mirror of
https://github.com/LCTT/TranslateProject.git
synced 2025-01-16 22:42:21 +08:00
Merge remote-tracking branch 'LCTT/master'
This commit is contained in:
Xingyu.Wang
commit
7fd4b81049
@ -1,39 +1,40 @@
|
||||
|
||||
如何在 Linux 中使用 Fio 来测评硬盘性能
|
||||
======
|
||||
|
||||
|
||||
|
||||
Fio（Flexible I/O Tester） 是一款由 Jens Axboe 开发的用于测评和压力/硬件验证的[免费开源][1]的软件
|
||||
Fio（Flexible I/O Tester） 是一款由 Jens Axboe 开发的用于测评和压力/硬件验证的[自由开源][1]的软件。
|
||||
|
||||
它支持 19 种不同类型的 I/O 引擎 (sync, mmap, libaio, posixaio, SG v3, splice, null, network, syslet, guasi, solarisaio, 以及更多), I/O 优先级（针对较新的 Linux 内核），I/O 速度，复刻或线程任务，和其他更多的东西。它能够在块设备和文件上工作。
|
||||
它支持 19 种不同类型的 I/O 引擎 (sync、mmap、libaio、posixaio、SG v3、splice、null、network、 syslet、guasi、solarisaio，以及更多)， I/O 优先级（针对较新的 Linux 内核），I/O 速度，fork 的任务或线程任务等等。它能够在块设备和文件上工作。
|
||||
|
||||
Fio 接受一种非常简单易于理解的文本格式作为任务描述。软件默认包含了许多示例任务文件。 Fio 展示了所有类型的 I/O 性能信息，包括完整的 IO 延迟和百分比。
|
||||
Fio 接受一种非常简单易于理解的文本格式的任务描述。软件默认包含了几个示例任务文件。 Fio 展示了所有类型的 I/O 性能信息，包括完整的 IO 延迟和百分比。
|
||||
|
||||
它被广泛的应用在非常多的地方，包括测评、QA，以及验证用途。它支持 Linux 、 FreeBSD 、 NetBSD、 OpenBSD、 OS X、 OpenSolaris、 AIX、 HP-UX、 Android 以及 Windows。
|
||||
它被广泛的应用在非常多的地方，包括测评、QA，以及验证用途。它支持 Linux 、FreeBSD 、NetBSD、 OpenBSD、 OS X、 OpenSolaris、 AIX、 HP-UX、 Android 以及 Windows。
|
||||
|
||||
在这个教程，我们将使用 Ubuntu 16 ，你需要拥有这台电脑的 sudo 或 root 权限。我们将完整的进行安装和 Fio 的使用。
|
||||
在这个教程，我们将使用 Ubuntu 16 ，你需要拥有这台电脑的 `sudo` 或 root 权限。我们将完整的进行安装和 Fio 的使用。
|
||||
|
||||
### 使用源码安装 Fio
|
||||
|
||||
我们要去克隆 Github 上的仓库。安装所需的依赖，然后我们将会从源码构建应用。首先，确保我们安装了 Git 。
|
||||
我们要去克隆 GitHub 上的仓库。安装所需的依赖，然后我们将会从源码构建应用。首先，确保我们安装了 Git 。
|
||||
|
||||
```
|
||||
sudo apt-get install git
|
||||
```
|
||||
|
||||
CentOS 用户可以执行下述命令：
|
||||
|
||||
```
|
||||
sudo yum install git
|
||||
```
|
||||
|
||||
现在，我们切换到 /opt 目录，并从 Github 上克隆仓库：
|
||||
现在，我们切换到 `/opt` 目录，并从 Github 上克隆仓库：
|
||||
|
||||
```
|
||||
cd /opt
|
||||
git clone https://github.com/axboe/fio
|
||||
```
|
||||
|
||||
你应该会看到下面这样的输出
|
||||
你应该会看到下面这样的输出：
|
||||
|
||||
```
|
||||
Cloning into 'fio'...
|
||||
@ -45,7 +46,7 @@ Resolving deltas: 100% (16251/16251), done.
|
||||
Checking connectivity... done.
|
||||
```
|
||||
|
||||
现在，我们通过在 opt 目录下输入下方的命令切换到 Fio 的代码目录：
|
||||
现在，我们通过在 `/opt` 目录下输入下方的命令切换到 Fio 的代码目录：
|
||||
|
||||
```
|
||||
cd fio
|
||||
@ -61,7 +62,7 @@ cd fio
|
||||
|
||||
### 在 Ubuntu 上安装 Fio
|
||||
|
||||
对于 Ubuntu 和 Debian 来说， Fio 已经在主仓库内。你可以很容易的使用类似 yum 和 apt-get 的标准包管理器来安装 Fio。
|
||||
对于 Ubuntu 和 Debian 来说， Fio 已经在主仓库内。你可以很容易的使用类似 `yum` 和 `apt-get` 的标准包管理器来安装 Fio。
|
||||
|
||||
对于 Ubuntu 和 Debian ，你只需要简单的执行下述命令：
|
||||
|
||||
@ -69,7 +70,8 @@ cd fio
|
||||
sudo apt-get install fio
|
||||
```
|
||||
|
||||
对于 CentOS/Redhat 你只需要简单执行下述命令：
|
||||
对于 CentOS/Redhat 你只需要简单执行下述命令。
|
||||
|
||||
在 CentOS ，你可能在你能安装 Fio 前需要去安装 EPEL 仓库到你的系统中。你可以通过执行下述命令来安装它：
|
||||
|
||||
```
|
||||
@ -124,23 +126,20 @@ Run status group 0 (all jobs):
|
||||
|
||||
Disk stats (read/write):
|
||||
sda: ios=0/0, merge=0/0, ticks=0/0, in_queue=0, util=0.00%
|
||||
|
||||
|
||||
```
|
||||
|
||||
### 执行随机读测试
|
||||
|
||||
我们将要执行一个随机读测试，我们将会尝试读取一个随机的 2GB 文件。
|
||||
|
||||
```
|
||||
|
||||
sudo fio --name=randread --ioengine=libaio --iodepth=16 --rw=randread --bs=4k --direct=0 --size=512M --numjobs=4 --runtime=240 --group_reporting
|
||||
|
||||
|
||||
```
|
||||
|
||||
你应该会看到下面这样的输出
|
||||
```
|
||||
你应该会看到下面这样的输出：
|
||||
|
||||
```
|
||||
...
|
||||
fio-2.2.10
|
||||
Starting 4 processes
|
||||
@ -176,15 +175,13 @@ Run status group 0 (all jobs):
|
||||
|
||||
Disk stats (read/write):
|
||||
sda: ios=521587/871, merge=0/1142, ticks=96664/612, in_queue=97284, util=99.85%
|
||||
|
||||
|
||||
```
|
||||
|
||||
最后，我们想要展示一个简单的随机读-写测试来看一看 Fio 返回的输出类型。
|
||||
|
||||
### 读写性能测试
|
||||
|
||||
下述命令将会测试 USB Pen 驱动器 (/dev/sdc1) 的随机读写性能：
|
||||
下述命令将会测试 USB Pen 驱动器 (`/dev/sdc1`) 的随机读写性能：
|
||||
|
||||
```
|
||||
sudo fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=random_read_write.fio --bs=4k --iodepth=64 --size=4G --readwrite=randrw --rwmixread=75
|
||||
@ -213,8 +210,6 @@ Disk stats (read/write):
|
||||
sda: ios=774141/258944, merge=1463/899, ticks=748800/150316, in_queue=900720, util=99.35%
|
||||
```
|
||||
|
||||
We hope you enjoyed this tutorial and enjoyed following along, Fio is a very useful tool and we hope you can use it in your next debugging activity. If you enjoyed reading this post feel free to leave a comment of questions. Go ahead and clone the repo and play around with the code.
