Filename Schema - put DATE after TITLE

[ryanwwest]

Hi, I was wondering if in the filename schema in https://github.com/novoid/Memacs/blob/master/docs/memacs_filenametimestamps.org, if it would be possible to put the timestamp after the 'title' of the file? For example, for 2010-12-31T23.59_Cookie_recipies.pdf instead have the file be recognized by Cookie_recipies_2010-12-31T23.59.pdf?

My reasoning is that I tried a similar schema from https://karl-voit.at/managing-digital-photographs/ of DATE-TITLE-TAGS.EXTENSION in the past but having the date first made it difficult to identify each file (mainly using the title), hard to select files in the terminal without fuzzy finders (where you have to type the start of the file first), and more frustrating to use in any other program that isn't set up for them. I love the idea in general and appreciate that DATE first creates a nice universal sort order, but I'd prefer to put TITLE first and am not sure if everything can be accomplished in the same way that way.

For example, can Org links still refer just to the (unique) timestamp, not necessarily the whole filename, if the filename doesn't start with the date?

[novoid]

Wow, my first "Github discussion" I guess. ;-)

I'm sorry, DATE-TITLE-TAGS.EXTENSION is my file name convention and I did not implement it that way so that its order can be changed easily.

However, I'm curious. What do you mean by "identify each file" or "hard to select files in the terminal without fuzzy finders"? I'm identifying files and selecting files all the time without any issue. Maybe I can help you there?

Related to Org links, you might want to read https://karl-voit.at/2022/02/10/lfile/ where I explain my current method to link files. It's not using Memacs any more. The only workflow I'm using myself that makes use of the Memacs filenametimestamp module is adding images to lazyblorg besides the usual "visualize files in the Memacs agenda" of course.

[ryanwwest]

Wow, my first "Github discussion" I guess. ;-)

It used to be issues for everything but if discussions are enabled I try to keep questions/theoretical stuff out to keep issue count down :) Of course it creates multiple places for sometimes similar discussions so maybe there are pros/cons.

[novoid]

ad me as a combination of Karl Voit, novoid, n0v0id (Twitter), publicvoit (homepage, Mastodon):

Unfortunately, I did not follow a consistent pattern with my aliases. :-(

Meanwhile, I try to use at least my real name in addition to my aliases so that people are able to link my account to my person.

ad shell UX:

I'm using more or less standard zsh in that regard, so we should both have the same UX. I often double click on the datestamp (I almost never use year-only date-stamps) and paste using the middle mouse in order to get a particular date into the current prompt line. As long as there aren't many files from the same day, that works fine. Of course, it also works fine for time-stamps as there are almost no files with the same time-stamp in my case. In case there are multiple files from the same day, multiple TABs do switch to completion mode suggesting one after another. That works file for me so far.

When I want to get a file from today, I use my AutoKey snippet $$dd which always inserts the day in form of "2023-09-26". So my workflow would be: "ls -la " + $$dd + TAB (or *) and get the appropriate result.

If you're only the Memacs module, you probably should be fine with a fork + adapting the regex. Adapting the line filename_timestamp_match = DATETIME_REGEX.match(file) and its regex might be enough already.

However, I'm using many tools around https://karl-voit.at/managing-digital-photographs/ which follow my filename convention.

[ryanwwest]

When I want to get a file from today, I use my AutoKey snippet $$dd which always inserts the day in form of "2023-09-26". So my workflow would be: "ls -la " + $$dd + TAB (or *) and get the appropriate result.

I do the same thing with the same ISO8601 formatted dates but it's alt/option-D across all systems. It's very valuable.

One reason why I asked about the flexible dates in filenames is because I want to design something that allows even more flexibility with dates and my data. I'd eventually like to build something that can read through markdown and Org files and search for YYYY-Q1, YYYY-MM and YYYY-MM-DD timestamps (optionally with additional time) at a heading, block/paragraph, and bullet point level and auto-include them into a sort of 'daily' journal unlike any other 'daily notes' system I've heard of to date. Meaning, if you added new thoughts to 5 notes files in a day, and 3 of them were important enough to warrant remembering by adding a date with our date-insert-shortcuts, then this tool would separately find each of the 3 files' headings(+content) / full paragraphs / bullet points (and sub-bullets) that listed that date and list them in a daily log as things that you edited, with hyperlinks to jump you to the source and maybe make it editable in-place. So in other words a sort of auto-journal where you don't have to think about actually journaling besides adding a timestamp if you're doing more than fixing typos or deleting stuff (which an alternative, git commits, might pick up as noise).

Is this a sort of thing that might integrate well as a new Memacs module (or overlap with an existing one or other tool)? I'm not familiar enough with Memacs - perhaps my idea would fit better with this (though they seem pretty related).

[novoid]

Well, it seems to be the case that you move the Memacs part where I'm using the Org agenda to the file level.

Memacs: open up the agenda, switch to the date you're interested in, activate the archive layer of your agenda in case you're using archive files for Memacs results and you've got your timeline view on your data.

You idea: open up a heading related to a date and see all the date-related linked files.

Valid approach but I prefer the standard agenda method. Furthermore, I would have no idea what to do with YYYY-Q1 files and day-based headings but that's an issue you need to resolve. ;-)

For now, I'll close this discussion. Feel free to post below if you have another thought related to that.

[ryanwwest]

Thanks for your thoughts. The Memacs agenda you describe where you switch to the date of interest and see what comes up is actually what I want too. I'll need to look up more deeply how it's displayed. It might be nice to jump to that date in the agenda view from any date of the SAME YYYY-MM-DD in an org file but that's not first priority. The biggest thing I want to figure out is finer granularity with which parts of the file are displayed in the agenda view as described above. I have some large files that have thoughts across many days and months so getting only the specific sections I wrote on a particular day would be nice (and editable from agenda view nicer).

[nbehrnd]

@ryanwwest Because of cookierecipes in the string of the file name, I initially thought about wanting to easy copy & past/yank (or move, delete) a bunch of files based on their name, rather than by the time stamp (date2name).

For one, where suitable, I use a file archive in similar shape as illustrated by the tree on Karl's blog post My Folder Hierarchy. Second, for simple navigation within my ~/.filetags_tagfilter/, I started to use the cross-platform vifm where a search by name (initiated by /) quickly narrows the selection if any of the string entered manually matches some portion of the file name. As shown below, the date/time stamps by date2name typically are not an obstacle to rapidly identify the files in question. For the purpose of the demonstration below, I replaced the default two-pane (Norton Commander like) display by a single column view:

[ryanwwest]

Cool, I love vim-based tools yet had somehow never heard of vifm. Partial filename matching like that is great. I'll check it out!

[nbehrnd]

@ryanwwest I just noticed (frankly, once set up, it was forgotten because it worked ever since) one has to explicitly request the partial filename matching in vifm. In my case, it likely is a combination of adjusting set suggestions, set ignorecase, set smartcase which is easiest documented by a share of the configuration file in the zip archive below. Altogether with calling a color style file (fetched from the program's project page, it yields a display as shown elsewhere.

example_vifm.zip