refactor app.py

app.py

@@ -236,16 +236,16 @@ EXAMPLES_TEST = [
 
 DESC_DICT = {
     "concatenate_pages": "Whether to merge pages",
-    "use_formula_recognition": "Whether to use formula recognition subpipeline. If used, the formula can be converted into Latex code. Otherwise, the formula part is the text recognition result.",
+    "use_formula_recognition": "Enable the formula recognition subpipeline. When enabled, formulas will be converted to LaTeX code; otherwise, they will be recognized as plain text.",
     "use_chart_recognition": "Use the PP-Chart2Table model to parse and convert the charts in the document into tables.",
-    "use_doc_orientation_classify": "Whether to use the document image orientation classification module. After use, you can correct distorted images, such as wrinkles, tilts, etc.",
-    "use_doc_unwarping": "Whether to use the document unwarping module. After use, you can correct distorted images, such as wrinkles, tilts, etc.",
-    "use_textline_orientation": "Whether to use the text line orientation classification module to support the distinction and correction of text lines of 0 degrees and 180 degrees.",
-    "use_region_detection": "Whether to use the layout region detection. After using it, it can handle complex layouts such as newspapers and magazines.",
-    "use_seal_recognition": "Whether to use seal text recognition subpipeline. After use, the seal text content in the document can be extracted.",
-    "use_table_recognition": "Whether to use table recognition subpipeline. If used, the table can be identified as a structured format (such as HTML). Otherwise, the table will be regarded as figure.",
+    "use_doc_orientation_classify": "Enable the document image orientation classification module. When enabled, you can correct distorted images, such as wrinkles, tilts, etc.",
+    "use_doc_unwarping": "Enable the document unwarping module. When enabled, you can correct distorted images, such as wrinkles, tilts, etc.",
+    "use_textline_orientation": "Enable the text line orientation classification module to support the distinction and correction of text lines of 0 degrees and 180 degrees.",
+    "use_region_detection": "Enable the layout region detection. When enabled, it can handle complex layouts such as newspapers and magazines.",
+    "use_seal_recognition": "Enable the seal text recognition subpipeline. When enabled, seals will be recognized as text; otherwise, they will be treated as figures.",
+    "use_table_recognition": "Enable the table recognition subpipeline. When enabled, tables will be extracted in HTML format; otherwise, they will be treated as figures.",
     "layout_threshold_nb": "The threshold used to filter out low confidence prediction results for the layout region, ranging from 0 to 1. If there are missed regions, this value can be appropriately lowered.",
-    "layout_nms": "Whether to use layout region detection model uses NMS post-processing. After using it, nested boxes or those with large intersections can be removed.",
+    "layout_nms": "Enable layout region detection model uses NMS post-processing. When enabled, nested boxes or those with large intersections can be removed.",
     "layout_unclip_ratio_nb": "Use this method to expand each region of ​​the layout. The larger the value, the larger the expanded region.",
     "text_det_limit_type": "[Short side] means to ensure that the shortest side of the image is not less than [Image side length limit for text detection], and [Long side] means to ensure that the longest side of the image is not greater than [Image side length limit for text detection].",
     "text_det_limit_side_len_nb": "For the side length limit of the text detection input image, for large images with dense text, if you want more accurate recognition, you should choose a larger size. This parameter is used in conjunction with the [Image side length limit type for text detection]. Generally, the maximum [Long side] is suitable for scenes with large images and text, and the minimum [Short side] is suitable for document scenes with small and dense images.",