How I plumbed in a custom find-grep scripts

[paddymul]

I wanted to write a custom find-grep script that did the following:

1. returned matches from the preferred file extensions (for python -> py,js, ts, jsx, tsx) and an exhaustive exclude list
2. then returned matches only from the exclude list (site-packages, node_modules...)

This lets me run a single consult-grep and get back all the relevant matches, with the most relevant matches returned at the top and more quickly

It was very difficult to plumb this into consult-mode. There are multiple builder functions that are nested. Eventually this is what I built. I'm totally open to feedback if there is a cleaner way to integrate this.

It would help if the elisp functions were a bit better documented. It took a while to realize that I was returning results in the wrong format.

(defvar paddy-find-script  "/Users/paddy/.emacs.d/personal/find_script.sh")
(defun my-parse-extensions (extensions)
  (format ".*\\.%s$" (regexp-opt extensions t)))

(defun my-parse-directories (directories)
  (format ".*\\.%s.*" (regexp-opt directories t)))

(defun paddy-py-find-grep-list (root-directory search-term)
  (let* ((py-preferred-extensions (my-parse-extensions '("py")))
	 (py-secondary-directories (my-parse-directories '(".venv" "foo")))
	 (py-secondary-extensions  (my-parse-extensions '("js" "ts" "jsx" "tsx")))
	 (py-never-directories (my-parse-directories '(".ruff" ".git" "pycache" "dist")))
	 (py-never-extensions (my-parse-extensions '("pyc" ".d.ts"))))
    (list paddy-find-script
	  root-directory  ; search_root $1
	  search-term     ; search_term $2
	  "" ; preferred_directories $3 - unused
	  py-preferred-extensions ; preferred_extensions $4 - unused
	  py-secondary-directories ; secondary_directories $5
	  py-secondary-extensions ; secondary_extensions $6
	  py-never-directories ; never_directories $7
	  py-never-extensions ; never_extensions $8
	  )))



;; works, accepts regexp and path
(defun paddy-consult-py-find-grep ()
  (interactive)
(cl-letf ((
       (symbol-function #'consult--grep-make-builder) (lambda (paths)
  "Build grep command line and grep across PATHS."
  (let* ((cmd (consult--build-args consult-grep-args))
         (type (if (consult--grep-lookahead-p (car cmd) "-P") 'pcre 'extended)))
    (lambda (input)
      (message "50 input %s" input)
      (pcase-let* ((`(,arg . ,opts) (consult--command-split input))
                   (flags (append cmd opts))
                   (ignore-case (or (member "-i" flags) (member "--ignore-case" flags))))
	(message "54 arg %s flags %s" input flags)
        (pcase-let ((`(,re . ,hl) (consult--compile-regexp arg type ignore-case)))
	  (let (
		(final-command (cons
				(paddy-py-find-grep-list (car paths)  (consult--join-regexps re type))
			       hl)))
	  (message "final-command %s" final-command)
	  final-command
	  ))))))))
  (call-interactively 'consult-grep)))

and the find-script

#!/bin/sh

## This commands runs a series of find commands that returns most wanted results first (quickly) and less wanted results later


search_root=$1
search_term=$2
preferred_directories=$3
preferred_extensions=$4

secondary_directories=$5
secondary_extensions=$6

never_directories=$7
never_extensions=$8

#for future os x installs
#brew install findutils

# we want gnu find, not bsd
#FIND_COMMAND=/opt/homebrew/bin/gfind
FIND_COMMAND="$(command -v gfind)"


#preferred search
$FIND_COMMAND $search_root\
	      -type f  -iregex $preferred_extensions \
	      -print0 \
	      -not \(  -iregex $secondary_directories \) \
	      -not \(  -iregex $secondary_extensions \) \
	      -not \(  -iregex $never_directories \) \
	      -not \(  -iregex $never_extensions \) \
    | xargs -0 grep --null --line-number  $search_term


$FIND_COMMAND $search_root\
	      -type f  -iregex $secondary_extensions \
	      -print0 \
	      -and \(  -iregex $secondary_directories \) \
              -not \(  -iregex $preferred_extensions \) \
	      -not \(  -iregex $never_directories \) \
	      -not \(  -iregex $never_extensions \) \
    | xargs -0 grep --null  --line-number $search_term