From 8979b754ddbac600f8e5744ffe96b9923505bc45 Mon Sep 17 00:00:00 2001
From: Shihaam Abdul Rahman <shihaam@shihaam.me>
Date: Sat, 3 Aug 2024 00:20:16 +0500
Subject: [PATCH] amount, currency, date, reference, account works

---
 ocr.sh     | 30 +++++++++++++++++++++++++++++-
 ocr_raw.sh | 12 ++++++++++++
 2 files changed, 41 insertions(+), 1 deletion(-)
 create mode 100755 ocr_raw.sh

diff --git a/ocr.sh b/ocr.sh
index e206b0b..344b837 100755
--- a/ocr.sh
+++ b/ocr.sh
@@ -2,4 +2,32 @@
 
 
 INPUT_IMAGE=$1
-tesseract $INPUT_IMAGE stdout
+
+raw_text=$(tesseract $INPUT_IMAGE stdout)
+
+referece=$(echo "$raw_text" | grep -o 'BLAZ[0-9]*')
+to_number=$(echo "$raw_text" | grep -oE '\b9[0-9]{16}\b|\b7[0-9]{12}\b')
+date=$(echo "$raw_text" | grep -oE '\b[0-3][0-9]/[0-1][0-9]/[0-9]{4}\b')
+time=$(echo "$raw_text" | grep -oE '\b[0-2][0-9]:[0-5][0-9]\b')
+currency=$(echo "$raw_text" | grep -oP '(?<=Amount\s)[A-Z]{3}')
+amount=$(echo "$raw_text" | grep -oP '(?<=Amount\s[A-Z]{3}\s)[0-9]{1,3}(?:,[0-9]{3})*(?:\.[0-9]{2})')
+
+json_output=$(cat <<EOF
+{
+  "Amount": "$amount",
+  "Currency": "$currency",
+  "Date": "$date $time",
+  "From": "$from",
+  "Message": "$message",
+  "referece": "$referece",
+  "Remarks": "$remarks",
+  "Status": "$status",
+  "To": {
+    "Account": "$to_number",
+    "Name": "$to_name"
+  }
+}
+EOF
+)
+
+echo $json_output
diff --git a/ocr_raw.sh b/ocr_raw.sh
new file mode 100755
index 0000000..9fc8882
--- /dev/null
+++ b/ocr_raw.sh
@@ -0,0 +1,12 @@
+#!/bin/bash
+
+
+INPUT_IMAGE=$1
+
+raw_text=$(tesseract $INPUT_IMAGE stdout)
+
+echo $raw_text
+echo ""
+echo +++real raw+++
+echo ""
+tesseract $INPUT_IMAGE stdout