|
||||
|
||||
我们希望你能喜欢这个教程并且享受接下来的内容，Fio 是一个非常有用的工具，并且我们希望你能在你下一次 Debugging 活动中使用到它。如果你喜欢这个文章，欢迎留下评论和问题。
|
||||
|
||||
|
||||
@ -224,7 +219,7 @@ via: https://wpmojo.com/how-to-use-fio-to-measure-disk-performance-in-linux/
|
||||
|
||||
作者：[Alex Pearson][a]
|
||||
译者：[Bestony](https://github.com/bestony)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
校对：[wxy](https://github.com/wxy)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
@ -1,34 +1,34 @@
|
||||
比特币是一个邪教 — Adam Caudill
|
||||
比特币是一个邪教
|
||||
======
|
||||
经过这些年，比特币社区已经发生了非常大的变化；社区成员从闭着眼睛都能讲解 [Merkle 树][1] 的技术迷们，变成了被一夜爆富欲望驱使的投机者和由一些连什么是 Merkle 树都不懂的人所领导的企图寻求 10 亿美元估值的区块链初创公司。随着时间的流逝，围绕比特币和其它加密货币形成了一个狂热，他们认为比特币和其它加密货币远比实际的更重要；他们相信常见的货币（法定货币）正在成为过去，而加密货币将从根本上改变世界经济。
|
||||
|
||||
每一年他们的队伍都在壮大，而他们对加密货币的看法也在变得更加宏伟，那怕是因为[使用新技术][2]而使它陷入困境的情况下。虽然我坚信设计优良的加密货币可以使金钱的跨境流动更容易，并且在大规模通胀的领域提供一个更稳定的选择，但现实情况是，我们并没有做到这些。实际上，正是价值的巨大不稳定性才使得投机者赚钱。那些宣扬美元和欧元即将死去的人，已经完全抛弃了对现实世界客观公正的看法。
|
||||
经过这些年，比特币社区已经发生了非常大的变化；社区成员从闭着眼睛都能讲解 [梅克尔树][1] 的技术迷们，变成了被一夜爆富欲望驱使的投机者和由一些连什么是梅克尔树都不懂的人所领导的企图寻求 10 亿美元估值的区块链初创公司。随着时间的流逝，围绕比特币和其它加密货币形成了一股热潮，他们认为比特币和其它加密货币远比实际的更重要；他们相信常见的货币（法定货币）正在成为过去，而加密货币将从根本上改变世界经济。
|
||||
|
||||
每一年他们的队伍都在壮大，而他们对加密货币的看法也在变得更加宏伟，那怕是对该技术的[新奇的用法][2]而使它陷入了困境。虽然我坚信设计优良的加密货币可以使金钱的跨境流动更容易，并且在大规模通胀的领域提供一个更稳定的选择，但现实情况是，我们并没有做到这些。实际上，正是价值的巨大不稳定性才使得投机者赚钱。那些宣扬美元和欧元即将死去的人，已经完全抛弃了对现实世界客观公正的看法。
|
||||
|
||||
### 一点点背景 …
|
||||
|
||||
比特币发行那天，我读了它的白皮书 —— 它使用有趣的 [Merkle 树][1] 去创建一个公共账簿和一个非常合理的共识协议 —— 由于它新颖的特性引起了密码学领域中许多人的注意。在白皮书发布后的几年里，比特币变得非常有价值，并由此吸引了许多人将它视为是一种投资，和那些认为它将改变一切的忠实追随者（和发声者）。这篇文章将讨论的正是后者。
|
||||
比特币发行那天，我读了它的白皮书 —— 它使用有趣的 [梅克尔树][1] 去创建一个公共账簿和一个非常合理的共识协议 —— 由于它新颖的特性引起了密码学领域中许多人的注意。在白皮书发布后的几年里，比特币变得非常有价值，并由此吸引了许多人将它视为是一种投资，和那些认为它将改变一切的忠实追随者（和发声者）。这篇文章将讨论的正是后者。
|
||||
|
||||
昨天，有人在推特上发布了一个最近的比特币区块的哈希，下面成千上万的推文和其它讨论让我相信，比特币已经跨越界线进入了真正的邪教领域。
|
||||
昨天（2018/6/20），有人在推特上发布了一个最近的比特币区块的哈希，下面成千上万的推文和其它讨论让我相信，比特币已经跨越界线进入了真正的邪教领域。
|
||||
|
||||
一切都源于 Mark Wilcox 的这个推文：
|
||||
一切都源于 Mark Wilcox 的[这个推文][9]：
|
||||
|
||||
> #00000000000000000021e800c1e8df51b22c1588e5a624bea17e9faa34b2dc4a
|
||||
> — Mark Wilcox (@mwilcox) June 19, 2018
|
||||
> [#00000000000000000021e800c1e8df51b22c1588e5a624bea17e9faa34b2dc4a][8]
|
||||
|
||||
张贴的这个值是 [比特币 #528249 号区块][3] 的哈希值。前导零是挖矿过程的结果；挖掘一个区块就是把区块内容与一个 nonce（和其它数据）组合起来，然后做哈希运算，并且它至少有一定数量的前导零才能被验证为有效区块。如果它不是正确的数字，你可以更换 nonce 再试。重复这个过程直到哈希值的前导零数量是正确的数字之后，你就有了一个有效的区块。让人们感到很兴奋的部分是接下来的 21e800。
|
||||
> — Mark Wilcox (@mwilcox) [June 19, 2018][9]
|
||||
|
||||
张贴的这个值是 [比特币 #528249 号区块][3] 的哈希值。前导零是挖矿过程的结果；挖掘一个区块就是把区块内容与一个<ruby>现时数<rt>nonce</rt></ruby>（和其它数据）组合起来，然后做哈希运算，并且它至少有一定数量的前导零才能被验证为有效区块。如果它不是正确的数字，你可以更换现时数再试。重复这个过程直到哈希值的前导零数量是正确的数字之后，你就有了一个有效的区块。让人们感到很兴奋的部分是接下来的 `21e800`。
|
||||
|
||||
一些人说这是一个有意义的编号，挖掘出这个区块的人实际上的难度远远超出当前所看到的，不仅要调整前导零的数量，还要匹配接下来的 24 位 —— 它要求非常强大的计算能力。如果有人能够以蛮力去实现它，这将表明有些事情很严重，比如，在计算或密码学方面的重大突破。
|
||||
|
||||
你一定会有疑问，为什么 21e800 如此重要 —— 一个你问了肯定会后悔的问题。有人说它是参考了 [E8 理论][4]（一个广受批评的提出标准场理论的论文），或是表示总共存在 2100000000 枚比特币（`21 x 10^8` 就是 2,100,000,000）。还有其它说法，因为太疯狂了而没有办法写出来。另一个重要的事实是，在前导零后面有 21e8 的区块平均每年被挖掘出一次 —— 这些从来没有人认为是很重要的。
|
||||
你一定会有疑问，为什么 `21e800` 如此重要 —— 一个你问了肯定会后悔的问题。有人说它是参考了 [E8 理论][4]（一个广受批评的提出标准场理论的论文），或是表示总共存在 2,100,000,000 枚比特币（`21 x 10^8` 就是 2,100,000,000）。还有其它说法，因为太疯狂了而没有办法写出来。另一个重要的事实是，在前导零后面有 21e8 的区块平均每年被挖掘出一次 —— 这些从来没有人认为是很重要的。
|
||||
|
||||
这就引出了有趣的地方：关于这是如何发生的[理论][5]。
|
||||
|
||||
* 一台量子计算机，它能以某种方式用不可思议的速度做哈希运算。尽管在量子计算机的理论中还没有迹象表明它能够做这件事。哈希是量子计算机认为很安全的东西之一。
|
||||
* 一台量子计算机，它能以某种方式用不可思议的速度做哈希运算。尽管在量子计算机的理论中还没有迹象表明它能够做这件事。哈希是量子计算机认为安全的东西之一。
|
||||
* 时间旅行。是的，真的有人这么说，有人从未来穿梭回到现在去挖掘这个区块。我认为这种说法太荒谬了，都懒得去解释它为什么是错误的。
|
||||
* 中本聪回来了。尽管事实上他的私钥没有任何活动，一些人从理论上认为他回来了，他能做一些没人能做的事情。这些理论是无法解释他如何做到的。
|
||||
|
||||
|
||||
|
||||
> 因此，总的来说（按我的理解）中本聪，为了知道和计算他做的事情，根据现代科学，他可能是以下之一：
|
||||
>
|
||||
> A) 使用了一台量子计算机
|
||||
@ -37,39 +37,35 @@
|
||||
>
|
||||
> — Crypto Randy Marsh [REKT] (@nondualrandy) [June 21, 2018][6]
|
||||
|
||||
如果你觉得所有的这一切听起来像 [命理学][7]，不止你一个人是这样想的。
|
||||
