### Install with mirror configurations

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/guide-to-contributor.md

Use mirrors to speed up the installation of Python and packages.

```sh
$ BOA_TUNA=1 npm install
```

```bash
$ export BOA_CONDA_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda # this is for miniconda
$ export BOA_CONDA_INDEX=https://pypi.tuna.tsinghua.edu.cn/simple                 # this is for pip
$ npm install
```

--------------------------------

### Ava Unit Test Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/rfcs/0001-framework-migration.md

Demonstrates unit testing with Ava, including array shuffling verification and mocking file system operations using Sinon. Requires setup for Ava and Sinon.

```javascript
// ava
import test from 'ava';
import * as fs from 'fs-extra';
import * as sinon from 'sinon';
import { shuffle } from './public';

test('array shuffle', (t) => {
  const array = [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ];
  shuffle(array);
  t.notDeepEqual(array, [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]);
  t.deepEqual(array.sort(), [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]);
});

// serial hook
test.serial.afterEach(() => {
  sinon.restore();
});

// serial case
test.serial('test a', (t) => {
  const fsReadJsonMock = sinon.stub(fs, 'readJson').resolves({});
  t.deepEqual(await fs.readJson('mockFileName.json'), {});
});

test.serial('test b', (t) => {
  const fsReadJsonMock = sinon.stub(fs, 'readJson').resolves(undefined);
  t.is(await fs.readJson('mockFileName.json'), undefined);
});

test.todo('some todo cases');
```

--------------------------------

### Install Pipcook Client

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/contributing/contribute-a-script.md

Install the Pipcook CLI globally to use Pipcook commands.

```sh
$ npm install @pipcook/cli -g
$ pipcook -v
2.0.0
```

--------------------------------

### Install Node.js model dependencies

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/pipcook-models.md

Navigate to the model directory and install required dependencies before usage.

```bash
$ cd output/nodejs
$ npm install # To install deps
```

--------------------------------

### Install Pipcook CLI

Source: https://context7.com/alibaba/pipcook/llms.txt

Global installation of the Pipcook command-line tool via npm or Docker.

```bash
# Install Pipcook CLI globally
npm install -g @pipcook/cli

# Verify installation
pipcook --version

# Alternative: Install via Docker
docker pull pipcook/pipcook:latest
docker run -it pipcook/pipcook:latest /bin/bash
```

--------------------------------

### Build Pipcook from source

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/guide-to-contributor.md

Install dependencies and compile the project.

```bash
$ npm install
$ npm run build
```

--------------------------------

### Serve a Model

Source: https://github.com/alibaba/pipcook/blob/main/docs/README.md

Start a local server to serve the trained model.

```shell
$ pipcook serve ./output
ℹ preparing framework
ℹ preparing scripts
ℹ preparing artifact plugins
ℹ initializing framework packages
Pipcook has served at: http://localhost:9091
```

--------------------------------

### Install Dependencies with Tuna Mirror

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/contributing/guide-to-contributor.md

Install dependencies using the Tuna mirror for faster downloads of Python and packages.

```bash
$ BOA_TUNA=1 npm install
```

--------------------------------

### Install Pipcook CLI

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/pipcook-tools.md

Install the Pipcook command-line interface globally using npm.

```sh
$ npm install @pipcook/cli -g
```

--------------------------------

### Install Dependencies with Custom Mirrors

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/contributing/guide-to-contributor.md

Install dependencies specifying custom miniconda and Python index mirrors.

```bash
$ export BOA_CONDA_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda # this is for miniconda
$ export BOA_CONDA_INDEX=https://pypi.tuna.tsinghua.edu.cn/simple                 # this is for pip
$ npm install
```

--------------------------------

### Install Pipcook CLI Beta Version

Source: https://github.com/alibaba/pipcook/blob/main/docs/meta/PROJECT_GUIDE.md

Use this command to install the beta version of the Pipcook CLI globally. This provides access to the latest experimental features and bug fixes.

```bash
$ npm install @pipcook/pipcook-cli@beta -g
```

```bash
$ npm install --tag beta @pipcook/pipcook-cli -g
```

--------------------------------

### Installing Boa via npm

Source: https://github.com/alibaba/pipcook/blob/main/docs/tutorials/using-python-functions-in-nodejs.md

Command to install the Boa package independently.

```bash
$ npm install @pipcook/boa
```

--------------------------------

### Install Project Dependencies

Source: https://github.com/alibaba/pipcook/blob/main/README.md

Install project dependencies using npm after cloning the repository. This command ensures all necessary packages are available.

```shell
$ npm install
```

--------------------------------

### Install Pipcook CLI

Source: https://github.com/alibaba/pipcook/blob/main/notebooks/pipcook_image_classification.ipynb

