Custom template for generated file names

[gimalay]

As part of section extraction, the extracted content is added to a new file. In the current state, the file name is generated randomly using a combination of alpha-numeric characters. However, some users may prefer to have custom file name templates.

The purpose of this discussion is to determine the necessary parameters for the new file name template. It could include date components and the title of the extracted section.

+1 this discussion to prioritize this.

[ficd0]

Suppose a section called "My Section" is extracted to a newfile. Perhaps the template could have elements like date, title, and id; so you can make a template like %title-%id to produce my-section-etnwu.md?

[ficd0]

ie. title could be normalized to kebab case with non-ascii characters removed; to avoid any funny business with whitespaces or emojis in the filenames

[gimalay]

Yes, something like this. Wiki links style files names (My Section.md) may be a valid option too.

We probably want to support different ID generation schemas too. Some people prefer current time based IDs some may prefer sequential IDs.

We can pick a couple most obvious schemas and define a simple config structure supporting these options. Building this should be very straightforward.

[ha1ix]

I made two issues related to this before I saw this discussion:
One is about simply changing the newly extracted file name to reflect 'Parent-ExtractedHeader.md' as a massive improvement over the randomly generated file name (which clutters file directory in a significant human-unparseable way.

The second suggests a more thorough frontmatter, I cite my own metadata yaml format and zora's toml frontmatter (zora is a static site generator written in rust). Finally I propose a new frontmatter template for IWE.

[gimalay]

hey @ha1ix, I just added custom templates for extracted file names. You can use the titles of the sections for the file names, including the parent section name. The details are here: https://github.com/iwe-org/iwe/blob/master/docs/feature-extract.md

[ha1ix]

I may be a luddite for using directories instead of one flat folder, but as I go through and experiment with reformatting some of my notes to have extractions, even if all of the extracts followed "Parent-ExtractedHeader.md" naming, my file list will be a mile long by the end of the day. Not so big a deal in the end, I am sure I can get used to it. However, I would not be opposed to IWE mkdir new folders for large extracts.

For example, I just tested extracing-subsections on a document with 25 subsections. It worked (like magic!) but it doubled the number of files in my root folder for a book in progress. I can re-architect my notes, reconsider my hygeine practices, start with folders to begin with for example, which all are probably much simpler than trying to get IWE to fix my not-clever note hierarchy. What would you do?

I have a folder called THEMES and inside of it notes for all the different themes "DEATH," "LOVE" , etc. In order to make use of IWE auto-extraction I would have to make DEATH, LOVE etc Folders instead of .md notes. Which may work just fine and is quite easy to implement moving forward.

If the generated file names follow "Parent-ExtractedHeading.md" as I suggest above, no problem re-organizing them after ward into seperate directories (good luck with the random character names!). Would running iwe normalize fix all of the relative paths of my moved notes?

additional 2c: I don't care about dates in my note titles generally. current project doesnt benefit from it, though I could imagine it being useful in the future for different types of projects or personal diary.

[gimalay]

@ha1ix thanks for your input. I'm glad you find iwe useful. Generated file name templates is something I want to add at some point. It's just requires some thinking around the structure of the config and possible file name templates.

Let me know If you have a good example of a config like this from another project or have you own suggestions.

[ha1ix]

For my use case generated file names should reflect the name of the parent note in some way, for example if Parent is called Cheese and the sub-section is called "Spicy Cheeses" then extracting "Spicy Cheeses" from "Cheese" should result in "Cheese-Spicy-Cheeses.md" as the new note.

[ha1ix]

Is automatic file renaming robust enough to handle changes from current random-charactes-naming to a future name-templates standard? Thinking practically about the possibilty of breaking my notes setup in the future.

[gimalay]

Custom templates for generated file names are added, the details can be found here:

https://github.com/iwe-org/iwe/blob/master/docs/feature-extract.md