。"
+        "en_US": "Inference the number in pic by onnx model.",
+        "zh_CN": "采用 onnx 模型推理图中数字。"
       },
       "args": [
         {
           "name": "image",
           "hidden": false,
           "optional": false,
           "control": "field",
-          "type": "string",
+          "type": "[]byte",
           "hint": {
             "en_US": "Input image",
             "zh_CN": "输入图像"
@@ -45,8 +45,8 @@
       "return": {
         "type": "string",
         "hint": {
-          "en_US": "Image Label",
-          "zh_CN": "图像标注"
+          "en_US": "Probability",
+          "zh_CN": "图像中的数字概率"
         }
       },
       "node": {

extensions/functions/mnist/mnist_test.go

@@ -2,7 +2,6 @@ package main
 
 import (
 	"fmt"
-	mqtt "github.com/eclipse/paho.mqtt.golang"
 	"github.com/lf-edge/ekuiper/contract/v2/api"
 	ort "github.com/yalue/onnxruntime_go"
 	_ "image/gif"
@@ -11,10 +10,8 @@ import (
 	"os"
 	"sync"
 	"testing"
-	"time"
 )
 
-// todo 测试文件仿照tf lite
 func Test_mnist_Exec(t *testing.T) {
 	type fields struct {
 		modelPath         string
@@ -62,76 +59,19 @@ func Test_mnist_Exec(t *testing.T) {
 		t.Run(tt.name, func(t *testing.T) {
 			f := &MnistFunc{
 				modelPath:         tt.fields.modelPath,
-				once:              tt.fields.once,
+				once:              sync.Once{},
 				inputShape:        tt.fields.inputShape,
 				outputShape:       tt.fields.outputShape,
 				sharedLibraryPath: tt.fields.sharedLibraryPath,
 				initModelError:    tt.fields.initModelError,
 			}
 
 			out, got1 := f.Exec(tt.args.in0, tt.args.args)
-			
+
 			if !got1 {
 				t.Errorf("Exec() error = %v, wantErr %v", got1, tt.want1)
 			}
 			fmt.Println(out)
 		})
 	}
 }
-
-/*
-➜  mnist git:(torch_dev_swx) ✗ go test -v -cover
-=== RUN   Test_mnist_Exec
-=== RUN   Test_mnist_Exec/test1
-Output probabilities:
-  0: 1.350922
-  1: 1.149244
-  2: 2.231948
-  3: 0.826893
-  4: -3.473754
-  5: 1.200287
-  6: -1.185765
-  7: -5.960128
-  8: 4.764542
-  9: -2.345179
- probably a 8, with probability 4.764542
-
------------------------------
-true --- PASS: Test_mnist_Exec (0.03s)
-    --- PASS: Test_mnist_Exec/test1 (0.03s)
-PASS
-coverage: 58.7% of statements
-ok      github.com/lf-edge/ekuiper/v2/extensions/functions/mnist        0.030s
-
-
-
-*/
-
-func TestPic(t *testing.T) {
-	const TOPIC = "tfdmnist"
-
-	images := []string{
-		"img.png",
-		// 其他你需要的图像
-	}
-	opts := mqtt.NewClientOptions().AddBroker("tcp://localhost:1883")
-	client := mqtt.NewClient(opts)
-	if token := client.Connect(); token.Wait() && token.Error() != nil {
-		panic(token.Error())
-	}
-	for _, image := range images {
-		fmt.Println("Publishing " + image)
-		payload, err := os.ReadFile(image)
-		if err != nil {
-			fmt.Println(err)
-			continue
-		}
-		if token := client.Publish(TOPIC, 0, false, payload); token.Wait() && token.Error() != nil {
-			fmt.Println(token.Error())
-		} else {
-			fmt.Println("Published " + image)
-		}
-		time.Sleep(1 * time.Second)
-	}
-	client.Disconnect(0)
-}