Installs the Pipcook CLI tool globally and creates a symbolic link for system-wide access.

```python
!npm install @pipcook/cli -g
!rm -f /usr/bin/pipcook
!ln -s /usr/local/lib/nodejs/node-v12.19.0-linux-x64/bin/pipcook /usr/bin/pipcook
```

--------------------------------

### Initialize Pipcook Environment

Source: https://github.com/alibaba/pipcook/blob/main/notebooks/pipcook_object_detection.ipynb

Downloads and installs Node.js 12.x and configures the system PATH for Pipcook operations.

```python
!wget -P /tmp https://nodejs.org/dist/v12.19.0/node-v12.19.0-linux-x64.tar.xz
!rm -rf /usr/local/lib/nodejs
!mkdir -p /usr/local/lib/nodejs
!tar -xJf /tmp/node-v12.19.0-linux-x64.tar.xz -C /usr/local/lib/nodejs
!sh -c 'echo "export PATH=/usr/local/lib/nodejs/node-v12.19.0-linux-x64/bin:\$PATH" >> /etc/profile'
!rm -f /usr/bin/node
!rm -f /usr/bin/npm
!ln -s /usr/local/lib/nodejs/node-v12.19.0-linux-x64/bin/node /usr/bin/node
!ln -s /usr/local/lib/nodejs/node-v12.19.0-linux-x64/bin/npm /usr/bin/npm
!npm config delete registry

import os
PATH_ENV = os.environ['PATH']
%env PATH=/usr/local/lib/nodejs/node-v12.19.0-linux-x64/bin:${PATH_ENV}
```

--------------------------------

### Generate Arithmetic Data and Create Dataset

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/contribute-a-script.md

A complete example showing data generation logic and the corresponding DataSource entry function.

```javascript
/**
 * Generate examples.
 *
 * Each example consists of a question, e.g., '123+456' and and an
 * answer, e.g., '579'.
 *
 * @param digits Maximum number of digits of each operand of the
 * @param numExamples Number of examples to generate.
 * @param invert Whether to invert the strings in the question.
 * @returns The generated examples and digit array.
 */
function generateData(digits, numExamples, invert = false) {
    const digitArray = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9'];
    const arraySize = digitArray.length;

  const output = [];
  const maxLen = digits + 1 + digits;

  const f = () => {
    let str = '';
    while (str.length < digits) {
      const index = Math.floor(Math.random() * arraySize);
      str += digitArray[index];
    }
    return parseInt(str);
  };

  const seen = new Set();
  while (output.length < numExamples) {
    const a = f();
    const b = f();
    const sorted = b > a ? [a, b] : [b, a];
    const key = sorted[0] + '`' + sorted[1];
    if (seen.has(key)) {
      continue;
    }
    seen.add(key);

    // Pad the data with spaces such that it is always maxLen.
    const q = `${a}+${b}`;
    const query = q + ' '.repeat(maxLen - q.length);
    let ans = (a + b).toString();
    // Answer can be of maximum size `digits + 1`.
    ans += ' '.repeat(digits + 1 - ans.length);

    if (invert) {
      throw new Error('invert is not implemented yet');
    }
    output.push([query, ans]);
  }
  return output;
}

/**
 * This is the entry of datasource script
 */
module.exports = async (options, context) => {
  // 1
  let {
    digits = '2',
    numExamples = '100'
  } = options;
  digits = parseInt(digits);
  numExamples = parseInt(numExamples);
  // 2
  const dataCook = context.dataCook;
  // 3
  const data = generateData(digits, numExamples);
  const split = Math.floor(numExamples * 0.9);
  const trainData = data.slice(0, split).map(item => ({ label: item[1], data: item[0] }));
  const testData = data.slice(split).map(item => ({ label: item[1], data: item[0] }));
  // 4
  const meta = {
    type: dataCook.Dataset.Types.DatasetType.General,
    size: {
      test: testData.length,
      train: trainData.length
    }
  };
  // 5
  return dataCook.Dataset.makeDataset({ trainData, testData }, meta);
};
```

--------------------------------

### Install Pipcook CLI

Source: https://github.com/alibaba/pipcook/blob/main/notebooks/pipcook_object_detection.ipynb

Installs the Pipcook command-line interface globally via npm and creates a symbolic link.

```python
!npm install @pipcook/cli -g --unsafe-perm
!rm -f /usr/bin/pipcook
!ln -s /usr/local/lib/nodejs/node-v12.19.0-linux-x64/bin/pipcook /usr/bin/pipcook
```

