Merge branch 'dev' into dev

kRHYME7 · web-flow · commit 90a31ac9ca11 · 2025-09-10T23:48:22.000+08:00
diff --git a/.github/workflows/lock.yml b/.github/workflows/lock.yml
@@ -0,0 +1,23 @@
+name: 'Close stale issues and PR'
+on:
+  schedule:
+    - cron: '30 1 * * *'
+
+jobs:
+  stale:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/stale@v9
+        with:
+          github-token: ${{secrets.GITHUB_TOKEN}}
+          stale-issue-message: 'This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.'
+          stale-issue-label: 'no-issue-activity'
+          stale-pr-message: 'This PR is stale because it has been open 45 days with no activity. Remove stale label or comment or this will be closed in 10 days.'
+          stale-pr-label: 'no-pr-activity'
+          close-issue-message: 'This issue was closed because it has been stalled for 5 days with no activity.'
+          days-before-issue-stale: 30
+          days-before-pr-stale: 45
+          days-before-issue-close: 5
+          days-before-pr-close: 10
+          exempt-pr-labels: 'awaiting-approval,work-in-progress'
+          exempt-issue-labels: 'awaiting-approval,work-in-progress,enhancement,bug,autorelease: pending'
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -6,10 +6,20 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 
 ## Changes
 
+### Changed
+
+- OCR: `imagemagick` screenshot preprocessing tuned for better recognition results
+
 ### Added
 
 - Turkish documentation.
 - No changes have been made to other codes.
+- OCR: `tesseract` now supports explicit language settings via `hyde/config.toml`:
+    ```toml
+    [screenshot.ocr]
+    tesseract_languages = ["eng"]
+    ```
+    To use text recognition bind `hyde-shell screenshot sc` to any hotkey.
 
 ## v25.9.1 (Unreleased)
 
diff --git a/Configs/.local/lib/hyde/screenshot.sh b/Configs/.local/lib/hyde/screenshot.sh
@@ -51,6 +51,9 @@ save_dir="${2:-$XDG_PICTURES_DIR/Screenshots}"
 save_file=$(date +'%y%m%d_%Hh%Mm%Ss_screenshot.png')
 annotation_tool=${SCREENSHOT_ANNOTATION_TOOL}
 annotation_args=("-o" "${save_dir}/${save_file}" "-f" "${temp_screenshot}")
+tesseract_default_language=("eng")
+tesseract_languages=("${SCREENSHOT_OCR_TESSERACT_LANGUAGES[@]:-${tesseract_default_language[@]}}")
+tesseract_languages+=("osd")
 
 if [[ -z "$annotation_tool" ]]; then
 	pkg_installed "swappy" && annotation_tool="swappy"
@@ -90,6 +93,54 @@ take_screenshot() {
 	fi
 }
 
+ocr_screenshot() {
+	local mode=$1
+	shift
+	local extra_args=("$@")
+
+	# execute grimblast with given args
+	if "$LIB_DIR/hyde/grimblast" "${extra_args[@]}" copysave "$mode" "$temp_screenshot"; then
+		if pkg_installed imagemagick; then
+			magick "${temp_screenshot}" \
+				-colorspace gray \
+				-contrast-stretch 0 \
+				-level 15%,85% \
+				-resize 400% \
+				-sharpen 0x1 \
+				-auto-threshold triangle \
+				-morphology close diamond:1 \
+				-deskew 40% \
+				"${temp_screenshot}"
+		else
+			notify-send -a "HyDE Alert" "OCR: imagemagick is not installed, recognition accuracy is reduced" -e -i "dialog-warning"
+		fi
+		tesseract_package_prefix="tesseract-data-"
+		tesseract_packages=("${tesseract_languages[@]/#/$tesseract_package_prefix}")
+		tesseract_packages+=("tesseract")
+		for pkg in "${tesseract_packages[@]}"; do
+			if ! pkg_installed "${pkg}"; then
+				notify-send -a "HyDE Alert" "$(echo -e "OCR: required package is not installed\n ${pkg}")" -e -i "dialog-error"
+				return 1
+			fi
+		done
+		tesseract_languages_prepared=$(IFS=+; echo "${tesseract_languages[*]}")
+		tesseract_output=$(tesseract \
+			--psm 6 \
+			--oem 3 \
+			-l ${tesseract_languages_prepared} \
+			"${temp_screenshot}" \
+			stdout
+			2>/dev/null
+		)
+		printf "%s" "$tesseract_output" | wl-copy
+		notify-send -a "HyDE Alert" "$(echo -e "OCR: ${#tesseract_output} symbols recognized\n\nLanguages used ${tesseract_languages[@]/#/'\n '}")" -i "${temp_screenshot}" -e
+		rm -f "${temp_screenshot}"
+	else
+		notify-send -a "HyDE Alert" "OCR: screenshot error" -e -i "dialog-error"
+		return 1
+	fi
+}
+
 pre_cmd
 