如果你觉得所有的这一切听起来像 <ruby>[命理学][7]<rt>numerology</rt></ruby>，不止你一个人是这样想的。
|
||||
|
||||
所有围绕有特殊意义的区块哈希的讨论，也引发了对在某种程度上比较有趣的东西的讨论。比特币的创世区块，它是第一个比特币区块，有一个不寻常的属性：早期的比特币要求哈希值的前 32 位是零；而创始区块的前导零有 43 位。因为由代码产生的创世区块从不会发布，它不知道它是如何产生的，也不知道是用什么类型的硬件产生的。中本聪有学术背景，因此可能他有比那个时候大学中常见设备更强大的计算能力。从这一点上说，只是对古怪的创世区块的历史有点好奇，仅此而已。
|
||||
所有围绕有特殊意义的区块哈希的讨论，也引发了对在某种程度上比较有趣的东西的讨论。比特币的创世区块，它是第一个比特币区块，有一个不寻常的属性：早期的比特币要求哈希值的前 32 <ruby>位<rt>bit</rt></ruby>是零；而创始区块的前导零有 43 位。因为产生创世区块的代码从未发布过，不知道它是如何产生的，也不知道是用什么类型的硬件产生的。中本聪有学术背景，因此可能他有比那个时候大学中常见设备更强大的计算能力。从这一点上说，只是对古怪的创世区块的历史有点好奇，仅此而已。
|
||||
|
||||
### 关于哈希运算的简单题外话
|
||||
|
||||
这种喧嚣始于比特币区块的哈希运算；因此理解哈希是什么很重要，并且要理解一个非常重要的属性，一个哈希是单向加密函数，它能够基于给定的数据创建一个伪随机输出。
|
||||
这种喧嚣始于比特币区块的哈希运算；因此理解哈希是什么很重要，并且要理解一个非常重要的属性，哈希是单向加密函数，它能够基于给定的数据创建一个伪随机输出。
|
||||
|
||||
这意味着什么呢？基于本文讨论的目的，对于每个给定的输入你将得到一个随机的输出。随机数有时看起来很有趣，很简单，因为它是随机的结果，并且人类大脑可以很容易从任何东西中找到顺序。当你从随机数据中开始查看顺序时，你就会发现有趣的事情 —— 这些东西毫无意义，因为它们只是简单地随机数。当人们把重要的意义归属到随机数据上时，它将告诉你很多这些参与者观念相关的东西，而不是数据本身。
|
||||
|
||||
### 币的邪教
|
||||
### 币之邪教
|
||||
|
||||
首先，我们来定义一组术语：
|
||||
|
||||
* 邪教：一个宗教崇拜和直接向一个特定的人或物虔诚的体系。
|
||||
* 宗教：有人认为是至高无上的追求或兴趣。
|
||||
* <ruby>邪教<rt>Cult</rt></ruby>：一个宗教崇拜和直接向一个特定的人或物虔诚的体系。
|
||||
* <ruby>宗教<rt>Religion</rt></ruby>：有人认为是至高无上的追求或兴趣。
|
||||
|
||||
<ruby>币之邪教<rt>Cult of the Coin</rt></ruby>有许多圣人，或许没有人比<ruby>中本聪<rt>Satoshi Nakamoto</rt></ruby>更伟大，他是比特币创始者（们）的假名。（对他的）狂热拥戴，要归因于他的能力和理解力远超过一般的研究人员，认为他的远见卓视无人能比，他影响了世界新经济的秩序。当将中本聪的神秘本质和未知的真实身份结合起来时，狂热的追随着们将中本聪视为一个真正值得尊敬的人物。
|
||||
|
||||
当然，除了追随其他圣人的追捧者之外，毫无疑问这些追捧者认为自己是正确的。任何对他们的圣人的批评都被认为也是对他们的批评。例如，那些追捧 EOS 的人，可能会视中本聪为一个开发了失败项目的黑客，而对 EOS 那怕是最轻微的批评，他们也会作出激烈的反应，之所以反应如此强烈，仅仅是因为攻击了他们心目中的神。那些追捧 IOTA 的人的反应也一样；还有更多这样的例子。
|
||||
|
||||
币的狂热追捧者中的许多圣人，或许没有人比中本聪更伟大，他是比特币创始人的假名。强力的护卫、赋予能力和理解力远超过一般的研究人员，认为他的远见卓视无人能比，他影响了世界新经济的秩序。当将中本聪的神秘本质和未知的真实身份结合起来时，狂热的追随着们将中本聪视为一个真正的值的尊敬的人物。
|
||||
|
||||
当然，除了追随其他圣人的追捧者之外，毫无疑问这些追捧者认为自己是正确的。任何对他们的圣人的批评都被认为也是对他们的批评。例如，那些追捧 EOS 的人，可能会认为中本聪是开发了一个失败项目的黑客，而对 EOS 那怕是最轻微的批评，他们也会作出激烈的反应，之所以反应如此强烈，仅仅是因为攻击了他们心目中的神。那些追捧 IOTA 的人的反应也一样；还有更多这样的例子。
|
||||
|
||||
这些追随着在讨论问题时已经失去了理性和客观，他们的狂热遮盖了他们的视野。任何对这些项目和项目背后的人的讨论，如果不是溢美之词，必然以某种程序的刻薄言辞结束，对于一个技术的讨论那种做法是毫无道理的。
|
||||
这些追随者在讨论问题时已经失去了理性和客观，他们的狂热遮盖了他们的视野。任何对这些项目和项目背后的人的讨论，如果不是溢美之词，必然以某种程序的刻薄言辞结束，对于一个技术的讨论那种做法是毫无道理的。
|
||||
|
||||
这很危险，原因很多：
|
||||
|
||||
* 开发者 & 研究者对缺陷视而不见。由于追捧者的大量赞美，这些参与开发的人对自己的能力开始膨胀，并将一些批评看作是无端的攻击 —— 因为他们认为自己是不可能错的。
|
||||
* 开发者 & 研究者对缺陷视而不见。由于追捧者的大量赞美，这些参与开发的人对自己的能力的看法开始膨胀，并将一些批评看作是无端的攻击 —— 因为他们认为自己是不可能错的。
|
||||
* 真正的问题是被攻击。技术问题不再被看作是需要去解决的问题和改进的机会，他们认为是来自那些想去破坏项目的人的攻击。
|
||||
* 用一枚币来控制他们。追随者们通常会结盟，而圣人仅有一个。承认其它项目的优越，意味着认同自己项目的缺陷或不足，而这是他们不愿意做的事情。
|
||||
* 阻止真实的进步。进化是很残酷的，它要求死亡，项目失败，以及承认这些失败的原因。如果忽视失败的教训，如果不允许那些应该去死亡的事情发生，进步就会停止。
|
||||
|
||||
|
||||
* 物以类聚，人以币分。追随者们通常会结盟到一起，而圣人仅有一个。承认其它项目的优越，意味着认同自己项目的缺陷或不足，而这是他们不愿意做的事情。
|
||||
* 阻止真实的进步。进化是很残酷的，死亡是必然会有的，项目可能失败，也要承认这些失败的原因。如果忽视失败的教训，如果不允许那些应该去死亡的事情发生，进步就会停止。
|
||||
|
||||
许多围绕加密货币和相关区块链项目的讨论已经开始变得越来越”有毒“，善意的人想在不受攻击的情况下进行技术性讨论越来越不可能。随着对真正缺陷的讨论，那些在其它环境中注定要失败的缺陷，在没有做任何的事实分析的情况下即刻被判定为异端已经成为了惯例，善意的人参与其中的代价变得极其昂贵。至少有些人已经意识到极其严重的安全漏洞，由于高“毒性”的环境，他们选择保持沉默。
|
||||
|
||||
@ -79,7 +75,10 @@
|
||||
|
||||
[注意：这种行为有许多例子可以引用，但是为了保护那些因批评项目而成为被攻击目标的人，我选择尽可能少的列出这种例子。我看到许多我很尊敬的人、许多我认为是朋友的人成为这种恶毒攻击的受害者 —— 我不想引起人们对这些攻击的注意和重新引起对他们的攻击。]
|
||||
|
||||
---
|
||||
关于作者：
|
||||
|
||||
我是一个资深应用安全顾问、研究员和具有超过 15 年的经验的软件开发者。我主要关注的是应用程序安全、安全通信和加密, 虽然我经常由于无聊而去研究新的领域。我通常会写了一些关于我的研究和安全、开发和软件设计，和我当前吸引了我注意力的爱好的文章。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
@ -88,7 +87,7 @@ via: https://adamcaudill.com/2018/06/21/bitcoin-is-a-cult/
|
||||
作者：[Adam Caudill][a]
|
||||
选题：[lujun9972](https://github.com/lujun9972)
|
||||
译者：[qhwdw](https://github.com/qhwdw)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
校对：[wxy](https://github.com/wxy)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
@ -100,3 +99,5 @@ via: https://adamcaudill.com/2018/06/21/bitcoin-is-a-cult/
|
||||
[5]:https://medium.com/@coop__soup/00000000000000000021e800c1e8df51b22c1588e5a624bea17e9faa34b2dc4a-cd4b67d446be