--------------------------------

### Pipcook Pipeline Schema v2.0 Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/rfcs/0000-new-pipeline.md

This example demonstrates a typical Pipcook pipeline configuration using the v2.0 schema, showcasing the array-based 'plugins' structure with various plugin types and their parameters.

```json
{
  "$id": "http://alibaba.github.io/pipcook/pipeline-2.0.json",
  "$schema": "http://json-schema.org/draft-07/schema",
  "default": {},
  "description": "The root schema comprises the entire JSON document.",
  "examples": [
    {
      "specVersion": "2.0",
      "plugins": [
        {
          "type": "dataCollect",
          "package": "@pipcook/plugins-chinese-poem-data-collect",
          "params": {
            "url": "foobar"
          }
        },
        {
          "type": "dataAccess",
          "package": "@pipcook/plugins-textline-data-access"
        },
        {
          "type": "dataProcess",
          "package": "@pipcook/plugins-textline-data-process",
          "params": {
            "arg1": "foobar"
          }
        },
        {
          "type": "dataProcess",
          "package": "@pipcook/plugins-textline-data-process",
          "params": {
            "arg2": "foobar"
          }
        },
        {
          "type": "modelDefine",
          "package": "@pipcook/plugins-tfjs-text-lstm-model-define"
        },
        {
          "type": "modelTrain",
          "package": "@pipcook/plugins-tfjs-text-lstm-model-train",
          "params": {
            "epochs": 200
          }
        },
        {
          "type": "modelEvaluate",
          "package": "@pipcook/plugins-tfjs-text-lstm-model-evaluate"
        },
        {
          "type": "deploy",
          "package": "@pipcook/plugins-nodejs-deploy-aws-ec2",
          "params": {
            "key": "..."
          }
        }
      ]
    }
  ],
  "required": [
    "plugins"
  ],
  "title": "The JSON schema for Pipcook Pipeline.",
  "type": "object",
  "properties": {
    "specVersion": {
      "$id": "#/properties/specVersion",
      "type": "string",
      "title": "The version schema",
      "description": "The pipeline spec version",
      "default": "",
      "examples": [
        "2.0"
      ]
    },
    "plugins": {
      "$id": "#/properties/plugins",
      "type": "array",
      "title": "The plugins schema",
      "description": "An explanation about the purpose of this instance.",
      "additionalItems": false,
      "items": {
        "$id": "#/properties/plugins/items",
        "anyOf": [
          {
            "$id": "#/properties/plugins/items/data-collect",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to collect the dataset from the source url.",
            "type": "object",
            "minItems": 1,
            "maxItems": 1,
            "properties": {
              "type": {
                "enum": [ "dataCollect" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "properties": {
                  "url": {
                    "type": "string",
                    "title": "the data source url.",
                    "optional": true
                  }
                },
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          },
          {
            "$id": "#/properties/plugins/items/data-access",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to create the common `UniDataset` object.",
            "type": "object",
            "minItems": 1,
            "maxItems": 1,
            "properties": {
              "type": {
                "enum": [ "dataAccess" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          },
          {
            "$id": "#/properties/plugins/items/data-process",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to process the sample/example in the `UniDataset`.",
            "type": "object",
            "minItems": 0,
            "properties": {
              "type": {
                "enum": [ "dataProcess" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "properties": {
                  "arg1": {
                    "type": "string",
                    "optional": true
                  },
                  "arg2": {
                    "type": "string",
                    "optional": true
                  }
                },
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          },
          {
            "$id": "#/properties/plugins/items/model-define",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to define the model.",
            "type": "object",
            "minItems": 1,
            "maxItems": 1,
            "properties": {
              "type": {
                "enum": [ "modelDefine" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          },
          {
            "$id": "#/properties/plugins/items/model-train",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to train the model.",
            "type": "object",
            "minItems": 1,
            "maxItems": 1,
            "properties": {
              "type": {
                "enum": [ "modelTrain" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "properties": {
                  "epochs": {
                    "type": "integer",
                    "title": "the number of epochs to train the model.",
                    "optional": true
                  }
                },
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          },
          {
            "$id": "#/properties/plugins/items/model-evaluate",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to evaluate the model.",
            "type": "object",
            "minItems": 1,
            "maxItems": 1,
            "properties": {
              "type": {
                "enum": [ "modelEvaluate" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          },
          {
            "$id": "#/properties/plugins/items/deploy",
            "required": [
              "type",
              "package"
            ],
            "title": "a type of plugin to deploy the model.",
            "type": "object",
            "minItems": 1,
            "maxItems": 1,
            "properties": {
              "type": {
                "enum": [ "deploy" ],
                "type": "string"
              },
              "package": {
                "type": "string",
                "title": "The package uri of the plugin",
                "default": ""
              },
              "params": {
                "default": {},
                "title": "The params of the plugin",
                "type": "object",
                "properties": {
                  "key": {
                    "type": "string",
                    "optional": true
                  }
                },
                "additionalProperties": true
              }
            },
            "additionalProperties": true
          }
        ]
      }
    }
  }
}
```

