请教关于如何调用静态文件以及LeapMotion插件的一点问题

[YIN-jw]

您好，抱歉再次打扰，关于静态文件想再请教您几个问题。
我在之后按照knn的安装步骤安装好了posenet插件，若有需要详细步骤在：https://github.com/YIN-jw/scratch3.0_posenet_install。其中在安装依赖时是直接yarn add @tensorflow/tfjs-core,yarn add @tensorflow-models/posenet（没有考虑tfjs库的版本号）。 之后能够在网页版中正常运行，但是knn插件又不能正常使用了。再重新yarn add @tensorflow/tfjs@0.10.0又能使用了，但是posenet不能使用。我猜想是因为knn插件使用的是tfjs@0.10.0（yarn add @tensorflow/tfjs@0.10.0）,posenet需要的版本比knn要高，所以造成了knn不能再使用。
目前对于这个问题没有想到解决办法，请问您可以给我一些建议吗？
另外，在使用LeapMotion这个插件时，出现的结果都是‘0’，想请问一下这个插件的用法是什么？

[summerscar]

你查一下yarn.lock knn tfjs的引用关系
在package.json 增加
"resolutions": {
"knn/{package}/{package}/ @tensorflow/tfjs": "0.1.0"
}
指定一下版本
参考 https://classic.yarnpkg.com/zh-Hans/docs/selective-version-resolutions/

leapmotion 需要一个硬件设备
https://www.ultraleap.com/product/leap-motion-controller/

[YIN-jw]

抱歉我回复的晚了，感谢您的回复，我还是不太清楚具体该怎么修改。查看了位于../scratch-vm下的yarn.lock文件，目前依赖是这样的：
"@tensorflow-models/knn-classifier@^1.2.2":
version "1.2.2"
resolved "https://registry.yarnpkg.com/@tensorflow-models/knn-classifier/-/knn-classifier-1.2.2.tgz#a5a9045b3d225a06e60f2b1cc2de56bdac6748e8"
integrity sha512-QRnkCf7ErOxSRtvJ6yCwhlLREPcBJGaXRanF46f0iY6ii3Sybjb6Ux0qnNPTrHZChD0izPa3Z4GQEgSAykiHkQ==

"@tensorflow-models/mobilenet@^2.0.4":
version "2.0.4"
resolved "https://registry.yarnpkg.com/@tensorflow-models/mobilenet/-/mobilenet-2.0.4.tgz#7f40771fcc650f50febd49be72372d39c15f6f95"
integrity sha512-fkcxDBldlYitDCV88xETwAQoBIC8xYKdDu80uBDuBZBfNC4k/jeVbZHM28WjvTDsJN7d7dYO4NbIkRAH0XHKSQ==

"@tensorflow-models/posenet@^2.2.1":
version "2.2.1"
resolved "https://registry.yarnpkg.com/@tensorflow-models/posenet/-/posenet-2.2.1.tgz#2f0a68d909842f59eec195f23c6f3d4cebc53fa8"
integrity sha512-n9/g6DfjAyrBTf/zt1haRCyWsgALxUCzg9/Ks3Y2mbYavRZVSCSTRPy/qlE5Hr4tLfyckGfDN14zmGTthNcg/g==

"@tensorflow/tfjs-converter@^1.7.1":
version "1.7.1"
resolved "https://registry.yarnpkg.com/@tensorflow/tfjs-converter/-/tfjs-converter-1.7.1.tgz#b96abbb1b38e96088709386cc21f4902809cabbf"
integrity sha512-CTQmTWAM+VqjxX1Qw9RhVrl32e1ITH2oVuWs/KgxXyyGNE9BLFQxFJhxQz/BPbxHFoIqZUK8RZiFrQudabx/lw==

"@tensorflow/tfjs-core@0.8.1":
version "0.8.1"
resolved "https://registry.yarnpkg.com/@tensorflow/tfjs-core/-/tfjs-core-0.8.1.tgz#d93e87302e29620906003b697d2d8596410ad712"
integrity sha512-3XeO7gAxeeIlTUItxiZXazuhM3QMUrIPMq8wPm8IPFU+ucUy7zAAmD1elQ9pDeXYvWtb+lMK5mNiU5CBXs3ACQ==
dependencies:
seedrandom "~2.4.3"

"@tensorflow/tfjs-core@1.7.0":
version "1.7.0"
resolved "https://registry.yarnpkg.com/@tensorflow/tfjs-core/-/tfjs-core-1.7.0.tgz#9207c8f2481c52a6a40135a6aaf21a9bb0339bdf"
integrity sha512-uwQdiklNjqBnHPeseOdG0sGxrI3+d6lybaKu2+ou3ajVeKdPEwpWbgqA6iHjq1iylnOGkgkbbnQ6r2lwkiIIHw==
dependencies:
"@types/offscreencanvas" "~2019.3.0"
"@types/seedrandom" "2.4.27"
"@types/webgl-ext" "0.0.30"
"@types/webgl2" "0.0.4"
node-fetch "~2.1.2"
seedrandom "2.4.3"

"@tensorflow/tfjs-core@^1.7.1":
version "1.7.1"
resolved "https://registry.yarnpkg.com/@tensorflow/tfjs-core/-/tfjs-core-1.7.1.tgz#145c3c70de0e0aef11ff7e9cadf7b3fa81ce13e2"
integrity sha512-aMbvvhgBFLGsV1uBFPlPvuv6OAjA8BzjyLDWHoss5Tgaj6S6PPt8xE5ea8+cUKqz5qNVGYmj03qR5S7SmJ3qnw==
dependencies:
"@types/offscreencanvas" "~2019.3.0"
"@types/seedrandom" "2.4.27"
"@types/webgl-ext" "0.0.30"
"@types/webgl2" "0.0.4"
node-fetch "~2.1.2"
seedrandom "2.4.3"