|
||||
[6]:https://twitter.com/nondualrandy/status/1009609117768605696?ref_src=twsrc%5Etfw
|
||||
[7]:https://en.wikipedia.org/wiki/Numerology
|
||||
[8]:https://twitter.com/hashtag/00000000000000000021e800c1e8df51b22c1588e5a624bea17e9faa34b2dc4a?src=hash&ref_src=twsrc%5Etfw
|
||||
[9]:https://twitter.com/mwilcox/status/1009160832398262273?ref_src=twsrc%5Etfw
|
||||
@ -1,4 +1,4 @@
|
||||
20 questions DevOps job candidates should be prepared to answer
|
||||
Translating by FelixYFZ 20 questions DevOps job candidates should be prepared to answer
|
||||
======
|
||||
|
||||
|
||||
|
||||
@ -0,0 +1,68 @@
|
||||
3 pitfalls everyone should avoid with hybrid multi-cloud, part 2
|
||||
======
|
||||
|
||||
|
||||
|
||||
This article was co-written with [Roel Hodzelmans][1].
|
||||
|
||||
Cloud hype is all around you—you're told it's critical to ensuring a digital future for your business. Whether you choose cloud, hybrid cloud, or hybrid multi-cloud, you have numerous decisions to make, even as you continue the daily work of enhancing your customers' experience and agile delivery of your applications (including legacy applications)—likely some of your business' most important resources.
|
||||
|
||||
In this series, we explain three pitfalls everyone should avoid when transitioning to hybrid multi-cloud environments. [In part one][2], we defined the different cloud types and explained the differences between hybrid cloud and multi-cloud. Here, in part two, we will dive into the first pitfall: Why cost is not always the best motivator for moving to the cloud.
|
||||
|
||||
### Why not?
|
||||
|
||||
When looking at hybrid or multi-cloud strategies for your business, don't let cost become the obvious motivator. There are a few other aspects of any migration strategy that you should review when putting your plan together. But often budget rules the conversations.
|
||||
|
||||
When giving this talk three times at conferences, we've asked our audience to answer a live, online questionnaire about their company, customers, and experiences in the field. Over 73% of respondents said cost was the driving factor in their business' decision to move to hybrid or multi-cloud.
|
||||
|
||||
But, if you already have full control of your on-premises data centers, yet perpetually underutilize and overpay for resources, how can you expect to prevent those costs from rolling over into your cloud strategy?
|
||||
|
||||
There are three main (and often forgotten, ignored, and unaccounted for) reasons cost shouldn't be the primary motivating factor for migrating to the cloud: labor costs, overcapacity, and overpaying for resources. They are important points to consider when developing a hybrid or multi-cloud strategy.
|
||||
|
||||
### Labor costs
|
||||
|
||||
Imagine a utility company making the strategic decision to move everything to the cloud within the next three years. The company kicks off enthusiastically, envisioning huge cost savings, but soon runs into labor cost issues that threaten to blow up the budget.
|
||||
|
||||
One of the most overlooked aspects of moving to the cloud is the cost of labor to migrate existing applications and data. A Forrester study reports that labor costs can consume [over 50% of the total cost of a public cloud migration][3]. Forrester says, "customer-facing apps for systems of engagement… typically employ lots of new code rather than migrating existing code to cloud platforms."
|
||||
|
||||
Step back and analyze what's essential to your customer success and move only that to the cloud. Then, evaluate all your non-essential applications and, over time, consider moving them to commercial, off-the-shelf solutions that require little labor cost.
|
||||
|
||||
### Overcapacity
|
||||
|
||||
"More than 80% of in-house data centers have [way more server capacity than is necessary][4]," reports Business Insider. This amazing bit of information should shock you to your core.