--------------------------------

### Install Pipcook via NPM

Source: https://github.com/alibaba/pipcook/blob/main/docs/INSTALL.md

Use this command to install Pipcook globally via NPM. This is the recommended method for most users.

```sh
$ npm install -g @pipcook/cli
```

--------------------------------

### Framework Directory Structure Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/intro-to-framework.md

This is a typical directory structure for a Pipcook framework package, including the framework description file, initialization script, and node_modules.

```sh
├── framework.json
├── index.js
└── node_modules
```

--------------------------------

### Pipcook Working Directory Structure

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/intro-to-pipeline.md

Example of the directory structure generated after pipeline execution.

```text
  ├── pipeline-config.json
  ├── cache
  ├── data
  ├── framework
  ├── model
  └── scripts
```

--------------------------------

### Install Python Packages for Boa

Source: https://context7.com/alibaba/pipcook/llms.txt

Manage Python dependencies using the 'bip' wrapper and configure PYTHONPATH for custom environments.

```bash
# Install Python packages using bip (Boa's pip wrapper)
./node_modules/.bin/bip install numpy
./node_modules/.bin/bip install tensorflow
./node_modules/.bin/bip install scikit-learn
./node_modules/.bin/bip install pandas

# For virtualenv or conda users, set PYTHONPATH
export PYTHONPATH=/Users/venv/lib/python3.7/site-packages

# Then use in JavaScript
const boa = require('@pipcook/boa');
const np = boa.import('numpy');
const sklearn = boa.import('sklearn');
```

--------------------------------

### Install Python Packages with bip

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/intro-to-boa.md

Use the 'bip' command, an alias for pip, to install Python packages within the Boa-managed conda environment.

```sh
./node_modules/.bin/bip install <package-name>
```

--------------------------------

### Jasmine Unit Test Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/rfcs/0001-framework-migration.md

Illustrates unit testing with Jasmine, focusing on verifying array shuffling. This example uses Jasmine's `describe` and `it` syntax.

```javascript
// jasmine
import { shuffle } from './public';

describe('public utils', () => {
  it('test if the array is shuffled', () => {
    const array = [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ];
    shuffle(array);
    expect(array).not.toEqual([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]);
    expect(array.sort()).toEqual([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]);
  });
}

```

--------------------------------

### Training Log Output

Source: https://github.com/alibaba/pipcook/blob/main/docs/tutorials/component-image-classification.md

Example of the console output generated during the model training process, showing epoch progress, loss, and accuracy metrics.

```text
Epoch 1/15
187/187 [==============================] - 12s 65ms/step - loss: 0.0604 - accuracy: 0.9823 - val_loss: 8.8755 - val_accuracy: 0.4112
Epoch 2/15
187/187 [==============================] - 11s 61ms/step - loss: 0.0056 - accuracy: 0.9993 - val_loss: 5.5883 - val_accuracy: 0.4925
Epoch 3/15
187/187 [==============================] - 11s 59ms/step - loss: 0.0107 - accuracy: 0.9980 - val_loss: 0.3830 - val_accuracy: 0.8388
...
187/187 [==============================] - 11s 61ms/step - loss: 3.0090e-05 - accuracy: 1.0000 - val_loss: 1.5646e-08 - val_accuracy: 1.0000
Epoch 14/15
187/187 [==============================] - 11s 61ms/step - loss: 5.1657e-05 - accuracy: 1.0000 - val_loss: 1.9073e-08 - val_accuracy: 1.0000
Epoch 15/15
187/187 [==============================] - 11s 61ms/step - loss: 5.1657e-05 - accuracy: 1.0000 - val_loss: 1.9073e-08 - val_accuracy: 1.0000
```

--------------------------------

### PascalVOC XML Annotation Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/spec/dataset.md

An example of an XML file used for annotating image data in the PascalVOC format. It includes details like folder, filename, image size, and object bounding boxes.

```xml
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<annotation>
  <folder>folder path</folder>
  <filename>image name</filename>
  <size>
    <width>width</width>
    <height>height</height>
  </size>
  <object>
    <name>category name</name>
    <bndbox> <!--this is not necessary for image classification problem-->
      <xmin>left</xmin>
      <ymin>top</ymin>
      <xmax>right</xmax>
      <ymax>bottom</ymax>
    </bndbox>
  </object>
</annotation>
```

--------------------------------

### Build Pipcook Project

Source: https://github.com/alibaba/pipcook/blob/main/README.md

Build the Pipcook project after installing dependencies. This command compiles the project for development or production use.

```shell
$ npm run build
```

--------------------------------

### Creating a TensorFlow model in Node.js

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/tutorials/using-python-functions-in-nodejs.md

