Header Search and/or Regex full-text search

[chawl3y]

Love the app! I've imported over 400 files into it and am having fun discovering all the features and tips.
I have several years worth of log files with entries broken down by date, each with a Markdown header (i.e. ## 2026-08-26). Occasionally I like to search for log entries from the same date a year ago (or two or three), but I'm unable to find results if I search for a partial string (I've tried -08-26, *-08-26).
Search does not seem to return any results for matches found in a Markdown header.

So, three questions:

• Does search utilize some kind of wildcard for fuzzy searching?
• Can search be made to return results found in Markdown headers?
• and, unrelated, but helpful, can search at some point in the future support regex searching?

Thanks again for the great work. This wiki has proven very useful.

[perber]

@chawl3y @chawl3y Thanks for your kind words. I guess the issue occurs because - is treated as a token in SQLite FTS.

I need to investigate this a bit further.

Usually, Markdown headers have more weight than the content, so yes, the search should find Markdown headers as well.

Currently regex is not planned.

[chawl3y]

Looking into a bit, it looks like the tokenizer

tokenize = "unicode61 tokenchars '-_/+#.'"

in internal/search/sqlite_index.go is optimizing searches for whole words (which is good)

Changing it to

tokenize="trigram"

would allow full text search of any 3-character string regardless of where it appears in the text.

But I think this might bloat the index storage, especially on huge amounts of files.

I'm not sure if I understand this completely, so this might be wrong.

I understand if the decision was made to keep the search index smaller. Searching, currently, is very performant on my 584 files (and growing!)

I'll see if I can figure out a workaround for my use-case.

Thanks again!