From 5c241d65ed4a6ec2bc3e5d75d6858ed6722f1b17 Mon Sep 17 00:00:00 2001
From: sommerfeld <sommerfeld@sommerfeld.dev>
Date: Wed, 13 May 2026 13:43:23 +0100
Subject: feat(sway): add dictate (whisper.cpp) and ocr (tesseract) keybinds

Push-to-talk dictation toggle on Super+i: parecord captures 16 kHz mono
WAV, whisper-cli transcribes (auto language), output is typed via wtype
and copied to the clipboard.

Region OCR on Super+Shift+o: slurp + grim feed tesseract (eng+por),
result lands in the clipboard with a notification preview.

Adds wtype to wayland.txt; tesseract (+eng/por data) and whisper.cpp +
the large-v3-turbo-q5_0 model package to extra.txt.
---
 dot_local/bin/executable_ocr | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)
 create mode 100644 dot_local/bin/executable_ocr

(limited to 'dot_local/bin/executable_ocr')

diff --git a/dot_local/bin/executable_ocr b/dot_local/bin/executable_ocr
new file mode 100644
index 0000000..6f6191f
--- /dev/null
+++ b/dot_local/bin/executable_ocr
@@ -0,0 +1,39 @@
+#!/usr/bin/env sh
+# OCR a screen region (default) or an image file → clipboard.
+#
+# Usage:
+#   ocr           # interactive: select a region with slurp, OCR it
+#   ocr <file>    # OCR an image file
+#
+# Requires: tesseract (+tesseract-data-eng, tesseract-data-por),
+#           grim, slurp, wl-clipboard, libnotify.
+#
+# Override languages via TESSERACT_LANG (e.g. TESSERACT_LANG=eng).
+
+set -eu
+
+lang="${TESSERACT_LANG:-eng+por}"
+
+if [ "${1:-}" ]; then
+  [ -r "$1" ] || { notify-send -u critical "📋 OCR" "Cannot read: $1"; exit 1; }
+  text="$(tesseract "$1" - -l "$lang" 2>/dev/null || true)"
+else
+  region="$(slurp 2>/dev/null)" || exit 0
+  text="$(grim -g "$region" - | tesseract - - -l "$lang" 2>/dev/null || true)"
+fi
+
+# Trim trailing whitespace per line; collapse runs of blank lines; drop
+# leading blanks.
+text="$(printf '%s\n' "$text" | awk '
+  { sub(/[[:space:]]+$/, "") }
+  NF { print; blank = 0; next }
+  !blank { print; blank = 1 }
+' | sed -e '/./,$!d')"
+
+if [ -z "$text" ]; then
+  notify-send -u low "📋 OCR" "No text detected."
+  exit 1
+fi
+
+printf '%s' "$text" | wl-copy
+notify-send -t 3000 "📋 OCR copied" "$(printf '%s' "$text" | head -c 200)"
-- 
cgit v1.3.1