notes by alifeee profile picture tagged git-diff (2) rss

return to notes / blog / website / weeknotes / linktree

here I may post some short, text-only notes, mostly about programming. source code.

tags: all (44) scripting (13) linux (5) bash (4) geojson (4) obsidian (4) android (3) github (3) html (3) jq (3) ............ see all (+63)

comparing EPC certificates with git-diff # prev single next top

tags: git-diff, scripting, housing • 484 'words', 145 secs @ 200wpm

Our house just got a new EPC certificate. You can (maybe) check yours on https://www.gov.uk/find-energy-certificate.

I'm interested in easy ways to see change. Trying to compare the old and new webpages by eye is hard, which leads me to text-diffing. I can copy the contents of the website to a file and compare them that way. Let's. I did a similar thing a while ago with computer benchmarks.

I manually create two files by copying the interesting bits of the webpage, called 1 and 2 (because who has time for .txt extensions). Then, I can run:

git diff --no-index -U1000 ~/1 ~/2 > diff.txt
cat diff.txt | sed -E 's#^\+(.*)#<ins>\1</ins>#' | sed -E 's#^-(.*)#<del>\1</del>#' | sed 's/^ //'

The latter command turns each into HTML by turning + lines into <ins> ("insert"), - into <del> ("delete"), and removing leading spaces on other lines. Then, I can whack the output into a simple HTML template:

<!DOCTYPE html>
<html>
  <head>
    <style>
    body { background: black; color: white; }
    pre { padding: 1rem; }
    del { text-decoration: none; color: red; }
    ins { text-decoration: none; color: green; }
    </style>
  </head>
  <body>
<pre>
diff goes here...
<del>del lines will be red</del>
<ins>ins lines will be green</ins>
</pre>
  </body>
</html>

The final output is something like this (personal information removed. don't doxx me.)

Energy rating
D

Valid until
05 February 2025
05 February 2035


Property type
Mid-terrace house
Total floor area
    130 square metres 
    123 square metres


This property’s energy rating is D. It has the potential to be C. 
This property’s energy rating is D. It has the potential to be B.


Features in this property


Window                Fully double glazed                                        Good
Roof                  Pitched, no insulation (assumed)                           Very poor
Roof                  Roof room(s), no insulation (assumed)                      Very poor
Roof                  Roof room(s), insulated (assumed)                          Good
Lighting              Low energy lighting in 64% of fixed outlets                Good
Lighting              Low energy lighting in all fixed outlets                   Very good
Secondary heating     None                                                       N/A


Primary energy use


The primary energy use for this property per year is 303 kilowatt hours per square metre (kWh/m2).
The primary energy use for this property per year is 252 kilowatt hours per square metre (kWh/m2).

Good job on us for having 100% low energy lighting fixtures, I guess...

Really, this is a complicated way to simplify something. I like simple things, so I like this.

back to top

comparing PCs with terminal commands # prev single next top

tags: pc-building, scripting, git-diff • 738 'words', 221 secs @ 200wpm

I was given an old computer. I'd quite like to make a computer to use in my studio, and take my tower PC home to play video games (mainly/only local coop games like Wilmot's Warehouse, Towerfall Ascension, or Unrailed, and occasionally Gloomhaven).

It's not the best, and I'd like to know what parts I would want to replace to make it suit my needs (which are vaguely "can use a modern web browser" without being slow).

By searching the web, I found these commands to collect hardware information for a computer:

uname -a # vague computer information
lscpu # cpu information
df -h # hard drive information
sudo dmidecode -t bios # bios information
free -h # memory (RAM) info
lspci -v | grep VGA -A11 # GPU info (1)
sudo lshw -numeric -C display # GPU info (2)

I also found these commands to benchmark some things:

sudo apt install sysbench glmark2
# benchmark CPU
sysbench --test=cpu run
# benchmark memory
sysbench --test=memory run
# benchmark graphics
glmark2

I put the output of all of these commands into text files for each computer, into a directory that looks like:

├── ./current
│   ├── ./current/benchmarks
│   │   ├── ./current/benchmarks/cpu
│   │   ├── ./current/benchmarks/gpu
│   │   └── ./current/benchmarks/memory
│   ├── ./current/bios
│   ├── ./current/cpu
│   ├── ./current/disks
│   ├── ./current/gpu
│   ├── ./current/memory
│   └── ./current/uname
└── ./new
    ├── ./new/benchmarks
    │   ├── ./new/benchmarks/cpu
    │   ├── ./new/benchmarks/gpu
    │   └── ./new/benchmarks/memory
    ├── ./new/bios
    ├── ./new/cpu
    ├── ./new/disks
    ├── ./new/gpu
    ├── ./new/memory
    └── ./new/uname
4 directories, 19 files

Then, I ran this command to generate a diff file to look at:

echo "<html><head><style>html {background: black;color: white;}del {text-decoration: none;color: red;}ins {color: green;text-decoration: none;}</style></head><body>" > compare.html
while read file; do
  f=$(echo "${file}" | sed 's/current\///')
  git diff --no-index --word-diff "current/${f}" "new/${f}" \
    | sed 's/\[\-/<del>/g' | sed 's/-\]/<\/del>/g' \
    | sed -E 's/\{\+/<ins>/g' | sed -E 's/\+\}/<\/ins>/g' \
    | sed '1s/^/<pre>/' | sed '$a</pre>'
done <<< $(find current/ -type f) >> compare.html
echo "</body></html>" >> compare.html

then I could open that html file and look very easily at the differences between the computers. Here is a snippet of the file as an example:

CPU(s):                   126
  On-line CPU(s) list:    0-110-5
Vendor ID:                AuthenticAMDGenuineIntel
  Model name:             AMD Ryzen 5 1600 Six-Core ProcessorIntel(R) Core(TM) i5-9400F CPU @ 2.90GHz
    CPU family:           236
    Model:                1158
    Thread(s) per core:   21
    Core(s) per socket:   6
    Socket(s):            1

Latency (ms):
         min:                                    0.550.71
         avg:                                    0.570.73
         max:                                    1.621.77
         95th percentile:                        0.630.74
         sum:                                 9997.519998.07

    glmark2 2021.02
=======================================================
    OpenGL Information
    GL_VENDOR:     AMDMesa
    GL_RENDERER:   AMD Radeon RX 580 Series (radeonsi, polaris10, LLVM 15.0.7, DRM 3.57, 6.9.3-76060903-generic)NV106
    GL_VERSION:    4.64.3 (Compatibility Profile) Mesa 24.0.3-1pop1~1711635559~22.04~7a9f319
...
[loop] fragment-loop=false:fragment-steps=5:vertex-steps=5: FPS: 9303213 FrameTime: 0.1074.695 ms
[loop] fragment-steps=5:fragment-uniform=false:vertex-steps=5: FPS: 8108144 FrameTime: 0.1236.944 ms
[loop] fragment-steps=5:fragment-uniform=true:vertex-steps=5: FPS: 7987240 FrameTime: 0.1254.167 ms
=======================================================
                                  glmark2 Score: 7736203

It seems like the big limiting factor is the GPU. Everything else seems reasonable to leave in there.

As ever, I find git diff --no-index a highly invaluable tool.

back to top