 case $1 in
@@ -106,23 +157,14 @@ m) # print focused monitor
 	take_screenshot "output"
 	;;
 sc) #? 󱉶 Use 'tesseract' to scan image then add to clipboard
-	check_package tesseract-data-eng tesseract
-	if ! GEOM=$(slurp); then
-		notify-send -a "HyDE Alert" "OCR preview: Invalid geometry" -e -i "dialog-error"
-		exit 1
-	fi
-	grim -g "${GEOM}" "${temp_screenshot}"
-	pkg_installed imagemagick && magick "${temp_screenshot}" -sigmoidal-contrast 10,50% "${temp_screenshot}"
-	tesseract "${temp_screenshot}" - | wl-copy
-	notify-send -a "HyDE Alert" "OCR preview" -i "${temp_screenshot}" -e
-	rm -f "${temp_screenshot}"
+	ocr_screenshot "area" "--freeze"
 	;;
 *) # invalid option
 	USAGE
 	;;
 esac
 
-[ -f "$temp_screenshot" ] && rm "$temp_screenshot"
+[ -f "${temp_screenshot}" ] && rm "${temp_screenshot}"
 
 if [ -f "${save_dir}/${save_file}" ]; then
 	notify-send -a "HyDE Alert" -i "${save_dir}/${save_file}" "saved in ${save_dir}"
diff --git a/Configs/.local/share/hyde/config.toml b/Configs/.local/share/hyde/config.toml
@@ -117,6 +117,10 @@ annotation_tool = "satty"  # Annotation tool for screenshots.
 annotation_pre_command = []  # Pre command for annotation tool.
 annotation_post_command = [""]  # Post command for annotation tool.
 
+# OCR configuration.
+[screenshot.ocr]
+tesseract_languages = ["eng"]  # Place desired languages to use for text recognition. To see installed languages run `tesseract --list-langs`.
+
 # Bookmark configuration.
 [rofi.bookmark]
 
diff --git a/Configs/.local/share/hyde/schema/config.toml b/Configs/.local/share/hyde/schema/config.toml
@@ -417,6 +417,15 @@ type      = "object"
 [properties.screenshot.properties.annotation_post_command.items]
     type = "string"
 
+[properties."screenshot.ocr"]
+    description = "OCR configuration."
+    type        = "object"
+
+[properties."screenshot.ocr".properties.tesseract_languages]
+    default     = [ "eng" ]
+    description = "Place desired languages to use for text recognition. To see installed languages run `tesseract --list-langs`."
+    type        = "array"
+
 [properties."rofi.bookmark"]
     description = "Bookmark configuration."
     type        = "object"
diff --git a/Configs/.local/share/hyde/schema/config.toml.json b/Configs/.local/share/hyde/schema/config.toml.json
@@ -457,6 +457,19 @@
                 }
             }
         },
+        "screenshot.ocr": {
+            "description": "OCR configuration.",
+            "type": "object",
+            "properties": {
+                "tesseract_languages": {
+                    "default": [
+                       "eng"
+                    ],
+                    "description": "Place desired languages to use for text recognition. To see installed languages run `tesseract --list-langs`.",
+                    "type": "array"
+                }
+            }
+        },
         "rofi.bookmark": {
             "description": "Bookmark configuration.",
             "type": "object",
diff --git a/Scripts/migrations/v25.9.1.sh b/Scripts/migrations/v25.9.1.sh
@@ -1,7 +1,7 @@
 #!/usr/bin/env sh
 
-echo "Replacing rofi-wayland to rofi"
+echo "Please be sure to update rofi"
+echo "Arch repo now support rofi-wayland inside rofi package"
+echo "If you are using arch linux or an arch based distro please remove rofi-wayland and install rofi"
+echo "sudo pacman -Syu should also do the trick"
 
-sudo pacman -Sy
-sudo pacman -Rns --noconfirm rofi-wayland
-sudo pacman -S --noconfirm rofi
