The First Cry of Atom Today is the first day of the rest of my life.

Change Jekyll theme

久しぶりにJekyllのテーマを変えてみた。どのテーマも半年くらい経つと飽きが来るのか更新頻度にも影響するので定期的にテーマを変えるようにしている。 おかげでGitHub Pagesのリポジトリの入れ替えというか大掃除にも慣れてきた。 まず新しいテーマをcloneしてくる。 $ git clone git[email protected]:LeNPaul/Lagrange.git CMSはJekyllでカスタムのドメインをGitHub Pagesで使っている場合は下記を最低限コピーしてくれば問題なく動く。 CNAME _posts _config.yml: ただしテーマによっては加筆修正必要。 テーマによっては各postのtagが読... Read more

Contribution to Software Design from gihyo

The other day, I contributed a software magazine “Software Design” from 技術評論社. I wrote about Treausure Data Service internal as much as possible. My targets of this article are mainly software engineers who are struggling to develop and maintain data platforms. Treasure Data Service is a cloud base service for those people. So knowing how to ... Read more

Introduction of dllib

I keep thinking about releasing self developed deep learning for a few years. That’s a good way to motivate my learning about deep learning. I’ve already done several times with n42 and neurallib. But they are not so much general and hard to extend. So I reinitialize the project and re-introduce brand new deep learning framework named dllib. ... Read more

AWS CLI with jq make things easy

AWS CLI is a utility for manipulating your infrastructure on AWS. It enables us to launch EC2 or check instances without leaving out your command line. But it is not easy to use because the output is JSON as default. Though JSON is a format which is useful for machine and also human, it is not suitable as an output of command line tools. Let me... Read more

High Availability of InfluxDB

I’m now considering how to make InfluxDB highly available. InfluxDB runs on single node by default. It is necessary to make high available system by ourselves. I checked this article. In this post there are 5 options to be considered in my opinion. Backups High-Availability with Nginx Sharding InfluxDB Relay Clustering InfluxDB is ... Read more