Tianke Youke

真的见鬼了，谁能想到我搜肠刮肚思考Charles为啥哪哪都是404，关了v2ray，关了Chrome的科学上网插件，关了系统vpn的自动配置，到处弄遍了，最后发现导致Charles不能用的是：

Jupyter！！！

去看了一下原因……

真没注意Jupyter的默认端口和Charles一样，都是8888，你说巧不巧！？

了解了一点点Nginx之后，我发现可以利用二级目录，随便把一些静态网站弄到服务器上来。大一跟几个小组同学在网页设计课一起撸的「指尖星空」网站终于可以拨云见日了——

指尖星空 https://jyzhu.top/fingerspace/

其实很简单，只要把整个静态网站用scp传到Ubuntu服务器上，然后在Nginx配置一下子目录的跳转就好了。

记录一下Nginx的配置吧，因为缺少系统的理解，我在这里绕了半天圈，看了一篇《Nginx虚拟目录alias和root目录》才理解此处应该设置alias。

太高兴了！糊里糊涂地搞了一堆SEO，我终于也出现在表层互联网世界了！

不过我到现在也搞不明白到底怎样做优化才是正确的，反正就是乱弄了一通，也没有什么精力……

不过也没关系……反正我的内容迄今为止也挺杂乱和没有干货的，整个网站好像也没什么价值。

搜「jyzhu.top」和「tianke youke」都可以搜到，虽然惨兮兮只被收录了域名和网站标题……但已经很不容易了……

Recently I am really tired. I got tooo much to do, that I couldn't finish. And my mental as well as physical health status is just in danger. I am always tired, exhausted, losing my hair, eager to sleep. Sometimes I couldn't do anything, but most of the time I just buried myself into the task in my hands, thinking of nothing. I don't dare to think too much. Because there's always a question hanging there, that is, what's the meaning of all of this? Why on earth am I worried about my TOEFL? Why do I try my best to learn, to do everything? What do these things targeting? A success? Some happiness? What is a success? Is there any happiness?

I don't dare to ask. Not only because I do believe that there's no answer, but also because all of my intimates just told me to not bother myself thinking of these useless things. But I have another question here: if these questions don't have answers, then aren't those trivial problems in daily life truly useless?

Hell. I don't want to talk about them too much. I don't have time. I am trapped in these trivial things, and I just wanna be trapped. It's better than die, isn't it?

Soooooo, I am gonna set several rules for myself. That's about what to do and how to live:

A robot to recover Rubik's cube, based on Arduino, using Python combined with Machine Learning and OpenCV.

这学期开了数据科学课，很开心。我校信管还是有点跟上时代，不至于跟着整个老旧专业的没落而没落，而是有转型打算的嘛。听c老师说甚至学院开会打算把整个专业改名，安上「大数据」，但后来发现如此的话教学内容也要大改而师资不够。说白了就是学院里边很多老师还是太软了，技术不大行，教不了硬的。

这课两周内把前四章基础的东西讲完了，数据导入和简单数据预处理之类的，比我想象中快，我还比较期待后面能学到新东西。

第五章开了个头，讲的是「选择和评价模型」，大抵是介绍机器学习中的常用模型，分清遇到不同数据分析需求时适用的模型。令我失望的是，真的真的只是讲了一下模型的种类，告诉你模型可以用来做什么，连一点点原理都没有涉及。我希望今天把课程所学温习，把上学期的实践课学到的机器学习原理整合进知识框架里，顺道把之前的笔记好好整理一下。如果合适的话，再发个公众号文章也不错（心心念念的广告费）。

不同于ml角度，把机器学习模型分为监督学习和无监督学习；ds是从业务问题的角度出发，把模型分为分类、打分、概率、排名、聚类等。我显然更喜欢前者的分类，相比较之下后者的分法互相交叉，令人迷惑。

1. R is easy. Half of Python half of Matlab.
2. R最大的简便就是把各种统计概率机器学习功能都打包好了嘛

作业题改错：

One:

1
2
3
4
5
6
7

#确定观察值相对较少的站点
agg<-aggregate(Station,list(Station),length)
minStation<-agg[agg$x==min(agg$x),]
minStation
#生成一个忽略这些站点的新数据框
newisit<-isit[!Station %in% minStation[,1],]
newisit

代码改来改去已经改迷糊了……所以从现在开始每次改了些什么都记录一下吧：

上上次：老师布置新添加一个话题之后，我简单复制了一份新爬虫，把topic参数改了一下，这样打算topic 1爬完后用新爬虫爬topic 2，互不干涉；但是老师找上来说两个话题必须同时爬才行。

所以上次：我把两个爬虫融合在一起，让topic参数不由手动指定，而根据URL自动识别，然后美滋滋觉得没问题了。但是今天一运行发现bug，这样两个话题爬到的数据都会存进同一份数据库里啊！

为了让两个话题分别存到2个数据库，这次：绞尽脑汁，还是分成了两个爬虫，两个爬虫分别连接2个redis数据库，使用2个pipelines存进2个MySQL数据库里。具体来说，

Recently I'm feeling very upset, even say, depressed. Especially when the sun goes down, and the sky turns dark, I almost wanna cry. I hope to get drunk once but after rethinking, I find it no chance. My baby won't let me do this.

I don't know specificly why I come here again. Perhaps it's because of the recent pressure from the English test, which I regarded as an impossible goal from the bottom of my heart. Once something impossible comes to my life, I cannot stand it at all.

I'm confused about why I should keep alive. I know this is a naive question and all the feelings are trivial.

So I won't talk about them anymore.

天地的雨下起来的时候
我坏了
烦躁的雨衣整齐叠在一边
心脏染上了一层蜡黄
我已装死很久了

在这场雨里
别再指望我解释什么
语言的甲板处处断裂
戏拟的修辞被扯成布条
你或许能听到
我的灵魂站在甲板上喃喃：
我也无法接受……我该如何契合
……至今的复杂性？