Shows how to import TensorFlow.js and define a Keras model by inheriting from Python's Model class. This example highlights the use of boa.kwargs() for specifying activation functions in layers.

```javascript
const boa = require('@pipcook/boa');
const tf = boa.import('tensorflow');
const { layers, Model } = tf.keras;

class TestModel extends Model {
  constructor() {
    super();
    this.conv1 = layers.Conv2D(32, 3, boa.kwargs({ activation: 'relu' }));
    this.flatten = layers.Flatten();
    this.d1 = layers.Dense(128, boa.kwargs({ activation: 'relu' }));
    this.d2 = layers.Dense(10, boa.kwargs({ activation: 'softmax' }));
  }
  call(x) {
    return this.conv1(x)
      .flatten(x)
      .d1(x).d2(x);
  }
}
```

--------------------------------

### Pull Pipcook Docker Image

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/INSTALL.md

Download the latest Pipcook Docker image from the Alibaba source repository. This is an alternative installation method.

```sh
$ docker pull registry.cn-beijing.aliyuncs.com/pipcook/pipcook:latest
```

--------------------------------

### Execute Training Pipeline

Source: https://github.com/alibaba/pipcook/blob/main/docs/tutorials/component-image-classification.md

Command to initiate the training process using a specified configuration file and output directory.

```shell
$ pipcook run image-classification.json --output ./classification
```

--------------------------------

### Serve Pipcook Model

Source: https://github.com/alibaba/pipcook/blob/main/README.md

Serve a trained Pipcook model for deployment. This command prepares the framework, scripts, and plugins, then initializes the packages.

```shell
$ pipcook serve ./output
```

--------------------------------

### Create VS Code Launch Configuration

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/contributing/contribute-a-script.md

Set up a launch.json file in VS Code to enable debugging of Pipcook run commands.

```json
{
  "version": "0.2.0",
  "configurations": [
    {
      "type": "pwa-node",
      "request": "launch",
      "name": "Launch Program",
      "skipFiles": [
        "<node_internals>/**"
      ],
      "runtimeExecutable": "pipcook",
      "runtimeArgs": [ "run", "${workspaceFolder}/debug/addition-rnn.json", "--output", "${workspaceFolder}/debug/workspace" ]
    }
  ]
}
```

--------------------------------

### 配置 Pipcook 安装源

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/faq/pipcook-framework.md

通过 pipcook init 命令指定 npm 客户端或使用清华源以加速安装过程。

```bash
pipcook init -c <npm client>
```

```bash
pipcook init --tuna
```

--------------------------------

### Pipcook Run Help

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/pipcook-tools.md

Display help information for the 'pipcook run' command to see available options.

```sh
$ pipcook run --help
```

--------------------------------

### Check Pipcook Version

Source: https://github.com/alibaba/pipcook/blob/main/notebooks/pipcook_object_detection.ipynb

Verifies that the Pipcook CLI is installed correctly by checking its version.

```python
!sudo pipcook -v
```

--------------------------------

### Run tests

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/guide-to-contributor.md

Execute the test suite for the entire project or specific packages.

```bash
$ npm test
```

```bash
$ ./node_modules/.bin/lerna run --scope <package_name>
```

--------------------------------

### Create VS Code Debug Configuration Files

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/contribute-a-script.md

Create the .vscode directory and the launch.json file for debugging in VS Code.

```sh
$ mkdir .vscode
$ touch .vscode/launch.json
```

--------------------------------

### Python Generator Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/intro-to-boa.md

A Python generator function that counts down from a given number to zero, yielding each number.

```python
def count_down(count):
  while count >= 0:
    yield count
    count -= 1
```

--------------------------------

### Pull Pipcook Docker Image

Source: https://github.com/alibaba/pipcook/blob/main/docs/INSTALL.md

Download the latest Pipcook Docker image. Ensure Docker is installed and running.

```sh
$ docker pull pipcook/pipcook:latest
```

--------------------------------

### Specify GPU for Training

Source: https://github.com/alibaba/pipcook/blob/main/docs/faq/plugins.md

Set the CUDA_VISIBLE_DEVICES environment variable to the desired GPU index before starting the Pipcook daemon.

```bash
$CUDA_VISIBLE_DEVICES
```

--------------------------------

### Implement a Basic DataSource Script

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/contribute-a-script.md

A template for a DataSource script entry function using the datacook context to create a dataset.

```javascript
/*
 * This is the entry of datasource script
 */
module.exports = async (options, context) => {
  let {
    // options from query, note that their types are `srting` or `string[]`
  } = options;
  // should get the instance of datacook from context
  const dataCook = context.dataCook;
  dataCook = context.dataCook; /**
   * create dataset here
   */
  const trainData = [ { label: 'label-1', data: 'data-1' }, { label: 'label-2', data: 'data-2' } ];
  const testData = [ { label: 'label-1', data: 'data-1' }, { label: 'label-2', data: 'data-2' } ];
  const meta = {
    type: dataCook.Dataset.Types.DatasetType.General,
    size: {
      test: testData.length,
      train: trainData.length
    }
  };
  return dataCook.Dataset.makeDataset({ trainData, testData }, meta);
};
```

