From ebafd93110133bdd03c4d64f6b901c3dc6035535 Mon Sep 17 00:00:00 2001
From: tretrauit <tretrauit@gmail.com>
Date: Sun, 7 Jan 2024 18:15:51 +0700
Subject: [PATCH] fix(katana): add workaround for text parsing

---
 swordfish/src/katana.rs | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/swordfish/src/katana.rs b/swordfish/src/katana.rs
index 718111b..6efce8e 100644
--- a/swordfish/src/katana.rs
+++ b/swordfish/src/katana.rs
@@ -57,6 +57,10 @@ fn fix_tesseract_string(text: &mut String) {
         text.remove(0);
         text.remove(0);
     }
+    // Remove the first character if it is not alphanumeric
+    if !TEXT_NUM_REGEX.is_match(text.clone().chars().nth(0).unwrap().to_string().as_str()) {
+        text.remove(0);
+    }
     // Workaround IR -> Ik
     // Maybe it only occurs if Ik is in the start of the string?
     // e.g. "IReda" -> "Ikeda"