From 10e6f489d0eaf0b40877a1f31cd28af7540134f4 Mon Sep 17 00:00:00 2001
From: Benoit Pierre <benoit.pierre@gmail.com>
Date: Sat, 1 Jun 2024 00:25:59 +0200
Subject: [PATCH] kopt: honor `TESSDATA_PREFIX` environment variable

Don't override it by default, but honor it if present.
---
 frontend/document/koptinterface.lua | 5 ++++-
 platform/android/llapp_main.lua     | 3 ---
 platform/cervantes/koreader.sh      | 3 ---
 platform/kindle/koreader.sh         | 3 ---
 platform/kobo/koreader.sh           | 3 ---
 platform/kobo/nickel.sh             | 2 +-
 platform/pocketbook/koreader.app    | 3 ---
 platform/remarkable/koreader.sh     | 3 ---
 platform/ubuntu-touch/koreader.sh   | 3 ---
 9 files changed, 5 insertions(+), 23 deletions(-)

diff --git a/frontend/document/koptinterface.lua b/frontend/document/koptinterface.lua
index c9fd8eb01..971e9a503 100644
--- a/frontend/document/koptinterface.lua
+++ b/frontend/document/koptinterface.lua
@@ -20,7 +20,10 @@ local util = require("util")
 
 local KoptInterface = {
     ocrengine = "ocrengine",
-    tessocr_data = DataStorage:getDataDir() .. "/data",
+    -- If `$TESSDATA_PREFIX` is set, don't override it: let libk2pdfopt honor it
+    -- (which includes checking for data in both `$TESSDATA_PREFIX/tessdata` and
+    -- in `$TESSDATA_PREFIX/` on more recent versions).
+    tessocr_data = not os.getenv('TESSDATA_PREFIX') and DataStorage:getDataDir().."/data" or nil,
     ocr_lang = "eng",
     ocr_type = 3, -- default 0, for more accuracy use 3
     last_context_size = nil,
diff --git a/platform/android/llapp_main.lua b/platform/android/llapp_main.lua
index 1242f6f55..c71e6ce58 100644
--- a/platform/android/llapp_main.lua
+++ b/platform/android/llapp_main.lua
@@ -16,9 +16,6 @@ end
 -- path to primary external storage partition
 local path = android.getExternalStoragePath()
 
--- set TESSDATA_PREFIX env var
-C.setenv("TESSDATA_PREFIX", path.."/koreader/data", 1)
-
 -- create fake command-line arguments
 -- luacheck: ignore 121
 if android.isDebuggable() then
diff --git a/platform/cervantes/koreader.sh b/platform/cervantes/koreader.sh
index ac9deb3da..d82a38489 100755
--- a/platform/cervantes/koreader.sh
+++ b/platform/cervantes/koreader.sh
@@ -53,9 +53,6 @@ if [ -n "${fail}" ] && [ "${fail}" -eq 0 ]; then
     exec ./koreader.sh "$@"
 fi
 
-# export trained OCR data directory
-export TESSDATA_PREFIX="data"
-
 # export dict directory
 export STARDICT_DATA_DIR="data/dict"
 
diff --git a/platform/kindle/koreader.sh b/platform/kindle/koreader.sh
index eb62de0c0..25b99a3e1 100755
--- a/platform/kindle/koreader.sh
+++ b/platform/kindle/koreader.sh
@@ -177,9 +177,6 @@ if [ -n "${fail}" ] && [ "${fail}" -eq 0 ]; then
     exec ./koreader.sh ${REEXEC_FLAGS} "${@}"
 fi
 
-# export trained OCR data directory
-export TESSDATA_PREFIX="data"
-
 # export dict directory
 export STARDICT_DATA_DIR="data/dict"
 
diff --git a/platform/kobo/koreader.sh b/platform/kobo/koreader.sh
index 6f8c7630b..75a1ef623 100755
--- a/platform/kobo/koreader.sh
+++ b/platform/kobo/koreader.sh
@@ -146,9 +146,6 @@ if [ -n "${fail}" ] && [ "${fail}" -eq 0 ]; then
     exec ./koreader.sh "${@}"
 fi
 
-# export trained OCR data directory
-export TESSDATA_PREFIX="data"
-
 # export dict directory
 export STARDICT_DATA_DIR="data/dict"
 
diff --git a/platform/kobo/nickel.sh b/platform/kobo/nickel.sh
index a89700bb2..7acf81f39 100755
--- a/platform/kobo/nickel.sh
+++ b/platform/kobo/nickel.sh
@@ -13,7 +13,7 @@ export QT_GSTREAMER_PLAYBIN_AUDIOSINK_DEVICE_PARAMETER=bluealsa:DEV=00:00:00:00:
 # shellcheck disable=SC2164
 cd /
 unset OLDPWD
-unset LC_ALL TESSDATA_PREFIX STARDICT_DATA_DIR EXT_FONT_DIR
+unset LC_ALL STARDICT_DATA_DIR EXT_FONT_DIR
 unset KO_DONT_GRAB_INPUT
 unset FBINK_FORCE_ROTA
 
diff --git a/platform/pocketbook/koreader.app b/platform/pocketbook/koreader.app
index 274720ca0..6a79f1574 100755
--- a/platform/pocketbook/koreader.app
+++ b/platform/pocketbook/koreader.app
@@ -78,9 +78,6 @@ ko_update_check() {
 # we're always starting from our working directory
 cd ${KOREADER_DIR} || exit
 
-# export trained OCR data directory
-export TESSDATA_PREFIX="data"
-
 # export dict directory
 export STARDICT_DATA_DIR="data/dict"
 
diff --git a/platform/remarkable/koreader.sh b/platform/remarkable/koreader.sh
index e4fa7f4dd..b271a6ce0 100755
--- a/platform/remarkable/koreader.sh
+++ b/platform/remarkable/koreader.sh
@@ -72,9 +72,6 @@ if [ -n "${fail}" ] && [ "${fail}" -eq 0 ]; then
     exec ./koreader.sh "${@}"
 fi
 
-# export trained OCR data directory
-export TESSDATA_PREFIX="data"
-
 # export dict directory
 export STARDICT_DATA_DIR="data/dict"
 
diff --git a/platform/ubuntu-touch/koreader.sh b/platform/ubuntu-touch/koreader.sh
index 34c1f6949..f1d0d65f0 100755
--- a/platform/ubuntu-touch/koreader.sh
+++ b/platform/ubuntu-touch/koreader.sh
@@ -7,9 +7,6 @@ KOREADER_DIR="${0%/*}"
 # we're always starting from our working directory
 cd "${KOREADER_DIR}" || exit
 
-# export trained OCR data directory
-export TESSDATA_PREFIX="data"
-
 # export external font directory
 export EXT_FONT_DIR="${HOME}/fonts"