--------------------------------

### Get Python Built-in Functions

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/intro-to-boa.md

Access Python's built-in functions like 'len' and 'range' through boa.builtins().

```javascript
const { len, range } = boa.builtins();
len([1, 2, 3]); // 3
len(range(0, 10)); // 10
```

--------------------------------

### Importing OS module and using its functions in Node.js

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/tutorials/using-python-functions-in-nodejs.md

Demonstrates importing the 'os' Python module using boa.import() and calling its functions like getpid(). It also shows how to use boa.kwargs() for keyword arguments when calling Python functions like os.makedirs().

```javascript
const boa = require('@pipcook/boa');
const os = boa.import('os');
console.log(os.getpid()); // prints the pid from python.
// using keyword arguments namely \`kwargs\`
os.makedirs('..', boa.kwargs({
  mode: 0x777,
  exist_ok: false,
}));
// using bult-in functions
const { range, len } = boa.builtins();
const list = range(0, 10); // create a range array
console.log(len(list)); // 10
console.log(list[2]); // 2
```

--------------------------------

### Configure YOLO Training Options

Source: https://github.com/alibaba/pipcook/blob/main/example/pipelines/README-CN.md

Define training parameters including batch size, epochs, and early-stopping patience.

```json
{
  "options": {
    "framework": "tfjs@3.8",
    "gpu": false,
    "train": {
      "epochs": 100,
      "batchSize": 16,
      "patience": 10
    }
  }
}
```

--------------------------------

### Configure Object Detection Pipeline (YOLO)

Source: https://context7.com/alibaba/pipcook/llms.txt

Sets up a pipeline for object detection using the YOLO architecture. Specify the datasource, dataflow scripts, model, and training options.

```json
{
  "specVersion": "2.0",
  "type": "ObjectDetection",
  "datasource": "https://cdn.jsdelivr.net/gh/imgcook/pipcook-script@9d210de/scripts/object-detection-yolo/build/datasource.js?format=pascalvoc&url=https://pipcook-cloud.oss-cn-hangzhou.aliyuncs.com/dataset/mask.zip",
  "dataflow": [
    "https://cdn.jsdelivr.net/gh/imgcook/pipcook-script@9d210de/scripts/object-detection-yolo/build/dataflow.js?size=416&size=416"
  ],
  "model": "https://cdn.jsdelivr.net/gh/imgcook/pipcook-script@9d210de/scripts/object-detection-yolo/build/model.js",
  "options": {
    "framework": "tfjs@3.8",
    "train": {
      "epochs": 100,
      "batchSize": 16,
      "patience": 10
    }
  }
}
```

--------------------------------

### CSV Format for Text Classification

Source: https://context7.com/alibaba/pipcook/llms.txt

Example of a CSV file used for text classification tasks. Each row contains a text sample and its corresponding category label.

```csv
name,category
"This product is amazing",positive
"Terrible quality, do not buy",negative
"Average product, nothing special",neutral
"Highly recommended!",positive
"Complete waste of money",negative
```

--------------------------------

### Simplified Data Source Configuration

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/contributing/contribute-a-script.md

When using the 'file' protocol for data sources in a pipeline, the protocol header can be omitted.

```json
{
  "datasource": "/data/my-datasource.js"
}
```

--------------------------------

### Run Pipcook Docker Container

Source: https://github.com/alibaba/pipcook/blob/main/docs/INSTALL.md

Start an interactive terminal session within the Pipcook Docker container. This allows you to run Pipcook commands inside the isolated environment.

```sh
$ docker run -it pipcook/pipcook:latest /bin/bash
```

--------------------------------

### PascalVOC XML Annotation for Object Detection

Source: https://context7.com/alibaba/pipcook/llms.txt

Example of an XML file structured in PascalVOC format for object detection tasks. It includes image metadata and bounding box annotations for objects.

```xml
<!-- annotations/train/image_name.xml -->
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<annotation>
  <folder>images</folder>
  <filename>image_name.jpg</filename>
  <size>
    <width>800</width>
    <height>600</height>
  </size>
  <object>
    <name>button</name>
    <bndbox>
      <xmin>100</xmin>
      <ymin>150</ymin>
      <xmax>200</xmax>
      <ymax>180</ymax>
    </bndbox>
  </object>
  <object>
    <name>input</name>
    <bndbox>
      <xmin>50</xmin>
      <ymin>200</ymin>
      <xmax>300</xmax>
      <ymax>230</ymax>
    </bndbox>
  </object>
</annotation>
```

