index notes
---
██╗███╗ ██╗██████╗ ███████╗██╗ ██╗ ██║████╗ ██║██╔══██╗██╔════╝╚██╗██╔╝ ██║██╔██╗ ██║██║ ██║█████╗ ╚███╔╝ ██║██║╚██╗██║██║ ██║██╔══╝ ██╔██╗ ██║██║ ╚████║██████╔╝███████╗██╔╝ ██╗ ╚═╝╚═╝ ╚═══╝╚═════╝ ╚══════╝╚═╝ ╚═╝
Also see tag index .
@index go to an index note on some topic
@idea idea for a draft, e.g. for a blog post
All files ( tree -f -I "images*|*.pdf|*.py|*.html" ).
TODO : Command for only listing directories.
Produce website from notes
Each note its own page (assuming you are in home folder):
for f in Zettelkasten/*.md; do pandoc -s --toc --number-sections --section-divs --lua-filter=filters/md_to_html.lua --css=https://www.edwinwenink.xyz/css/twocolumn.css --mathml $f -o site/${f%.md}.html; done
One page website:
pandoc --toc --metadata title="Archive Fever" --metadata author="Edwin Wenink" --number-sections --section-divs --css twocolumn.css ./**/*.md -o archivefever.html
With link to my website:
pandoc --toc --metadata title="<a href=https://www.edwinwenink.xyz/>Archive Fever</a>" --metadata author="<a href=https://www.edwinwenink.xyz/~ejw>Edwin Wenink</a>" --number-sections --section-divs --css twocolumn.css ./**/*.md -o archivefever.html
Does NOT work anymore in new pandoc version, unfortunately. So:
- remove explicit html in metadata
- add -s standalone flag; css does not load anymore otherwise.
pandoc -s --toc --metadata title="Archive Fever" --metadata author="Edwin Wenink" --number-sections --section-divs --css twocolumn.css ./**/*.md -o archivefever.html
Now let's start adding a lua filter:
```` pandoc -s --toc --metadata title="Archive Fever" --metadata author="Edwin Wenink" --number-sections --section-divs --css twocolumn.css --lua-filter fix-links.lua ./**/*.md -o archivefever.html
I got stuck.
The problem is that the section ID I want the backlink to point to is derived from the \#header, whereas the backlink points to the filename on disk.
Idea:
1. When parsing the link
2. Open filename on disk
3. Parse it's header
4. Convert that to its section-id
5. Make that the link target.
Alternatively, do not do a one-page website.
In that case, I need:
1. Something to produce the overal TOC (probably hard)
2. Add a link on every page pointing back
3. Adjust makefile to find the css file... from subdirectories.
## Visualize backlinks.json
Call `py visualize_backlinks.py` (assuming you have `backlinks.json`) and then run:
dot -K sfdp backlinks.gv -Tpdf -O ```
- K-flag selects algorithm
- fdp and sfdp seem to work well, everything stays readable. Probably best if the graphs get more complex.
- twopi is nice and compact; overlapping circles show a tight connection
- T-flag sets output format
- O-flag automatically sets the output name based on the input name. So output will be
backlinks.gz.pdf
Instructions for creating index
Linux
Make a tree with:
tree -f -I "images*|*.pdf|*.py|*.html"
The -f option keeps full relative path listings, which is necessary for gf
The -I option excludes everything matching the given pattern.
Aim to keep .md and perhaps the occasional .org
Windows
On Windows, run dir /s /b | findstr .md and then :%s/C:\\Users\\Edwin Wenink\\Documents\\Notes\\//g
Copy with yy:
.!dir /s /b | findstr .md
Then paste in command line with
Search and replace:
:%s/C:\Users\Edwin Wenink\Documents\Notes\//g
A quick way to deal with spaces is to visually select the whole line before doing gf.
Making a nice tree is a bitch on Windows due to annoying formatted PITA output.