|
||||
|
||||
What exactly is "way more" in this context?
|
||||
|
||||
One hint comes from Deutsche Bank CTO Pat Healey, presenting at Red Hat Summit 2017. He talks about ordering hardware for the financial institution's on-premises data center, only to find out later that [usage numbers were in the single digits][5].
|
||||
|
||||
Healey is not alone; many companies have these problems. They don't do routine assessments, such as checking electricity, cooling, licensing, and other factors, to see how much capacity they are using on a consistent basis.
|
||||
|
||||
### Overpaying
|
||||
|
||||
Companies are paying an average of 36% more for cloud services than they need to, according to the Business Insider article mentioned above.
|
||||
|
||||
One reason is that public cloud providers enthusiastically support customers coming agnostically into their cloud. As customers leverage more of the platform's cloud-native features, they reach a monetary threshold, and technical support drops off dramatically.
|
||||
|
||||
It's a classic case of vendor lock-in, where the public cloud provider knows it is cost-prohibitive for the customer to migrate off its cloud, so it doesn't feel compelled to provide better service.
|
||||
|
||||
### Coming up
|
||||
|
||||
In part three of this series, we'll discuss the second of three pitfalls that everyone should avoid with hybrid multi-cloud. Stay tuned to learn why you should take care with moving everything to the cloud.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/6/reasons-move-to-cloud
|
||||
|
||||
作者：[Eric D.Schabell][a]
|
||||
选题：[lujun9972](https://github.com/lujun9972)
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/eschabell
|
||||
[1]:https://opensource.com/users/roelh
|
||||
[2]:https://opensource.com/article/18/4/pitfalls-hybrid-multi-cloud
|
||||
[3]:https://www.techrepublic.com/article/labor-costs-can-make-up-50-of-public-cloud-migration-is-it-worth-it/
|
||||
[4]:http://www.businessinsider.com/companies-waste-62-billion-on-the-cloud-by-paying-for-storage-they-dont-need-according-to-a-report-2017-11
|
||||
[5]:https://youtu.be/SPRUJ5Z-Aew
|
||||
@ -1,3 +1,5 @@
|
||||
Translating by jessie-pang
|
||||
|
||||
Why moving all your workloads to the cloud is a bad idea
|
||||
======
|
||||
|
||||
|
||||
@ -1,3 +1,5 @@
|
||||
Translatin by imquanquan
|
||||
|
||||
Here are some amazing advantages of Go that you don’t hear much about
|
||||
============================================================
|
||||
|
||||
@ -220,4 +222,4 @@ via: https://medium.freecodecamp.org/here-are-some-amazing-advantages-of-go-that
|
||||
[19]:https://golang.org/src/encoding/json/encode.go
|
||||
[20]:https://tour.golang.org/
|
||||
[21]:https://github.com/kirillrogovoy/
|
||||
[22]:https://twitter.com/krogovoy
|
||||
[22]:https://twitter.com/krogovoy
|
||||
|
||||
@ -0,0 +1,87 @@
|
||||
3 pitfalls everyone should avoid with hybrid multicloud
|
||||
======
|
||||
|
||||
|
||||
This article was co-written with [Roel Hodzelmans][1].
|
||||
|
||||
We're all told the cloud is the way to ensure a digital future for our businesses. But which cloud? From cloud to hybrid cloud to hybrid multi-cloud, you need to make choices, and these choices don't preclude the daily work of enhancing your customers' experience or agile delivery of the applications they need.
|
||||
|
||||
This article is the first in a four-part series on avoiding pitfalls in hybrid multi-cloud computing. Let's start by examining multi-cloud, hybrid cloud, and hybrid multi-cloud and what makes them different from one another.
|
||||
|
||||
### Hybrid vs. multi-cloud
|
||||
|
||||
There are many conversations you may be having in your business around moving to the cloud. For example, you may want to take your on-premises computing capacity and turn it into your own private cloud. You may wish to provide developers with a cloud-like experience using the same resources you already have. A more traditional reason for expansion is to use external computing resources to augment those in your own data centers. The latter leads you to the various public cloud providers, as well as to our first definition, multi-cloud.
|
||||
|
||||
#### Multi-cloud
|
||||
|
||||
Multi-cloud means using multiple clouds from multiple providers for multiple tasks.
|
||||
|
||||
![Multi-cloud][3]
|
||||
|
||||
Figure 1. Multi-cloud IT with multiple isolated cloud environments
|
||||
|
||||
Typically, multi-cloud refers to the use of several different public clouds in order to achieve greater flexibility, lower costs, avoid vendor lock-in, or use specific regional cloud providers.
|
||||
|
||||
A challenge of the multi-cloud approach is achieving consistent policies, compliance, and management with different providers involved.
|
||||
|
||||
Multi-cloud is mainly a strategy to expand your business while leveraging multi-vendor cloud solutions and spreading the risk of lock-in. Figure 1 shows the isolated nature of cloud services in this model, without any sort of coordination between the services and business applications. Each is managed separately, and applications are isolated to services found in their environments.
|
||||
|
||||
#### Hybrid cloud
|
||||
|
||||
Hybrid cloud solves issues where isolation and coordination are central to the solution. It is a combination of one or more public and private clouds with at least a degree of workload portability, integration, orchestration, and unified management.
|
||||
|
||||
![Hybrid cloud][5]
|
||||
|
||||
Figure 2. Hybrid clouds may be on or off premises, but must have a degree of interoperability
|
||||
|
||||
The key issue here is that there is an element of interoperability, migration potential, and a connection between tasks running in public clouds and on-premises infrastructure, even if it's not always seamless or otherwise fully implemented.
|
||||
|
||||
If your cloud model is missing portability, integration, orchestration, and management, then it's just a bunch of clouds, not a hybrid cloud.
|
||||
|
||||
The cloud environments in Fig. 2 include at least one private and public cloud. They can be off or on premises, but they have some degree of the following:
|
||||
|
||||
* Interoperability
|
||||
* Application portability
|
||||
* Data portability
|
||||
* Common management
|
||||
|
||||
|
||||
|
||||
As you can probably guess, combining multi-cloud and hybrid cloud results in a hybrid multi-cloud. But what does that look like?
|
||||
|
||||
### Hybrid multi-cloud
|
||||
|
||||
Hybrid multi-cloud pulls together multiple clouds and provides the tools to ensure interoperability between the various services in hybrid and multi-cloud solutions.
|
||||
|
||||
![Hybrid multi-cloud][7]
|
||||
|
||||
Figure 3. Hybrid multi-cloud solutions using open technologies
|
||||
|
||||
Bringing these together can be a serious challenge, but the result ensures better use of resources without isolation in their respective clouds.
|
||||
|
||||
Fig. 3 shows an example of hybrid multi-cloud based on open technologies for interoperability, workload portability, and management.
|
||||
|
||||
### Moving forward: Pitfalls of hybrid multi-cloud
|
||||
|
||||
In part two of this series, we'll look at the first of three pitfalls to avoid with hybrid multi-cloud. Namely, why cost is not always the obvious motivator when determining how to transition your business to the cloud.
