Basically I would like to be able to change the note directory to manage seperate knowledge bases, for example:
- Personal Wiki
- Static Site
- Wiki for Project
- School Project public knowledge base etc.
This should be fed in as a -c option.
Refer to this line here
I should move the emojis to the end so that they don’t interfere when using tty
As opposed to having no feedback and just operating on STDIN and STDOUT. Could a similar effect be acheived using /tmp or /dev/shm?
see the stoml package.
Add Recommended Aliases to the readme as well as fish functions.
should I restructure the subfunctions to accept only 1 argument either help or directory and fail in the absence of any given argument?
- Much Simpler to implement and maintain
- Less clean to pass in functions
- Less Extensible
When this issue is resolved switch back to MDCat
just like html-dir add dw-dir using this script:
~/bin/ImportHtml2Dokuwiki.sh
The links in the html-dir export are absolute, a second pass is needed to make them relative, this will be necessary for everything except for self-contained stuff.
Make a seperate function for self contained and relative stuff.
Currently html-dir CANNOT be used to import into mediawiki because even thtough the files get pulled in the links are broken.
calling pandoc ./file.html --extract-media='./media' -o ./file.html is necessary to import a zip file into mediawiki.
<a href=”file:NoteFind.sh::ramtmp=”$(mktemp -p @@html:@@dev/shm@@html:@@)”“>In the SkimAndGrep function there should be export options for HTML mathml [fn:1] as well as odt[fn:2]
I should change the emacs call to open emacsclient instead of emacs
Completed this by calling help and exiting if the arguments are empty in cadmus.
See here
Recoll should be called with regard to a config, the idea being i could run:
cadmus -c personalcadmus -c unicadmus -c work
and have different note bases
Take the command query and change white spaces into | characters, this way search terms will be highlighted in matches.
These desperately need to be standardised like search and NoteFind have been
Currently all the tag scripts call a relative script, not in the PATH, maybe this is actually OK? it means 0 install…
WikiJS uses a broken yaml implementation but they can be extracted with some ripgrep, this should represent a further option in FilterNotesByTMSUTag.sh.
rg --pcre2 '(?<=[---\n[\s\S]ags:).*[,\s|:\s][a-z]+' -t markdown -o | sd -s ':' ' ' | sd -s ',' ' ' | sed s/^/tmsu\ tag\ /Should this do grep or recoll?
Probably grep just in case recoll ever brakes.
As opposed to using its own implementation.
Cadmus should deal with the clipboard as a frontend, relying on scripts that print to STDOUT, this will promote modularity.
If a notename that you point to (like choose from fzf) has the same name as a folder (think julia.md) then the link you get back is a fucking mess.
This doesn’t work for $$, fix that.
Fixed this by changing this
FILE="$(basename ${FILEPATH})"to this
[[ -f "${FILEPATH}" ]] && FILE="$(basename ${FILEPATH})" || exit 0I had some notes on my mediawiki that might help me figure this out.
Use FPM to make multiple packages bundling everything together.
Consider also just using install --help and doing something like this or like this.
| Number | Description |
| 3 | File not found |
| 4 | directory not found |
| 5 | unrecognized option |
| 6 | Missing Dependency |
Exit Code Number ∉ {1, 2, 126, 127, 128, 130, 255}
[fn:2] This could so something like pandoc -o /tmp/myfile.odt; xdg-open /tmp/myfile.odt
[fn:1] Think for using Thunderbird