How to use ASCII characters to search accented characters in consult-line?

[rodrigomorales1]

Let me explain a minimal working example:

I saved the contents of the code block below to /tmp/a.el.

(use-package consult
  :bind
  ("C-l" . consult-line))

I saved the contents of the code block below to /tmp/a.txt. Note that the last four lines have the accented characters: á (U+00E1, LATIN SMALL LETTER A WITH ACUTE) and é (U+00E9, LATIN SMALL LETTER E WITH ACUTE). These two characters are part of the Spanish alphabet.

a1
a2
e1
e2
á1
á2
é1
é2

I started Emacs using the following command:

$ emacs --quick --eval='(load "/tmp/a.el")' --eval='(find-file "/tmp/a.txt")'

I pressed M-x consult-line RET. I inserted a in the minibuffer and I pressed TAB. The buffer *Completions* was shown on top of the minibuffer, its content is shown below

Click on a completion to select it.
In this buffer, type RET to select the completion near point.

2 possible completions:
1 a1
2 a2

I want the results to be:

Click on a completion to select it.
In this buffer, type RET to select the completion near point.

2 possible completions:
1 a1
2 a2
5 á1
6 á2

That is, I want to use ASCII characters to search accented characters. How to accomplish this behavior with consult-line?

Motivation

I'm currently editing files in a language that use the accented characters (e.g. áéíóú). I wish I could use ASCII characters to match accented characters when using consult-line.

Additional information

I found some sexps on the Internet that might be relevant to this issue.

The following sexp was retrieved from this comment at Emacs Stack Exchange.

(char-fold-to-regexp "a")

\(?:a[̀-̄̆-̨̣̥̊̌̏̑]\|[aªà-åāăąǎǟǡǻȁȃȧᵃḁạảấầẩẫậắằẳẵặₐⓐａ𝐚𝑎𝒂𝒶𝓪𝔞𝕒𝖆𝖺𝗮𝘢𝙖𝚊]\)

The following sexp was retrieved from this answer at Emacs Stack Exchange.

(char-to-string (car (get-char-code-property ?á 'decomposition)))

a

[minad]

Hint: Search for char-fold-to-regexp in the Orderless documentation.

[rodrigomorales1]

Thanks, after searching related information on Orderless. I found two solutions:

Solution 1: Use % as a suffix for words in the prompt of consult-line

I saved the contents of the following code block to /tmp/a.el.

(use-package vertico
  :config
  (vertico-mode 1))

(use-package orderless
  :init
  (setq completion-styles '(orderless)))

(use-package consult
  :bind
  ("C-l" . consult-line))

I saved the contents of the following code block to /tmp/a.txt.

día mano fruta
avión batería vegetal
mercado día tienda

I started Emacs using the following command:

emacs --quick --eval='(load "/tmp/a.el")' --eval='(find-file "/tmp/a.txt")'

I searched dia% and the lines candidates were: 1 día mano fruta and 3 mercado día tienda.

Solution 2: Add char-fold-to-regexp to orderless-matching-styles

I learned this from this comment at Reddit.

I saved the contents of the following code block to /tmp/a.el.

(use-package vertico
  :config
  (vertico-mode 1))

(use-package orderless
  :init
  (setq completion-styles '(orderless))
  (setq orderless-matching-styles
        (list #'orderless-literal #'orderless-regexp #'char-fold-to-regexp)))

(use-package consult
  :bind
  ("C-l" . consult-line))

I saved the contents of the following code block to /tmp/a.txt.

día mano fruta
avión batería vegetal
mercado día tienda

I started Emacs using the following command:

emacs --quick --eval='(load "/tmp/a.el")' --eval='(find-file "/tmp/a.txt")'

I searched dia and the candidates were: 1 día mano fruta and 3 mercado día tienda.

[minad]

Note that solution 2 is slow due the costly regexps, therefore we don't use this configuration by default in Orderless.