|
||||
|
||||
This article is based on "[3 pitfalls everyone should avoid with hybrid multi-cloud][8]," a talk the authors will be giving at [Red Hat Summit 2018][9], which will be held May 8-10 in San Francisco. [Register by May 7][9] to save US$ 500 off of registration. Use discount code **OPEN18** on the payment page to apply the discount.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/4/pitfalls-hybrid-multi-cloud
|
||||
|
||||
作者：[Eric D.Schabell][a]
|
||||
选题：[lujun9972](https://github.com/lujun9972)
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/eschabell
|
||||
[1]:https://opensource.com/users/roelh
|
||||
[3]:https://opensource.com/sites/default/files/u128651/multi-cloud.png (Multi-cloud)
|
||||
[5]:https://opensource.com/sites/default/files/u128651/hybrid-cloud.png (Hybrid cloud)
|
||||
[7]:https://opensource.com/sites/default/files/u128651/hybrid-multicloud.png (Hybrid multi-cloud)
|
||||
[8]:https://agenda.summit.redhat.com/SessionDetail.aspx?id=153892
|
||||
[9]:https://www.redhat.com/en/summit/2018
|
||||
@ -1,218 +0,0 @@
|
||||
BriFuture is translating
|
||||
|
||||
Twitter Sentiment Analysis using NodeJS
|
||||
============================================================
|
||||
|
||||
|
||||
|
||||
|
||||
If you want to know how people feel about something, there is no better place than Twitter. It is a continuous stream of opinion, with around 6,000 new tweets being created every second. The internet is quick to react to events and if you want to be updated with the latest and hottest, Twitter is the place to be.
|
||||
|
||||
Now, we live in an age where data is king and companies put Twitter's data to good use. From gauging the reception of their new products to trying to predict the next market trend, analysis of Twitter data has many uses. Businesses use it to market their product that to the right customers, to gather feedback on their brand and improve or to assess the reasons for the failure of a product or promotional campaign. Not only businesses, many political and economic decisions are made based on observation of people's opinion. Today, I will try and give you a taste of simple [sentiment analysis][1] of tweets to determine whether a tweet is positive, negative or neutral. It won't be as sophisticated as those used by professionals, but nonetheless, it will give you an idea about opinion mining.
|
||||
|
||||
We will be using NodeJs since JavaScript is ubiquitous nowadays and is one of the easiest languages to get started with.
|
||||
|
||||
### Prerequisite:
|
||||
|
||||
* NodeJs and NPM installed
|
||||
|
||||
* A little experience with NodeJs and NPM packages
|
||||
|
||||
* some familiarity with the command line.
|
||||
|
||||
Alright, that's it. Let's get started.
|
||||
|
||||
### Getting Started
|
||||
|
||||
Make a new directory for your project and go inside the directory. Open a terminal (or command line). Go inside the newly created directory and run the `npm init -y` command. This will create a `package.json` in your directory. Now we can install the npm packages we need. We just need to create a new file named `index.js` and then we are all set to start coding.
|
||||
|
||||
### Getting the tweets
|
||||
|
||||
Well, we want to analyze tweets and for that, we need programmatic access to Twitter. For this, we will use the [twit][2] package. So, let's install it with the `npm i twit` command. We also need to register an App through our account to gain access to the Twitter API. Head over to this [link][3], fill in all the details and copy ‘Consumer Key’, ‘Consumer Secret’, ‘Access token’ and ‘Access Token Secret’ from 'Keys and Access Token' tabs in a `.env` file like this:
|
||||
|
||||
```
|
||||
# .env
|
||||
# replace the stars with values you copied
|
||||
CONSUMER_KEY=************
|
||||
CONSUMER_SECRET=************
|
||||
ACCESS_TOKEN=************
|
||||
ACCESS_TOKEN_SECRET=************
|
||||
|
||||
```
|
||||
|
||||
Now, let's begin.
|
||||
|
||||
Open `index.js` in your favorite code editor. We need to install the `dotenv`package to read from `.env` file with the command `npm i dotenv`. Alright, let's create an API instance.
|
||||
|
||||
```
|
||||
const Twit = require('twit');
|
||||
const dotenv = require('dotenv');
|
||||
|
||||
dotenv.config();
|
||||
|
||||
const { CONSUMER_KEY
|
||||
, CONSUMER_SECRET
|
||||
, ACCESS_TOKEN
|
||||
, ACCESS_TOKEN_SECRET
|
||||
} = process.env;
|
||||
|
||||
const config_twitter = {
|
||||
consumer_key: CONSUMER_KEY,
|
||||
consumer_secret: CONSUMER_SECRET,
|
||||
access_token: ACCESS_TOKEN,
|
||||
access_token_secret: ACCESS_TOKEN_SECRET,
|
||||
timeout_ms: 60*1000
|
||||
};
|
||||
|
||||
let api = new Twit(config_twitter);
|
||||
|
||||
```
|
||||
|
||||
Here we have established a connection to the Twitter with the required configuration. But we are not doing anything with it. Let's define a function to get tweets.
|
||||
|
||||
```
|
||||
async function get_tweets(q, count) {
|
||||
let tweets = await api.get('search/tweets', {q, count, tweet_mode: 'extended'});
|
||||
return tweets.data.statuses.map(tweet => tweet.full_text);
|
||||
}
|
||||
|
||||
```
|
||||
|
||||
This is an async function because of the `api.get` the function returns a promise and instead of chaining `then`s, I wanted an easy way to extract the text of the tweets. It accepts two arguments -q and count, `q` being the query or keyword we want to search for and `count` is the number of tweets we want the `api` to return.
|
||||
|
||||
So now we have an easy way to get the full texts from the tweets. But we still have a problem, the text that we will get now may contain some links or may be truncated if it's a retweet. So we will write another function that will extract and return the text of the tweets, even for retweets and remove the links if any.
|
||||
|
||||
```
|
||||
function get_text(tweet) {
|
||||
let txt = tweet.retweeted_status ? tweet.retweeted_status.full_text : tweet.full_text;
|
||||
return txt.split(/ |\n/).filter(v => !v.startsWith('http')).join(' ');
|
||||
}
|
||||
|
||||
async function get_tweets(q, count) {
|
||||
let tweets = await api.get('search/tweets', {q, count, 'tweet_mode': 'extended'});
|
||||
return tweets.data.statuses.map(get_text);
|
||||
}
|
||||
|
||||
```
|
||||
|
||||
So, now we have the text of tweets. Our next step is getting the sentiment from the text. For this, we will use another package from `npm` - [`sentiment`][4]package. Let's install it like the other packages and add to our script.
|
||||
|
||||
```
|
||||
const sentiment = require('sentiment')
|
||||
|
||||
```
|
||||
|
||||
Using `sentiment` is very easy. We will just have to call the `sentiment`function on the text that we want to analyze and it will return us the comparative score of the text. If the score is below 0, it expresses a negative sentiment, a score above 0 is positive and 0, as you may have guessed, is neutral. So based on this, we will print the tweets in different colors - green for positive, red for negative and blue for neutral. For this, we will use the [`colors`][5] package. Let's install it like the other packages and add to our script.
|
||||
|
||||
```
|
||||
const colors = require('colors/safe');
|
||||
|
||||
```
|
||||
|
||||
Alright, now let us bring it all together in a `main` function.
