~nobiot

Recent activity

[PATCH] GitHub issue #135 denote-get-path-by-id 26 days ago

From Noboru Ota to ~protesilaos/denote

Hi Prot,

Please find attached the patch I promised.

I tested the cases:

1. Multiple files exist for a single ID (html and Org)
2. [[denote:ID]] refers to a non-Denote file (e.g. a jpeg) -- the
   original issue reported in #135

– nobiot

Re: Exploring your Denote collection a month ago

From Noboru Ota to ~protesilaos/denote

> The discussion on getting Denote stored in the Org Roam database
> is a bit strange to me. Why not just use Org Roam?

Peter, I share this sentiment, too. I guess, though, probably from a
different angle from yours. My suggestion is also "Why not just use Org
Roam?" for database and existing visualizing capability -- but of
course, another avenue you have been exploring is a great addition: "Why
not direcly use files created by Denote?".

Denote can be configured to create and work within normal Org files with
using Org-ID (instead of Denote-specific "denote:" link type in
Org). Org-roam can then store these normal Org files with the ID.

All of this is possible because they all have a good, composable

Re: Exploring your Denote collection a month ago

From Noboru Ota to ~protesilaos/denote

Elias Storms <elias.storms@gmail.com> writes:

> The good news is that Org-roam UI can be modified to include the
> `denote:` type links as Peter pointed out here:
> https://github.com/org-roam/org-roam-ui/issues/272
> It's as easy as redefining a function.

Ah, our messages passed each other.  I see Peter has an easy workaround.
Sure; that's a good one, too.

You too have a good day.

—nobiot

Re: Exploring your Denote collection a month ago

From Noboru Ota to ~protesilaos/denote

Elias Storms <elias.storms@gmail.com> writes:

> Indeed, we'd need to look at the internals of Org-roam (but most of it
> goes over my head).

If I were to make Org-roam work with the 'denote:' type link, I'd
override function 'org-roam-db-insert-link', so that the db could
register 'denote:' as an 'id:' Semantically both are IDs, so this is
logically consistent to me.

I suspect that changing this initial part might just do it (informed by
my experience of implementing Md-roam):

    (defun org-roam-db-insert-link (link)

Re: Exploring your Denote collection a month ago

From Noboru Ota to ~protesilaos/denote

Elias Storms <elias.storms@gmail.com> writes:

> Org-roam does pick up the Denote IDs this way, but unfortunately the default
> `[[denote:id][Title]]` type of links aren't registered, resulting in a
> graph without links. I probably need to use `id:` type links for this to
> work.
>
> So perhaps some middleware is needed? One that would copy the Denote
> directory and modify each file so it has the property drawer and
> replaces links. Not sure what's best here.

No you do not need a middleware.

My simple suggestion is to *not* use 'denote:id' links, and simply use

Re: Exploring your Denote collection a month ago

From Noboru Ota to ~protesilaos/denote

Noboru Ota <me@nobiot.com> writes:
> If you are creating a new file via `denote`, you'd need to save the
> buffer and run `org-roam-db-sync`. I have just tried it with `emacs -Q`
> and it worked on my end.
> [...omission...]
> There should be a more elegant way to automate adding new files to the
> db, but I do not know a good way at the moment.  On my end, I use my own
> `md-roam` to interface markdown files with `org-roam` and "it just
> works" -- sorry, this is just a way to say "I just got off work and do
> not have a mental capacity to remember / work out how I do this
> exactly".

Sorry, a quick correction:

Re: Exploring your Denote collection a month ago

From Noboru Ota to ~protesilaos/denote

Elias Storms <elias.storms@gmail.com> writes:

> Nobiot, you mentioned using Org-roam on Denote notes:
>
>> All you need is to change the front matter configuration to use "id"
>> instead of Denote's default "identifier". Org-id will take the id
>> property. Denote's backlinks work without using "denote:" link type --
>> backlinks are searched with regexp for the id.
>
> Can you elaborate? The Org-roam database doesn't seem to pick up Denote
> identifiers, even when they are indicated with `#+id:` in the
> front-matter. Or do you mean to change the link type?

EFLS, happy to.  Instead of the keyword `#+id:`, you would need to use a

Re: Exploring your Denote collection 2 months ago

From nobiot to ~protesilaos/denote

Peter Prevos <peter@prevos.net> writes:

> Ideally, I like some Emacs code that converts the Denote folder to a
> JSON, like this:
>
> { "nodes": [
>      { "id": "20230123T085900", "name": "Denote example" },
>      { "id": "20230123T085930", "name": "A linked note" }
>    ],
>    "links": [
>      { "source": "20230123T085900", "target": "20230123T085930"}
>    ],
>    "tags": [
>     { "id": "20230123T085900", "tag": "denote" },

Re: Exploring your Denote collection 2 months ago

From nobiot to ~protesilaos/denote

Hi Peter
Nice work!

> 1. Build middleware so that Denote users can use the wonderful
> org-roam-ui package (perhaps after Emacs 29 lands with database
> support).

Denote works with Org-roam without a middleware for org files. I don not
mean to discourage you from developing a middleware with additional
features you wish to add. I would just like to point out that you have a
starting point that you can already play with.

All you need is to change the front matter configuration to use "id"
instead of Denote's default "identifier". Org-id will take the id

Re: [PATCH] feat: thing-at-point 'denote-id 3 months ago

From Noboru to ~protesilaos/denote

By the way, this can wait until we finalize the 
next release; perhaps you need a bit more 
headspace to focus on updating the manual,
etc…

> On 11. Dec 2022, at 16:21, Noboru <me@nobiot.com> wrote:
> 
> 
>> Very well!  Before proceeding with this, maybe we can do something about
>> the 'denote-link--link-at-point-string' and thus have a more general
>> approach?
> 
> Of course, we can take a look. I don’t seem to able to just get the identifier without the “denote:” string. Is there a good way with this function?
>