Update app.py

app.py

@@ -5,7 +5,6 @@ import csv
 import os
 import cv2
 import numpy as np
-import pandas as pd
 import easyocr
 import keras_ocr
 from paddleocr import PaddleOCR
@@ -79,27 +78,36 @@ def generate_ocr(method, img):
 
     return text_output, label
 
-# Save extracted text to JSON & CSV
+# Save extracted text to JSON & CSV (in the specified repository)
 def save_results(text, label):
     data = {"text": text, "label": label}
 
+    # Ensure the repository exists
+    repo_path = "./wnmnd/ocr-llm-test"
+    if not os.path.exists(repo_path):
+        os.makedirs(repo_path)
+
+    # Define the full file paths for JSON and CSV
+    results_json = os.path.join(repo_path, "ocr_results.json")
+    results_csv = os.path.join(repo_path, "ocr_results.csv")
+
     # Save to JSON
-    if not os.path.exists(RESULTS_JSON):
-        with open(RESULTS_JSON, "w") as f:
-            json.dump([], f)  # Initialize an empty list if file doesn't exist
-    with open(RESULTS_JSON, "r+") as f:
-        content = json.load(f)  # Load current content
-        content.append(data)  # Append new data
-        f.seek(0)  # Go back to the start of the file
-        json.dump(content, f, indent=4)  # Save updated content
+    if not os.path.exists(results_json):
+        with open(results_json, "w") as f:
+            json.dump([], f)
+    with open(results_json, "r+") as f:
+        content = json.load(f)
+        content.append(data)
+        f.seek(0)
+        json.dump(content, f, indent=4)
 
     # Save to CSV
-    file_exists = os.path.exists(RESULTS_CSV)
-    with open(RESULTS_CSV, "a", newline="") as f:
+    file_exists = os.path.exists(results_csv)
+    with open(results_csv, "a", newline="") as f:
         writer = csv.DictWriter(f, fieldnames=["text", "label"])
         if not file_exists:
-            writer.writeheader()  # Write header if file is new
-        writer.writerow(data)  # Append new data
+            writer.writeheader()
+        writer.writerow(data)
 
 # Gradio Interface
 image_input = gr.Image()