|
||||
|
||||
```
|
||||
async function main() {
|
||||
let keyword = \* define the keyword that you want to search for *\;
|
||||
let count = \* define the count of tweets you want *\;
|
||||
let tweets = await get_tweets(keyword, count);
|
||||
for (tweet of tweets) {
|
||||
let score = sentiment(tweet).comparative;
|
||||
tweet = `${tweet}\n`;
|
||||
if (score > 0) {
|
||||
tweet = colors.green(tweet);
|
||||
} else if (score < 0) {
|
||||
tweet = colors.red(tweet);
|
||||
} else {
|
||||
tweet = colors.blue(tweet);
|
||||
}
|
||||
console.log(tweet);
|
||||
}
|
||||
}
|
||||
|
||||
```
|
||||
|
||||
And finally, execute the `main` function.
|
||||
|
||||
```
|
||||
main();
|
||||
|
||||
```
|
||||
|
||||
There you have it, a short script of analyzing the basic sentiments of a tweet.
|
||||
|
||||
```
|
||||
\\ full script
|
||||
const Twit = require('twit');
|
||||
const dotenv = require('dotenv');
|
||||
const sentiment = require('sentiment');
|
||||
const colors = require('colors/safe');
|
||||
|
||||
dotenv.config();
|
||||
|
||||
const { CONSUMER_KEY
|
||||
, CONSUMER_SECRET
|
||||
, ACCESS_TOKEN
|
||||
, ACCESS_TOKEN_SECRET
|
||||
} = process.env;
|
||||
|
||||
const config_twitter = {
|
||||
consumer_key: CONSUMER_KEY,
|
||||
consumer_secret: CONSUMER_SECRET,
|
||||
access_token: ACCESS_TOKEN,
|
||||
access_token_secret: ACCESS_TOKEN_SECRET,
|
||||
timeout_ms: 60*1000
|
||||
};
|
||||
|
||||
let api = new Twit(config_twitter);
|
||||
|
||||
function get_text(tweet) {
|
||||
let txt = tweet.retweeted_status ? tweet.retweeted_status.full_text : tweet.full_text;
|
||||
return txt.split(/ |\n/).filter(v => !v.startsWith('http')).join(' ');
|
||||
}
|
||||
|
||||
async function get_tweets(q, count) {
|
||||
let tweets = await api.get('search/tweets', {q, count, 'tweet_mode': 'extended'});
|
||||
return tweets.data.statuses.map(get_text);
|
||||
}
|
||||
|
||||
async function main() {
|
||||
let keyword = 'avengers';
|
||||
let count = 100;
|
||||
let tweets = await get_tweets(keyword, count);
|
||||
for (tweet of tweets) {
|
||||
let score = sentiment(tweet).comparative;
|
||||
tweet = `${tweet}\n`;
|
||||
if (score > 0) {
|
||||
tweet = colors.green(tweet);
|
||||
} else if (score < 0) {
|
||||
tweet = colors.red(tweet);
|
||||
} else {
|
||||
tweet = colors.blue(tweet)
|
||||
}
|
||||
console.log(tweet)
|
||||
}
|
||||
}
|
||||
|
||||
main();
|
||||
```
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://boostlog.io/@anshulc95/twitter-sentiment-analysis-using-nodejs-5ad1331247018500491f3b6a
|
||||
|
||||
作者：[Anshul Chauhan][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://boostlog.io/@anshulc95
|
||||
[1]:https://en.wikipedia.org/wiki/Sentiment_analysis
|
||||
[2]:https://github.com/ttezel/twit
|
||||
[3]:https://boostlog.io/@anshulc95/apps.twitter.com
|
||||
[4]:https://www.npmjs.com/package/sentiment
|
||||
[5]:https://www.npmjs.com/package/colors
|
||||
[6]:https://boostlog.io/tags/nodejs
|
||||
[7]:https://boostlog.io/tags/twitter
|
||||
[8]:https://boostlog.io/@anshulc95
|
||||
@ -1,3 +1,5 @@
|
||||
FSSlc translating
|
||||
|
||||
View The Contents Of An Archive Or Compressed File Without Extracting It
|
||||
======
|
||||
|
||||
|
||||
@ -1,3 +1,4 @@
|
||||
Translating by leemeans
|
||||
Setting Up a Timer with systemd in Linux
|
||||
======
|
||||
|
||||
|
||||
@ -0,0 +1,216 @@
|
||||
用 NodeJS 进行 Twitter 情感分析
|
||||
============================================================
|
||||
|
||||
|
||||
|
||||
|
||||
如果你想知道大家对某件事情的看法，Twitter 是最好的地方了。Twitter 是观点持续不断的涌现出来的地方，每秒钟大概有 6000 条新 Twitter 发送出来。因特网上的发展很快，如果你想与时俱进或者跟上潮流，Twitter 就是你要去的地方。
|
||||
|
||||
现在，我们生活在一个数据为王的时代，很多公司都善于运用 Twitter 上的数据。根据测量到的他们新产品的人气，尝试预测之后的市场趋势，分析 Twitter 上的数据有很多用处。通过数据，商人把产品卖给合适的用户，收集关于他们品牌和改进的反馈，或者获取他们产品或促销活动失败的原因。不仅仅是商人，很多政治和经济上的决定是在观察人们意见的基础上所作的。今天，我会试着让你感受下关于 Twitter 的简单 [情感分析][1]，判断这个 Twitter 是正能量，负能量还是中性的。这不会像专业人士所用的那么复杂，但至少，它会让你知道挖掘观念的想法。
|
||||
|
||||
我们将使用 NodeJs，因为 JavaScript 太常用了，而且它还是最容易入门的语言。
|
||||
|
||||
### 前置条件:
|
||||
|
||||
* 安装了 NodeJs 和 NPM
|
||||
|
||||
* 有 NodeJs 和 NPM 包的经验
|
||||
|
||||
* 熟悉命令行。
|
||||
|
||||
好了，就是这样。开始吧
|
||||
|
||||
### 开始
|
||||
|
||||
为了你的项目新建一个目录，进入这个目录下面。打开终端（或是命令行）。进入刚创建的目录下面，运行命令 `npm init -y`。这会在这个目录下创建一个 `package.json` 文件。现在我们可以安装需要的 npm 包了。只需要创建一个新文件，命名为 `index.js` 然后我们就完成了初始的编码。
|
||||
|
||||
### 获取 tweets
|
||||
|
||||
好了，我们想要分析 Twitter ，为了实现这个目的，我们需要获取 Twitter 的标题。为此，我们要用到 [twit][2] 包。因此，先用 `npm i wit` 命令安装它。我们还需要在 APP 上注册账户，用来访问 Twitter 的 API。点击这个 [链接][3]，填写所有项目，从 “Keys and Access Token” 标签页中复制 “Consumer Key”，“Consumer Secret”，“Access token” 和 “Access Token Secret” 这几项到 `.env` 文件中，就像这样：
|
||||
|
||||
```
|
||||
# .env
|
||||
# replace the stars with values you copied
|
||||
CONSUMER_KEY=************
|
||||
CONSUMER_SECRET=************
|
||||
ACCESS_TOKEN=************
|
||||
ACCESS_TOKEN_SECRET=************
|
||||
|
||||
```
|
||||
|
||||
现在开始。
|
||||
|
||||
用你最喜欢的代码编辑器打开 `index.js`。我们需要用 `npm i dotenv` 命令安装 `dotenv` 包来读取 `.env` 文件。好了，创建一个 API 实例。
|
||||
|
||||
```
|
||||
const Twit = require('twit');
|
||||
const dotenv = require('dotenv');
|
||||
|
||||
dotenv.config();
