1 | # `alwaysai` [![Build Status](https://travis-ci.com/alwaysai/cli.svg?branch=master)](https://travis-ci.com/alwaysai/cli)
|
2 | The [alwaysAI](https://alwaysai.co) command-line interface (CLI)
|
3 |
|
4 | ## Development strategy
|
5 | TODO Talk up front about the different development strategies because that would influence where the developer does things???
|
6 |
|
7 | - On the board
|
8 | - Host + ssh to board
|
9 | - Host + container
|
10 |
|
11 | ## Prerequisites
|
12 | [Node.js](https://en.wikipedia.org/wiki/Node.js) is an open-source, cross-platform JavaScript run-time environment.
|
13 | Install Node.js using:
|
14 |
|
15 | - For Mac/Linux [nvm](https://github.com/nvm-sh/nvm#installation-and-update)
|
16 | - For Windows [nvm-windows](https://github.com/coreybutler/nvm-windows)
|
17 |
|
18 | Now open up a *new* terminal and do:
|
19 | ```
|
20 | $ node --version
|
21 | ```
|
22 | It should print something like "v12.2.0".
|
23 |
|
24 |
|
25 | ## Install the `alwaysai` CLI
|
26 | Use the `npm` package manager, which ships with Node.js, to install the alwaysAI CLI:
|
27 | ```
|
28 | $ npm install --global alwaysai
|
29 | ```
|
30 |
|
31 | If you installed Node.js using the installer from [nodejs.org](https://nodejs.org/en/) instead of using `nvm`, you may need to do:
|
32 | ```
|
33 | $ sudo npm install --global alwaysai
|
34 | ```
|
35 | and enter your OS password.
|
36 |
|
37 | Verify that the CLI has been installed by running:
|
38 | ```
|
39 | $ alwaysai version
|
40 | ```
|
41 | It should print e.g. `0.2.10`.
|
42 |
|
43 | ### Log in
|
44 | [Sign up](https://alwaysai.co/signup) for an alwaysAI account and complete the first-time login flow throw the web. Now authenticate yourself with the CLI:
|
45 | ```
|
46 | $ alwaysai user login
|
47 | ```
|
48 | You'll be prompted to enter your username and password. You're ready to start developing your first alwaysAI deep-learning computer vision application!
|
49 |
|
50 | ### Initialize your application
|
51 | An alwaysAI application is a versioned collection of models, configuration files, and python code. In a terminal do:
|
52 |
|
53 | ```
|
54 | $ mkdir my-app
|
55 | $ cd foo
|
56 | $ alwaysai app init
|
57 | ```
|
58 | The `app init` command prompts you to configure your current directory as alwaysAI application. It creates the application configuration file `alwaysai.app.json` as well as a stub application `app.py` and `requirements.txt` if there isn't one already.
|
59 |
|
60 | ### Find a model
|
61 | The alwaysAI model catalog is a collection of computer vision (CV) models that you can use in your application:
|
62 | ```
|
63 | $ alwaysai app models search mobilenet
|
64 | @alwaysai/MobileNetSSD
|
65 | ```
|
66 |
|
67 | ```py
|
68 | TODO PYTHON
|
69 | ```
|
70 |
|
71 | ### Add a model to your application configuration
|
72 | You can add a model to your application using the command-line interface like so:
|
73 | ```
|
74 | $ alwaysai app models add alwaysai/MobileNetSSD
|
75 | Added alwaysai/MobileNetSSD
|
76 | ```
|
77 | Now you'll see in your application config file:
|
78 |
|
79 | ```json
|
80 | {
|
81 | "id": "alice/my-app",
|
82 | "version": "0.0.0",
|
83 | "models": {
|
84 | "alwaysai/MobileNetSSD": "4.0.8"
|
85 | },
|
86 | }
|
87 | ```
|
88 |
|
89 | Models get pulled down from the alwaysAI cloud by the `app install` and/or `app target install` commands...
|
90 |
|
91 |
|
92 | ### Install the application
|
93 | ```
|
94 | $ alwaysai app install
|
95 | TODO
|
96 | ```
|
97 |
|
98 | ## More information
|
99 | If you encounter any bugs or have any questions or feature requests, please don't hesitate to file an issue or submit a pull request on this project's repository on GitHub.
|
100 |
|
101 | ## Related
|
102 |
|
103 | ## License
|
104 | MIT © [alwaysAI, Inc.](https://alwaysai.co)
|