This post introduces how to complie latest R for RStidio, and install RStudio Server in a Debian host.
While installing the R package “hdf5r“ version 1.3.3, I encountered the error:
fatal error: hdf5.h: No such file or directory
My OS is Debian buster/10.6, and the package “libhdf5-dev” has been installed. It looks R cannot find the path of hdf5 header and library files.
Aria2 creats a “.aria2” file for each object it downloading. It can be used
to resume the download if necessary.
A couple of days ago, I received an email from the Let’s Encrypt and was told the certificate of my domain will expire soon. This domain has been encrypted by certbot and the automatic renewal worked well at that time. About one month ago, I added it to cloudflare CDN. It may hindered the certificate renew.
So I spent some time to fix the problem.
Most websites supports HTTPS right now. So I decided to re-configure my download server to HTTPS in weekend, and update aria2 to latest version.
- OpenVZ VPS
- Debian 9.12
- Kernel 2.6.32
- aria2 1.35.0
- AriaNg 1.1.4
- Letsencrypt/certbot 0.28.0
- Nginx 1.10.3
Viewing and annotation of a huge phylogenetic tree, especially with hundreds or even thousamds taxa, is always a problem. This post provides a method to split a phylogenetic tree into multiple pages of a PDF file by pdfposter.
Since PyMOL 2.x released, there were no binary installer avaliable for Windows.
PyMOL Wiki provides the way to install open source PyMOL in Windows under Python 2.7.
This post provided the method to install PyMOL 2.x by conda (Anaconda or Minoconda) in Windows.
Conda is quite convenient for non-root users to install softwares in Linux system. But sometimes there is no packages available in anaconda repository and we have to install these softwares from source. This post introduces the way to compile and install softwares from source, use the Kyoto Cabinet Perl module as an example.
几年前就已经搭建了基于树莓派 2B+ 的下载机，不过由于宽带没有外网 IP ，一直没有实现外网的访问和控制。最近花了几天空闲时间，搭建好基于 frp 的内网穿透服务。这里把搭建过程记录下来，与大家交流。
BEAST (and BEAST 2, of course) is widely uesed for phylogenetic inference. But it is quite complicated. This post represents some useful tips while using BEAST.
Beagle-lib accelerates Bayesian and Maximun Likelihood phylogenetic software amazingly. Especially the beagle-lib v3.x improves 20-50% than version 2.x.
Although CUDA and OpenCL improves BEAST computing significantly, how to install beagle-lib without an Nvidia GPU card?
./configure --without-cuda --without-opencl
It will show these warning messages:
configure: WARNING: OpenCL not found or disabled. OpenCL implementation will not be built. If OpenCL support is desired, check the path to OpenCL and specify --with-opencl=/path/to/opencl
It’s OK. Simply continue with:
And do not forget run test with:
Sometimes it’s not possible to install beagle-lib in the default system library path /usr/local/lib. It could be resolved by setting the environment variable
For example, to install beagle-lib into /data/lib, in configure step:
Then append to .bashrc:
to enable the environment change.