hyiltiz/vim-plugins-profile

:clock4: Profile Vim's plugins, generate awesome statistics and optimize (n)vim startup time

TL;DR

# Perl version as an example:
perl <(curl -sSL https://raw.githubusercontent.com/hyiltiz/vim-plugins-profile/master/vim-plugins-profile.pl)

# or Ruby
ruby <(curl -sSL https://raw.githubusercontent.com/hyiltiz/vim-plugins-profile/master/vim-plugins-profile.rb)

# or Python (add -p flat to plot a bar chart)
python <(curl -sSL https://raw.githubusercontent.com/hyiltiz/vim-plugins-profile/master/vim-plugins-profile.py)

# or R
bash <(curl -sSL https://raw.githubusercontent.com/hyiltiz/vim-plugins-profile/master/vim-plugins-profile.sh)

Here is a screenshot to have a quick look at what this is all about.

Here is a peek at the profiling result for my plugins:

Generating vim startup profile...
Parsing vim startup profile...
Crunching data and generating profile plot ...

Your plugins startup profile graph is saved
as `profile.png` under current directory.

==========================================
Top 10 Plugins That Slows Down Vim Startup
==========================================
   1	105.13	"vim-colorschemes"
   2	42.661	"vim-easytags"
   3	31.173	"vim-vendetta"
   4	22.02	"syntastic"
   5	13.362	"vim-online-thesaurus"
   6	7.888	"vim-easymotion"
   7	6.931	"vim-airline"
   8	6.608	"YankRing.vim"
   9	5.266	"nerdcommenter"
  10	5.017	"delimitMate"
==========================================
Done!

Story

If you use vim-plug (or other amazing plugin manager of your choice) to install your vim (gvim or macvim) plugins, then chances are high that it gets addictive. You will find yourself with several dozens of useful plugins.

vim-plug (and NeoBundle) offers you to load your plugins on-demand (lazy-loading). But which needs fine tuning? Well, using vim's built-in profiling vim --startuptime you can get a timing for all function calls during startup. However, the data is for each functions. You will have to figure out the math, and make sure those functions calls are form the same plugins. Even some sorting might help, but sorting the timing for each functions does not really make sense because it is really time of the plugins (but not the functions) that you really care about.

I am poor at doing mental math, even for simple sums. However, with the power of a simple bash script and R, we can get all we want.

This utility automatically detects your plugins directory, and does the rest of the hard work for you.

Supported Plugin-Managers

Here is the list of supported managers. Hopefully, your favourite plugin manager is among the list. If not, or if you prefer to manage your own plguins (using symlinks, of course), we could still adjust the code.

• vim-plug
• NeoBundle
• Vundle
• Pathogen

Installation

This is NOT a vim plugin! This is simply a profiler for your vim plugins that are installed through various plugin managers such as vim-plug.

Download the .zip here and then simply run the bash script:

bash ./vim-plugins-profile.sh  # calls R after pre-processing

# Use Perl powers! Built-in RegEx, no dependencies.
perl ./vim-plugins-profile.pl

# Or Python if you are from the other camp.
python ./vim-plugins-profile.py # -p flag plots a bar chart

# Alternatively use Ruby powers! Less dependency, graph with ASCII art
ruby ./vim-plugins-profile.rb

# To use an alternative executable such as neovim, pass it as the first argument.
ruby ./vim-plugins-profile.rb nvim

Then open the profile.png file for the result! It is that simple.

You can run it even without installation:

Dependency

Nothing. Well, at least Bash or Ruby or Perl or Python, but most systems already come with those pre-installed already.

If not (e.g. in M$ Windows systems), then you will need to install several tools before you can run this.

For the Perl Version:

• Git for Windows: The Git Bash contains a minimal Perl installation which is sufficient to run vim-plugins-profile.pl.

For the Ruby Version:

• Bash (Cygwin, or Git for Windows will also work).
• Ruby 2.3 (other version might as well just work. If not, you can repurt an Issue then I'll fix it)

To produce the eye-candy graphs, you can use either R, Perl + GNUplot or Python.

For R, the script prompts whether it should install the R:ggplot2 package if you already have R. Here are the list of dependencies for it:

• R
• R:ggplot2 (the ggplot2 package for R)

For Python, you can use either python2 or python3. If you have matplotlib (optional) installed, then you can even generate the bar plot. Implementation for people from the Python camp is merged from @bchretien. It also supports a custom command to run in the exec mode. Feel free to hack your way!