lfp

commit 03ee7bbb016621559434e19f72daba45383ff954

Author: Connor Etherington <[email protected]>

Date:   Wed Oct 25 10:35:45 2023 +0200

    Auto-Commit Update - 20231025

---

 usr/share/lfp/ocr/ocr.js | 6 +++---

 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/usr/share/lfp/ocr/ocr.js b/usr/share/lfp/ocr/ocr.js

index c87a373..e4af7b0 100755

--- a/usr/share/lfp/ocr/ocr.js

+++ b/usr/share/lfp/ocr/ocr.js

@@ -24,7 +24,9 @@ const sh = (cmd, opts) => new Promise((resolve, reject) => {

 const ocr = async (imagePath) => {

   const Type = await sh(`file -b --mime-type ${imagePath}`);

-  if (Type === 'PDF document') {

+  if (await sh(`file -bL --mime-encoding ${imagePath}`) === 'utf-8') {

+    return await fsp.readFile(imagePath, 'utf8');

+  } else if (Type === 'PDF document') {

     await sh(`convert ${imagePath} -background white ${imagePath}.png`);

     const res = await tesseract.recognize(`${imagePath}.png`, config);

     await fs.unlinkSync(`${imagePath}.png`);

@@ -32,8 +34,6 @@ const ocr = async (imagePath) => {

   } else if(Type === 'PNG image data' || Type === 'JPEG image data' || Type === 'TIFF image data' || Type === 'image/x-portable-bitmap' || Type === 'image/png' || Type === 'image/jpeg' || Type === 'image/tiff' || Type === 'PC bitmap') {

     const image = await fsp.readFile(imagePath);

     return await tesseract.recognize(image, config);

-  } else if (Type === 'Unicode text') {

-    return await fsp.readFile(imagePath, 'utf8');

   } else {

     throw new Error('Unknown image type');

   }