Minor changes

app.py

@@ -84,7 +84,7 @@ def crop_image(pil_img, detection_result, padding=30):
     width = image.shape[1]
     height = image.shape[0]
     # print(width, height)
-    for i, result in enumerate(detection_result):
+    for idx, result in enumerate(detection_result):
         class_id = int(result[5])
         score = float(result[4])
         min_x = result[0]
@@ -113,7 +113,7 @@ def crop_image(pil_img, detection_result, padding=30):
 
         crop_images.append(crop_image)
 
-        cv2.rectangle(image, (x1, y1), (x2, y2), color=(0, 0, 255))
+        cv2.rectangle(image, (x1, y1), (x2, y2), color=(0, 0, 255), thickness=2)
         cv2.putText(image, f'{score:.2f}', (x1, y1), cv2.FONT_HERSHEY_SIMPLEX, fontScale=0.5, color=(255, 0, 0))
 
     return crop_images, cv_to_PIL(image)
@@ -148,7 +148,7 @@ def convert_stucture(page_tokens, pil_img, structure_result):
     bboxes = []
     scores = []
     labels = []
-    for i, result in enumerate(structure_result):
+    for idx, result in enumerate(structure_result):
         class_id = int(result[5])
         score = float(result[4])
         min_x = result[0]
@@ -206,12 +206,12 @@ def visualize_ocr(pil_img, ocr_result):
     plt.gcf().set_size_inches(20, 20)
     ax = plt.gca()
 
-    for i, result in enumerate(ocr_result):
+    for idx, result in enumerate(ocr_result):
         bbox = result['bbox']
         text = result['text']
-        rect = patches.Rectangle(bbox[:2], bbox[2]-bbox[0], bbox[3]-bbox[1], linewidth=2, edgecolor='red', facecolor='none', linestyle="-")
+        rect = patches.Rectangle(bbox[:2], bbox[2]-bbox[0], bbox[3]-bbox[1], linewidth=2, edgecolor='red', facecolor='none', linestyle='-')
         ax.add_patch(rect)
-        ax.text(bbox[0], bbox[3], text, horizontalalignment='left', verticalalignment='bottom', color='blue', fontsize=10)
+        ax.text(bbox[2], bbox[3], text, horizontalalignment='left', verticalalignment='bottom', color='blue', fontsize=7)
 
     plt.xticks([], [])
     plt.yticks([], [])
@@ -224,17 +224,6 @@ def visualize_ocr(pil_img, ocr_result):
 
     return PIL.Image.open(img_buf)
 
-    # image = PIL_to_cv(pil_img)
-    # for i, res in enumerate(ocr_result):
-    #     bbox = res['bbox']
-    #     x1 = int(bbox[0])
-    #     y1 = int(bbox[1])
-    #     x2 = int(bbox[2])
-    #     y2 = int(bbox[3])
-    #     cv2.rectangle(image, (x1, y1), (x2, y2), color=(255, 0, 0))
-    #     cv2.putText(image, res['text'], (x1, y1), cv2.FONT_HERSHEY_SIMPLEX, fontScale=0.25, color=(0, 0, 255))
-    # return cv_to_PIL(image)
-
 
 def get_bbox_decorations(data_type, label):
     if label == 0:
@@ -266,7 +255,7 @@ def visualize_structure(pil_img, structure_result):
     plt.gcf().set_size_inches(20, 20)
     ax = plt.gca()
 
-    for i, result in enumerate(structure_result):
+    for idx, result in enumerate(structure_result):
         class_id = int(result[5])
         score = float(result[4])
         min_x = result[0]
@@ -301,7 +290,7 @@ def visualize_structure(pil_img, structure_result):
         rect = patches.Rectangle(bbox[:2], bbox[2]-bbox[0], bbox[3]-bbox[1],
                                     linewidth=linewidth,
                                     edgecolor=color, facecolor='none',
-                                    linestyle="--")
+                                    linestyle='--')
         ax.add_patch(rect)
 
     plt.xticks([], [])
@@ -311,7 +300,7 @@ def visualize_structure(pil_img, structure_result):
     for class_name in structure_class_names[:-1]:
         color, alpha, linewidth, hatch = get_bbox_decorations('recognition', structure_class_map[class_name])
         legend_elements.append(
-            Patch(facecolor=color, edgecolor=color, label=class_name, hatch=hatch, alpha=alpha)
+            Patch(facecolor='none', edgecolor=color, linewidth=linewidth, linestyle='--', label=class_name, hatch=hatch)
         )
 
     plt.legend(handles=legend_elements, bbox_to_anchor=(0.5, -0.02), loc='upper center', borderaxespad=0,
@@ -326,7 +315,7 @@ def visualize_structure(pil_img, structure_result):
 
 
 def visualize_cells(pil_img, cells):
-    plt.imshow(pil_img, interpolation="lanczos")
+    plt.imshow(pil_img, interpolation='lanczos')
     plt.gcf().set_size_inches(20, 20)
     ax = plt.gca()
 
@@ -618,7 +607,7 @@ def main():
                 str_cols[2].subheader('Structure result')
                 str_cols[3].subheader('Cells result')
 
-                for i, img in enumerate(crop_images):
+                for idx, img in enumerate(crop_images):
                     str_cols = st.columns(4)
 
                     vis_img = visualize_image(img)