From 6d853ff27f1ff8db72f06a19a9748c215f705fd7 Mon Sep 17 00:00:00 2001
From: stormdragon2976 <stormdragon2976@gmail.com>
Date: Mon, 5 Dec 2022 20:07:17 -0500
Subject: [PATCH] Very experimental changes with clipboard_translator.sh, now
 requires sqlite3 and perl.

---
 audiogame-manager.sh           |  7 ++-
 speech/clipboard_translator.sh | 87 +++++++++++++++++++++-------------
 2 files changed, 60 insertions(+), 34 deletions(-)

diff --git a/audiogame-manager.sh b/audiogame-manager.sh
index ab56c7c..a306a24 100755
--- a/audiogame-manager.sh
+++ b/audiogame-manager.sh
@@ -739,8 +739,11 @@ game_launcher() {
         if [[ "$game" =~ laser-breakout ]]; then
             pgrep -u "$USER" nvda2speechd &> /dev/null || ${XDG_DATA_HOME:-$HOME/.local/share}/audiogame-manager/nvda2speechd &
         fi
-        if [[ "$game" =~ ^bokurano-daibouken\||bokurano-daibouken-2\| ]]; then
-            "${0%/*}/speech/clipboard_translator.sh" play.exe &
+        if [[ "$game" =~ ^bokurano-daibouken-2\| ]]; then
+            "${0%/*}/speech/clipboard_translator.sh" play.exe bokurano-daibouken2 &
+        fi
+        if [[ "$game" =~ ^bokurano-daibouken\| ]]; then
+            "${0%/*}/speech/clipboard_translator.sh" play.exe bokurano-daibouken &
         fi
         if [[ "$game" =~ shadow-line ]]; then
             find "${WINEPREFIX}/drive_c/" -type f -name 'nvdaControllerClient.dll' -exec rm -v "{}" \;
diff --git a/speech/clipboard_translator.sh b/speech/clipboard_translator.sh
index 3445913..4122723 100755
--- a/speech/clipboard_translator.sh
+++ b/speech/clipboard_translator.sh
@@ -1,54 +1,77 @@
 #!/usr/bin/env bash
 
-cleanup() {
-    sort -uo "$dictionaryFile" "$dictionaryFile"
-}
+# Modified from the script at:
+#  https://gist.github.com/fdietze/6768a0970d7d732b7fbd7930ccceee2a
 
-trap cleanup EXIT
+set -Eeuo pipefail # https://vaneyckt.io/posts/safer_bash_scripts_with_set_euxo_pipefail/#:~:text=set%20%2Du,is%20often%20highly%20desirable%20behavior.
+shopt -s expand_aliases
 
-# Clear the clipboard so you don't accidently send personal info to the translator.
-echo "" | xclip -d "${DISPLAY:-:0}" -selection clipboard 2> /dev/null
-export dictionaryFile="${XDG_CONFIG_HOME:-$HOME/.config}/storm-games/audiogame-manager/translations.txt"
-if [[ ! -r "$dictionaryFile" ]]; then
-    touch "$dictionaryFile"
-fi
 
-if [[ $# -ne 1 ]]; then
-    echo "Usage: $0 application name."
+if [[ $# -ne 2 ]]; then
+    echo "Usage: $0 \"application name\" \"file name\"."
+    exit 1
 fi
 
 # Wait for the application to start
 while ! pgrep -u "$USER" ^$1 &> /dev/null ; do
     sleep 0.05
 done
+    
+fileName="${2,,}"
+fileName="${fileName//[[:space:]]/-}.sqlite"
+translationFile="${XDG_CACHE_HOME:-$HOME/.cache}/audiogame-manager/${fileName}"
 
 # Read so long as the application is running
 while pgrep -u "$USER" ^$1 &> /dev/null ; do
-    newText=""
-    translatedText=""
-    newText="$(xclip -d "${DISPLAY:-:0}" -selection clipboard -o 2> /dev/null)"
-    sleep 0.01
-    if [[ "${#newText}" -lt 1 ]]; then
-        continue
-    fi
-    if [[ "${newText}" =~ ^[0-9A-Za-z[:space:][:punct:]]+$ ]]; then
-        spd-say -- "$newText"
+    sleep 0.05
+    text="$(xclip -d "${DISPLAY:-:0}" -selection clipboard -o 2> /dev/null)"
+    if [[ "${text}" =~ ^[0-9A-Za-z[:space:][:punct:]]+$ ]]; then
+        spd-say -- "$text"
         echo "" | xclip -d "${DISPLAY:-:0}" -selection clipboard 2> /dev/null
         continue
     fi
-    export newText
-    translatedText="$(awk -v originalText="${newText})==:" 'BEGIN { exitCode = 1; FS = "\\)==:\\[" }
-        $0~originalText { print $2; exitCode = 0; exit}
-        END { exit exitCode }' "$dictionaryFile")"
-    if [[ "${#translatedText}" -ge 1 ]]; then
-        spd-say -- "$translatedText"
-        echo "" | xclip -d "${DISPLAY:-:0}" -selection clipboard 2> /dev/null
+    if [[ -z "$text" ]]; then
         continue
     fi
-    translatedText="$(trans -no-autocorrect -b -- "$newText")"
-    if ! [[ "${newText}" =~ ^[0-9[:punct:]]+$ ]]; then
-        echo "${newText})==:[${translatedText}" >> "$dictionaryFile"
+    q() { sqlite3 -column -init "" "$translationFile" "$@" 2> /dev/null ; }
+
+
+    if [[ ! -s "$translationFile" ]]; then
+        rm -f "$translationFile"
+        cat << EOF | sqlite3 -init "" "$translationFile" 2> /dev/null
+.bail on
+
+CREATE TABLE translations(
+  text TEXT NOT NULL,
+  translation TEXT NOT NULL,
+
+  PRIMARY KEY (text)
+);
+CREATE INDEX translations_text_idx ON translations (text);
+EOF
     fi
-    spd-say -- "$translatedText"
+
+
+    # https://en.wikipedia.org/wiki/Unicode_equivalence#Combining_and_precomposed_characters
+    # https://www.effectiveperlprogramming.com/2011/09/normalize-your-perl-source/
+    alias nfc="perl -MUnicode::Normalize -CS -ne 'print NFC(\$_)'" # composed characters
+
+    # Normalize different unicode space characters to the same space
+    # https://stackoverflow.com/a/43640405
+    alias normalize_spaces="perl -CSDA -plE 's/[^\\S\\t]/ /g'"
+    alias normalize_unicode="normalize_spaces | nfc"
+
+    textEscaped="$(echo "$text" | sed "s/'/''/g" | normalize_unicode)" # escape single quotes for sqlite
+    translated="$(q "SELECT translation FROM translations WHERE text = '$textEscaped' LIMIT 1" | sed 's/\s*$//')"
+
+    if [[ -z "$translated" ]]; then
+        translated="$(trans -no-autocorrect -no-warn -brief "$text" | head -1 | sed 's/\s*$//' | normalize_unicode)"
+        if [[ -n "$translated" ]]; then
+            translatedEscaped="$(echo "$translated" | sed "s/'/''/g")"
+            q "INSERT OR IGNORE INTO translations (text, translation) VALUES ('$textEscaped', '$translatedEscaped')"
+        fi
+    fi
+
+    spd-say -- "$translated"
     echo "" | xclip -d "${DISPLAY:-:0}" -selection clipboard 2> /dev/null
 done