--------------------------------

### Specify Dataset Format (Local Path)

Source: https://github.com/alibaba/pipcook/blob/main/example/pipelines/README.md

Configure the datasource to use a local dataset directory with the PascalVoc format. The 'url' parameter is set to 'file://' followed by the local path.

```json
{
  "datasource": "https://cdn.jsdelivr.net/gh/imgcook/pipcook-script@9d210de/scripts/object-detection-yolo/build/datasource.js?format=pascalvoc&url=file:///path/to/dataset-directory"
}
```

--------------------------------

### Serve Models via HTTP

Source: https://context7.com/alibaba/pipcook/llms.txt

Deploy trained models as REST APIs for real-time predictions, with options for custom ports and debugging.

```bash
# Serve model on default port 9091
pipcook serve ./output
# Output:
# ℹ preparing framework
# ℹ preparing scripts
# ℹ preparing artifact plugins
# ℹ initializing framework packages
# Pipcook has served at: http://localhost:9091

# Serve on custom port
pipcook serve ./output --port 8080

# Serve with debug mode
pipcook serve ./model-workspace --port 3000 --debug

# Test the prediction endpoint (for image classification)
curl -X POST http://localhost:9091/predict \
  -F "file=@./test-image.jpg"
```

--------------------------------

### Python Functions in Node.js Worker Threads

Source: https://github.com/alibaba/pipcook/blob/main/docs/manual/intro-to-boa.md

Example demonstrating how to run Python functions in a separate Node.js worker thread using '@pipcook/boa' and 'worker_threads'. This allows non-blocking execution of Python code.

```javascript
const { Worker, isMainThread, workerData, parentPort } = require('worker_threads');
const boa = require('@pipcook/boa');
const pybasic = boa.import('tests.base.basic'); // a Python example
const { SharedPythonObject, symbols } = boa;

class Foobar extends pybasic.Foobar {
  hellomsg(x) {
    return `hello <${x}> on ${this.test}(${this.count})`;
  }
}

if (isMainThread) {
  const foo = new Foobar();
  const worker = new Worker(__filename, {
    workerData: {
      foo: new SharedPythonObject(foo),
    },
  });
  let alive = setInterval(() => {
    const ownership = foo[symbols.GetOwnershipSymbol]();
    console.log(`ownership should be ${expectedOwnership}.`);
  }, 1000);

  worker.on('message', state => {
    if (state === 'done') {
      console.log('task is completed');
      setTimeout(() => {
        clearInterval(alive);
        console.log(foo.ping('x'));
      }, 1000);
    }
  });
} else {
  const { foo } = workerData;
  console.log(`worker: get an object${foo} and sleep 5s in Python`);
  foo.sleep(); // this is a blocking function which is implemented at Python to sleep 1s
  
  console.log('python sleep is done, and sleep in nodejs(thread)');
  setTimeout(() => parentPort.postMessage('done'), 1000);
}
```

--------------------------------

### Import Python Modules with Boa

Source: https://context7.com/alibaba/pipcook/llms.txt

Demonstrates importing and using Python modules like 'os', 'numpy', and 're' directly within Node.js using the Boa Python bridge. Ensure '@pipcook/boa' is installed.

```javascript
const boa = require('@pipcook/boa');

// Import Python system modules
const os = boa.import('os');
console.log(os.getpid()); // prints the pid from python

// Import third-party Python packages
const np = boa.import('numpy');
const tf = boa.import('tensorflow');

// Create numpy arrays
const arr = np.array([[1, 2, 3], [4, 5, 6]], np.int32);
console.log(arr[0]); // 1

// Reshape arrays
const x1 = np.arange(15).reshape(3, 5);
console.log(x1.shape); // (3, 5)

// Use regex module
const re = boa.import('re');
const m = re.search('(?<=abc)def', 'abcdef');
console.log(m.group(0)); // 'def'
```

--------------------------------

### Run Node.js with experimental loaders

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/manual/intro-to-boa.md

Commands to execute Node.js applications using the Pipcook Boa ESM loader.

```bash
$ node --experimental-loader @pipcook/boa/esm/loader.mjs app.mjs
```

```bash
$ node --experimental-modules --experimental-loader @pipcook/boa/esm/loader.mjs app.mjs
```

--------------------------------

### Clone the Pipcook repository

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/guide-to-contributor.md

Initial step to obtain the source code from GitHub.

```bash
$ git clone git@github.com:alibaba/pipcook.git
```

--------------------------------

### Dataset Directory Structure for PascalVOC

Source: https://context7.com/alibaba/pipcook/llms.txt

Illustrates the expected directory layout for datasets using the PascalVOC format, including separate folders for annotations and images.

```tree
dataset/
├── annotations/
│   ├── train/
│   │   └── image_name.xml
│   ├── validation/
│   └── test/
└── images/
    └── image_name.jpg
```

