Have you ever heard of the top command in linux. Yes, the one that lists out currently running processes with a lot of other meta information about usage. A typical top command output, when piped to a file, looks like this:
Whoa! A lot of information being thrown here, but not exactly human readable. Sometimes when I am using a tool in linux, I often spin up the OG top command to look at how much the tool is using my memory as compared to other processes. I want a clear readable format which I can use to view the result of the top command so that I can embed it wherever I like, be it a blog or an issue. Basically I want something like this:
top - 00:26:50 up 15:33
1 user
load average: 0.34
0.69
0.84
Tasks: 370 total
1 running
279 sleeping
0 stopped
0 zombie
%Cpu(s): 5.2 us
1.5 sy
0.1 ni
92.3 id
0.8 wa
0.0 hi
0.1 si
0.0 st
KiB Mem : 16262016 total
9118020 free
1652820 used
5491176 buff/cache
KiB Swap:
0 total
0 free
0 used. 14034404 avail Mem
PID
USER
PR
NI
VIRT
RES
SHR
S
%CPU
%MEM
TIME+
COMMAND
15353
angad
20
0
57824
4240
3476
R
12.5
0.0
0:00.02
top
15312
angad
20
0
273360
34332
12680
S
6.2
0.2
0:00.24
vim
1
root
20
0
225944
9512
6544
S
0.0
0.1
0:42.51
systemd
2
root
20
0
0
0
0
S
0.0
0.0
0:00.07
kthreadd
4
root
0
-20
0
0
0
I
0.0
0.0
0:00.00
kworker/0:+
6
root
0
-20
0
0
0
I
0.0
0.0
0:00.00
mm_percpu_+
7
root
20
0
0
0
0
S
0.0
0.0
0:00.12
ksoftirqd/0
8
root
20
0
0
0
0
I
0.0
0.0
0:30.62
rcu_sched
9
root
20
0
0
0
0
I
0.0
0.0
0:00.00
rcu_bh
Aaah now this is more like it. Informative yet soothing to the eyes. Often, data visualization is limited by the perceptive short term memory load by the sheer volume of the displayed information. More times than not, we can mould the information a certain way so that it is more perceptive. Markdown comes to the rescue here!
The How
In this series, I am going to tell you how I converted the garbled top command output to a human readable form using the following steps:
-d: This is the delay time interval (5 seconds here)
-n: This specifies the number of cycles to run. Since we only want a snapshot, we specify 1.
-b: Notice that when piped to a file, the top command prints out a lot of control characters. This can be avoided by using batch mode, which removes there characters for further processing.
Whatever comes after the pipe is to make sure we only get a specified number of lines in the top output (15 here). We can specify any number here, or remove this option for the full top output
We are using the > symbol to take the output and pipe it to a file called myfile. Use vim to open this file, and we will proceed.
Converting metadata to a list
We talked about macros in the last entry of this series. We are going to create a macro for converting the first section of the top command to a bulleted list.
Press q + a to start recording the macro in the a key. The following set of functions need to be performed sequentially:
Find all commas and add a new line instead, with a tab and a *, to specify that the item is a sub element of the list:
:%s/,/\r\t\*/g
Note that here we have used the \r symbol instead of \n as \n appends an extra character at the end of the newline. g is used to specify that this action needs to be performed on every occurrence of , in a line.
Cleanup the extra space in the last point:
KiB Swap: 0 total
* 0 free
* 0 used. 13797256 avail Mem
For this, search for Kib, then delete the space till the 0 reaches the : and press enter to go into a new line. Insert a *. Cleanup the extra space in the last two fields also. To get this:
KiB Swap:
* 0 total
* 0 free
* 0 used. 13797256 avail Mem
Insert a bullet behind every field by searching and pre-pending the line with a *.
Now we have something that looks like this:
top - 00:26:50 up 15:33
1 user
load average: 0.34
0.69
0.84
Tasks: 370 total
1 running
279 sleeping
0 stopped
0 zombie
%Cpu(s): 5.2 us
1.5 sy
0.1 ni
92.3 id
0.8 wa
0.0 hi
0.1 si
0.0 st
KiB Mem : 16262016 total
9118020 free
1652820 used
5491176 buff/cache
KiB Swap:
0 total
0 free
0 used. 14034404 avail Mem
We have our first section done and recorded in macro 'a'. To paste the contents of the macro in the current file, you can do " + a + p. You can use this output to create your custom function for performing the actions specified and adding them to a keybinding if you want.
Converting process data to a table
Under a metaphorical microscope, when we highlight all of the hidden characters, the process data looks something like this:
Converting this to a table format will include the following steps:
Search for "PID" using /PID.
Paste the following line below the line that contains PID:
Since we will be replacing spaces with | to make a table, we have to make sure none of our process names have spaces. Till now I have only seen the Web Content process having a space. Simply replace it:
:%s/Web Content/Web-Content/g
Go to the line with PID. Now from that line till the end of the document delete the space from each line beginning:
" The + suggests that match one or more space" .,$ means from current line till the end of the document:.,$s/^\s\+//
Convert each line ending to a |:
" $ is the line ending character:.,$s/$/|/
Convert each line beginning with a |:
" ^ is the line beginning character:.,$s/^/|/
Convert one or more spaces to a |:
:.,$s/\s\+/|/g
And you are good to go. The process data now looks like this:
