Update PPOCRLabel

Update PPOCRLabel

PPOCRLabel/PPOCRLabel.py

@@ -2336,13 +2336,6 @@ class MainWindow(QMainWindow):
                     else:
                         labeldict[file] = []
 
-        # if len(labeldict) != len(csv_paths):
-        #     msg = 'ERROR, box label and excel label are not in the same number\n' + \
-        #           'box label: ' + str(len(labeldict)) + '\n' + \
-        #           'excel label: ' + str(len(csv_paths)) + '\n' + \
-        #           'Please check the label.txt and tableRec_excel_output\n'
-        #     QMessageBox.information(self, "Information", msg)
-        #     return
         train_split, val_split, test_split = partitionDialog.getDataPartition()
         # check validate
         if train_split + val_split + test_split > 100:
@@ -2365,14 +2358,8 @@ class MainWindow(QMainWindow):
             filename, _ = os.path.splitext(os.path.basename(image_path))
             csv_path = os.path.join(TableRec_excel_dir, filename + '.xlsx')
             if not os.path.exists(csv_path):
-                msg = 'ERROR, Can not find ' + csv_path
-                QMessageBox.information(self, "Information", msg)
-                return
+                continue
 
-            # read xlsx file, convert to HTML
-            # xd = pd.ExcelFile(csv_path)
-            # df = xd.parse()
-            # structure = df.to_html(index = False)
             excel = xlrd.open_workbook(csv_path)
             sheet0 = excel.sheet_by_index(0)  # only sheet 0
             merged_cells = sheet0.merged_cells # (0,1,1,3) start row, end row, start col, end col
@@ -2384,7 +2371,6 @@ class MainWindow(QMainWindow):
 
             token_list = convert_token(html_list)
 
-
             # load box annotations
             cells = []
             for anno in labeldict[image_path]:

PPOCRLabel/libs/utils.py

@@ -221,10 +221,10 @@ def convert_token(html_list):
                 token_list.append("<td")
                 if 'colspan' in col:
                     _, n = col.split('colspan=')
-                    token_list.append(" colspan=\"{}\"".format(n))
+                    token_list.append(" colspan=\"{}\"".format(n[0]))
                 if 'rowspan' in col:
                     _, n = col.split('rowspan=')
-                    token_list.append(" rowspan=\"{}\"".format(n))
+                    token_list.append(" rowspan=\"{}\"".format(n[0]))
                 token_list.extend([">", "</td>"])
         token_list.append("</tr>")
     token_list.append("</tbody>")

PPOCRLabel/resources/strings/strings-en.properties

@@ -112,4 +112,4 @@ keyDialogTip=Enter object label
 keyChange=Change Box Key
 TableRecognition=Table Recognition
 cellreRecognition=Cell Re-Recognition
-exportJSON=export JSON(PubTabNet)
+exportJSON=Export Excel Label(PubTabNet)

PPOCRLabel/resources/strings/strings-zh-CN.properties

@@ -84,7 +84,7 @@ mhelp=帮助
 iconList=缩略图
 detectionBoxposition=检测框位置
 recognitionResult=识别结果
-creatPolygon=多边形标注
+creatPolygon=多点标注
 drawSquares=正方形标注
 rotateLeft=图片左旋转90度
 rotateRight=图片右旋转90度