--------------------------------

### Text Dataset CSV Format Example

Source: https://github.com/alibaba/pipcook/blob/main/docs/spec/dataset.md

This shows the standard CSV format for text datasets in Pipcook. The first column contains the text content, and the second column specifies the category, with ',' as the delimiter and no header row.

```csv
name, category
prod1, type1
prod2, type2
prod3, type2
prod4, type1
```

--------------------------------

### Scaffold a new Pipcook script project

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/contribute-a-script.md

Use the npm init command to generate a template project structure for Pipcook scripts.

```sh
$ npm init pipcook-script all my-scripts js && cd my-scripts
$ tree -L 1
.
├── LICENSE
├── README.md
├── debug
├─ node_modules
├── package-lock.json
├── package.json
├── src
└─ webpack.config.js
$ tree ./src
./src
├── dataflow.js
├── datasource.js
├── index.js
└── model.js
```

--------------------------------

### Predict Button Click Handler

Source: https://github.com/alibaba/pipcook/blob/main/packages/cli/serve-resource/text/index.html

This JavaScript code handles the prediction request when the 'Predict' button is clicked. It retrieves the input text, sends a GET request to the 'predict' endpoint with the input as a query parameter, and displays the server's response in the element with the ID 'result'. Ensure jQuery is included for this functionality.

```javascript
$(document).ready(function () {
    document.getElementById('input_text').focus();
    $('#predict').click(function (event) {
        event.preventDefault();
        var input = document.getElementById('input_text').value;
        $.get({
            url: 'predict?input=' + input,
            cache: false,
            contentType: false,
            processData: false,
            success: function (resp) {
                $('#result').html(JSON.stringify(resp, null, ' '));
            },
        });
    });
});
```

--------------------------------

### 运行 Pipcook 训练任务

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/faq/pipcook-framework.md

使用 pipcook run 命令启动训练，支持本地路径或远程 URL 作为配置文件。

```bash
pipcook run <url>
```

--------------------------------

### Train Machine Learning Models

Source: https://context7.com/alibaba/pipcook/llms.txt

Train models using pipeline configuration files with support for remote URLs, local files, and custom training options.

```bash
# Train image classification model with MobileNet
pipcook train https://cdn.jsdelivr.net/gh/alibaba/pipcook@main/example/pipelines/image-classification-mobilenet.json -o ./output

# Train from local pipeline file
pipcook run /path/to/your/pipeline-config.json -o ./my-model

# Train with custom options
pipcook train pipeline.json \
  --output ./classification-output \
  --npmClient npm \
  --registry https://registry.npmjs.org \
  --debug

# Expected output:
# Epoch 1/15
# 187/187 [==============================] - 12s 65ms/step - loss: 0.0604 - accuracy: 0.9823
# ...
# Epoch 15/15
# 187/187 [==============================] - 11s 61ms/step - loss: 5.1657e-05 - accuracy: 1.0000
```

--------------------------------

### Train a Pipeline

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/README.md

Execute a machine learning pipeline using a JSON configuration file.

```sh
$ pipcook train https://cdn.jsdelivr.net/gh/alibaba/pipcook@main/example/pipelines/image-classification-mobilenet.json -o output
```

--------------------------------

### Run a pipeline

Source: https://github.com/alibaba/pipcook/blob/main/docs/contributing/guide-to-contributor.md

Execute a specific pipeline using the provided tool script.

```bash
$ sh tools/run_pipeline.sh <pipeline_name>
```

--------------------------------

### Configure Training Options for Image Classification

Source: https://github.com/alibaba/pipcook/blob/main/example/pipelines/README.md

Set training options such as the framework, GPU usage, and number of epochs. GPU is enabled by default.

```json
{
  "options": {
    "framework": "tfjs@3.8",
    "gpu": false,
    "train": {
      "epochs": 10
    }
  }
}
```

--------------------------------

### Run All Pipcook Tests

Source: https://github.com/alibaba/pipcook/blob/main/docs/zh-cn/contributing/guide-to-contributor.md

Execute all tests for the Pipcook project.

```bash
$ npm test
```

--------------------------------

### Train a Model

Source: https://github.com/alibaba/pipcook/blob/main/docs/README.md

Train a model using a specified pipeline configuration file.

```shell
$ pipcook train https://cdn.jsdelivr.net/gh/alibaba/pipcook@main/example/pipelines/image-classification-mobilenet.json -o ./output
```

--------------------------------

### Manage Context with with()

Source: https://github.com/alibaba/pipcook/blob/main/docs/tutorials/using-python-functions-in-nodejs.md

Uses the with-statement to manage local context, ensuring resources are properly entered and exited.

```python
with(localcontext()) {  
  # balabala  
}
```