"@tensorflow/tfjs-layers@0.5.0":
version "0.5.0"
resolved "https://registry.yarnpkg.com/@tensorflow/tfjs-layers/-/tfjs-layers-0.5.0.tgz#0ec8d07b46863a162d3e0c60fae3d1087d8aa3ce"
integrity sha512-C7vNAaHoe0+Nk6A1wRZ9iKLjDhZqFMCXEPO9fVz4cRWl2sVa/foYOxR/gLn6OYFlTZADA5D7fRK24KqsmZ7/SA==

"@tensorflow/tfjs@0.10.0":
version "0.10.0"
resolved "https://registry.yarnpkg.com/@tensorflow/tfjs/-/tfjs-0.10.0.tgz#c64f8740b46c2ba734dc0564d37e6a2c33055511"
integrity sha512-8nTBG9APxafEMETega1vZ5V8IzSH6205yIcButLSS++OtuCLGwTsdCFuUMaf5+RnXdxhcJaGWawc+eT55Wk5vg==
dependencies:
"@tensorflow/tfjs-core" "0.8.1"
"@tensorflow/tfjs-layers" "0.5.0"

在package.json文件中是这样的：
"dependencies": {
"@tensorflow-models/knn-classifier": "^1.2.2",
"@tensorflow-models/mobilenet": "^2.0.4",
"@tensorflow-models/posenet": "^2.2.1",
"@tensorflow/tfjs": "0.10.0",
"@tensorflow/tfjs-converter": "^1.7.1",
"@tensorflow/tfjs-core": "^1.7.1",
"@vernier/godirect": "1.5.0",
"arraybuffer-loader": "^1.0.6",
"atob": "2.1.2",
"babel-polyfill": "^6.26.0",
"btoa": "1.2.1",
"canvas": "^2.6.1",
"canvas-toBlob": "1.0.0",
"decode-html": "2.0.0",
"diff-match-patch": "1.0.4",
"face-api.js": "^0.22.2",
"format-message": "6.2.1",
"htmlparser2": "3.10.0",
"immutable": "3.8.1",
"jszip": "^3.1.5",
"minilog": "3.1.0",
"nets": "3.2.0",
"scratch-parser": "5.0.0",
"scratch-sb1-converter": "0.2.7",
"scratch-translate-extension-languages": "0.0.20190416132834",
"socket.io-client": "2.0.4",
"text-encoding": "0.7.0",
"worker-loader": "^1.1.1"
},

我尝试在上一行的逗号之后添加：
"resolutions": {
"knn//@tensorflow/tfjs": "0.10.0"
}
"resolutions": {
"knn/{package}/{package}/@tensorflow/tfjs": "0.10.0"
}
"resolutions": {
"scratch3_knn//@tensorflow/tfjs": "0.10.0"
}
"resolutions": {
"@tensorflow/knn-classifier/@tensorflow/tfjs": "0.10.0"
}

之后webpack-dev-server --https报错Unexpected string in JSON at position2513。看了相关的资料也没弄懂原因，能麻烦再指点一下吗？

[YIN-jw]

好像格式有点问题，修正：两个 ‘/’ 之间是用的 ‘*’ 号。

[summerscar]

我用knn和faceapi存在依赖问题是这样解决的
"resolutions": {
"face-api.js/tfjs-image-recognition-base/@tensorflow/tfjs-core": "1.2.9"
}

我不确定是否和你是同一个问题

[YIN-jw]

抱歉再请问一下，您在安装posenet插件时有遇到和knn插件版本不同，knn插件不能使用的情况吗？（我这边 knn和faceapi两个插件却可以同时使用）
刚开始安装knn插件用的是@tensorflow/tfjs@0.10.0（包括tfjs-core@0.8.1和tfjs-layers@0.5.0）,之后安装posenet插件使用了@tensorflow/tfjs-core@1.7.2和@tensorflow/tfjs-converter@1.7.2，将之前的版本覆盖了。
我在package.json中添加
"resolutions": {
"@tensorflow/knn-classifier/@tensorflow/tfjs-core": "0.8.1"
}对版本进行限制，webpack-dev-server --https可以运行通过了，但是控制台显示：

Uncaught Error: 'realDivide' not yet implemented or not found in the registry. Did you forget to import the kernel?
at ha (tf-core.esm.js:17)
at o.t.realDivide (tf-core.esm.js:17)
at tf-core.esm.js:17
at tf-core.esm.js:17
at tf-core.esm.js:17
at t.scopedRun (tf-core.esm.js:17)
at t.tidy (tf-core.esm.js:17)
at f (tf-core.esm.js:17)
at tf-core.esm.js:17
at t.scopedRun (tf-core.esm.js:17)]
tf-core.esm.js:17 Uncaught TypeError: Cannot read property 'backend' of undefined
at t.moveData (tf-core.esm.js:17)
at t.get (tf-core.esm.js:17)
at tf-core.esm.js:17
at Array.map ()
at o.runWebGLProgram (tf-core.esm.js:17)
at tf-core.esm.js:17
at Object.kernelFunc (tf-core.esm.js:17)
at f (tf-core.esm.js:17)
at tf-core.esm.js:17
at t.scopedRun (tf-core.esm.js:17)
感觉可能是版本没有限制正确或者是其他原因，感觉离解决很近了，不知道是不是哪里还没有考虑详细。