|
||||
|
||||
const { CONSUMER_KEY
|
||||
, CONSUMER_SECRET
|
||||
, ACCESS_TOKEN
|
||||
, ACCESS_TOKEN_SECRET
|
||||
} = process.env;
|
||||
|
||||
const config_twitter = {
|
||||
consumer_key: CONSUMER_KEY,
|
||||
consumer_secret: CONSUMER_SECRET,
|
||||
access_token: ACCESS_TOKEN,
|
||||
access_token_secret: ACCESS_TOKEN_SECRET,
|
||||
timeout_ms: 60*1000
|
||||
};
|
||||
|
||||
let api = new Twit(config_twitter);
|
||||
|
||||
```
|
||||
|
||||
这里已经用所需的配置文件建立了到 Twitter 上的连接。但我们什么事情都没做。先定义个获取 Twitter 的函数：
|
||||
|
||||
```
|
||||
async function get_tweets(q, count) {
|
||||
let tweets = await api.get('search/tweets', {q, count, tweet_mode: 'extended'});
|
||||
return tweets.data.statuses.map(tweet => tweet.full_text);
|
||||
}
|
||||
|
||||
```
|
||||
|
||||
这是个 async 函数，因为 `api.get` 函数返回一个 promise 对象，而不是 `then` 链，我想通过这种简单的方式获取推文。它接收两个参数 -q 和 count，`q` 是查询或者我们想要搜索的关键字，`count` 是让这个 `api` 返回的 Twitter 数量。
|
||||
|
||||
目前为止我们拥有了一个从 Twitter 上获取完整文本的简单方法，我们要获取的文本中可能包含某些连接或者原推文可能已经被删除了。所以我们会编写另一个函数，获取并返回即便是转发的 Twitter 的文本，，并且删除其中存在的链接。
|
||||
|
||||
```
|
||||
function get_text(tweet) {
|
||||
let txt = tweet.retweeted_status ? tweet.retweeted_status.full_text : tweet.full_text;
|
||||
return txt.split(/ |\n/).filter(v => !v.startsWith('http')).join(' ');
|
||||
}
|
||||
|
||||
async function get_tweets(q, count) {
|
||||
let tweets = await api.get('search/tweets', {q, count, 'tweet_mode': 'extended'});
|
||||
return tweets.data.statuses.map(get_text);
|
||||
}
|
||||
|
||||
```
|
||||
|
||||
现在我们拿到了文本。下一步是从文本中获取情感。为此我们会使用 `npm` 中的另一个包 —— [`sentiment`][4]。让我们像安装其他包那样安装 `sentiment`，添加到脚本中。
|
||||
|
||||
```
|
||||
const sentiment = require('sentiment')
|
||||
|
||||
```
|
||||
|
||||
`sentiment` 用起来很简单。我们只用把 `sentiment` 函数用在我们想要分析的文本上，它就能返回文本的相对分数。如果分数小于 0，它表达的就是消极情感，大于 0 的分数是积极情感，而 0，如你所料，表示中性的情感。基于此，我们将会把 tweets 打印成不同的颜色 —— 绿色表示积极，红色表示消极，蓝色表示中性。为此，我们会用到 [`colors`][5] 包。先安装这个包，然后添加到脚本中。
|
||||
|
||||
```
|
||||
const colors = require('colors/safe');
|
||||
|
||||
```
|
||||
|
||||
好了，现在把所有东西都整合到 `main` 函数中。
|
||||
|
||||
```
|
||||
async function main() {
|
||||
let keyword = \* define the keyword that you want to search for *\;
|
||||
let count = \* define the count of tweets you want *\;
|
||||
let tweets = await get_tweets(keyword, count);
|
||||
for (tweet of tweets) {
|
||||
let score = sentiment(tweet).comparative;
|
||||
tweet = `${tweet}\n`;
|
||||
if (score > 0) {
|
||||
tweet = colors.green(tweet);
|
||||
} else if (score < 0) {
|
||||
tweet = colors.red(tweet);
|
||||
} else {
|
||||
tweet = colors.blue(tweet);
|
||||
}
|
||||
console.log(tweet);
|
||||
}
|
||||
}
|
||||
|
||||
```
|
||||
|
||||
最后，执行 `main` 函数。
|
||||
|
||||
```
|
||||
main();
|
||||
|
||||
```
|
||||
|
||||
就是这样，一个简单的分析 tweet 中的基本情感的脚本。
|
||||
|
||||
```
|
||||
\\ full script
|
||||
const Twit = require('twit');
|
||||
const dotenv = require('dotenv');
|
||||
const sentiment = require('sentiment');
|
||||
const colors = require('colors/safe');
|
||||
|
||||
dotenv.config();
|
||||
|
||||
const { CONSUMER_KEY
|
||||
, CONSUMER_SECRET
|
||||
, ACCESS_TOKEN
|
||||
, ACCESS_TOKEN_SECRET
|
||||
} = process.env;
|
||||
|
||||
const config_twitter = {
|
||||
consumer_key: CONSUMER_KEY,
|
||||
consumer_secret: CONSUMER_SECRET,
|
||||
access_token: ACCESS_TOKEN,
|
||||
access_token_secret: ACCESS_TOKEN_SECRET,
|
||||
timeout_ms: 60*1000
|
||||
};
|
||||
|
||||
let api = new Twit(config_twitter);
|
||||
|
||||
function get_text(tweet) {
|
||||
let txt = tweet.retweeted_status ? tweet.retweeted_status.full_text : tweet.full_text;
|
||||
return txt.split(/ |\n/).filter(v => !v.startsWith('http')).join(' ');
|
||||
}
|
||||
|
||||
async function get_tweets(q, count) {
|
||||
let tweets = await api.get('search/tweets', {q, count, 'tweet_mode': 'extended'});
|
||||
return tweets.data.statuses.map(get_text);
|
||||
}
|
||||
|
||||
async function main() {
|
||||
let keyword = 'avengers';
|
||||
let count = 100;
|
||||
let tweets = await get_tweets(keyword, count);
|
||||
for (tweet of tweets) {
|
||||
let score = sentiment(tweet).comparative;
|
||||
tweet = `${tweet}\n`;
|
||||
if (score > 0) {
|
||||
tweet = colors.green(tweet);
|
||||
} else if (score < 0) {
|
||||
tweet = colors.red(tweet);
|
||||
} else {
|
||||
tweet = colors.blue(tweet)
|
||||
}
|
||||
console.log(tweet)
|
||||
}
|
||||
}
|
||||
|
||||
main();
|
||||
```
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://boostlog.io/@anshulc95/twitter-sentiment-analysis-using-nodejs-5ad1331247018500491f3b6a
|
||||
|
||||
作者：[Anshul Chauhan][a]
|
||||
译者：[BriFuture](https://github.com/BriFuture)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://boostlog.io/@anshulc95
|
||||
[1]:https://en.wikipedia.org/wiki/Sentiment_analysis
|
||||
[2]:https://github.com/ttezel/twit
|
||||
[3]:https://boostlog.io/@anshulc95/apps.twitter.com
|
||||
[4]:https://www.npmjs.com/package/sentiment
|
||||
[5]:https://www.npmjs.com/package/colors
|
||||
[6]:https://boostlog.io/tags/nodejs
|
||||
[7]:https://boostlog.io/tags/twitter
|
||||
[8]:https://boostlog.io/@anshulc95
|
||||
Loading…
Reference in New Issue
Block a user