update pptc as libreoffice impress

desktop_env/evaluators/metrics/slides.py

@@ -139,6 +139,7 @@ def compare_pptx_files(file1_path, file2_path, **options):
     examine_number_of_slides = options.get("examine_number_of_slides", True)
     examine_shape = options.get("examine_shape", True)
     examine_text = options.get("examine_text", True)
+    examine_indent = options.get("examine_indent", True)
     examine_font_name = options.get("examine_font_name", True)
     examine_font_size = options.get("examine_font_size", True)
     examine_font_bold = options.get("examine_font_bold", True)
@@ -148,6 +149,7 @@ def compare_pptx_files(file1_path, file2_path, **options):
     examine_strike_through = options.get("examine_strike_through", True)
     examine_alignment = options.get("examine_alignment", True)
     examine_bottom_position = options.get("examine_bottom_position", False)
+    examine_bullets = options.get("examine_bullets", True)
 
     # compare the number of slides
     if len(prs1.slides) != len(prs2.slides) and examine_number_of_slides:
@@ -174,6 +176,12 @@ def compare_pptx_files(file1_path, file2_path, **options):
                     if para1.alignment != para2.alignment and examine_alignment:
                         return 0
                     # check if the runs are the same
+                    if para1.text != para2.text and examine_text:
+                        return 0
+
+                    if para1.level != para2.level and examine_indent:
+                        return 0
+
                     for run1, run2 in zip(para1.runs, para2.runs):
 
                         # check if the font properties are the same
@@ -200,7 +208,40 @@ def compare_pptx_files(file1_path, file2_path, **options):
                                 'strike', 'noStrike') and examine_strike_through:
                             return 0
 
-                        # fixme: Actually there are more properties to be compared, but we cannot get them through pptx
+                        def _extract_bullets(xml_data):
+                            root = ET.fromstring(xml_data)
+
+                            namespaces = {
+                                'a': 'http://schemas.openxmlformats.org/drawingml/2006/main',
+                                'p': 'http://schemas.openxmlformats.org/presentationml/2006/main',
+                            }
+
+                            bullets = []
+
+                            for paragraph in root.findall('.//a:p', namespaces):
+                                pPr = paragraph.find('a:pPr', namespaces)
+                                if pPr is not None:
+                                    lvl = pPr.get('lvl')
+                                    buChar = pPr.find('a:buChar', namespaces)
+                                    char = buChar.get('char') if buChar is not None else "No Bullet"
+                                    buClr = pPr.find('a:buClr/a:srgbClr', namespaces)
+                                    color = buClr.get('val') if buClr is not None else "No Color"
+                                else:
+                                    lvl = "No Level"
+                                    char = "No Bullet"
+                                    color = "No Color"
+
+                                text = "".join(t.text for t in paragraph.findall('.//a:t', namespaces))
+
+                                bullets.append((lvl, char, text, color))
+
+                            return bullets
+
+                    if _extract_bullets(run1.part.blob.decode('utf-8')) != _extract_bullets(
+                            run2.part.blob.decode('utf-8')) and examine_bullets:
+                        return 0
+
+                    # fixme: Actually there are more properties to be compared, we can add them later via parsing the xml data
 
     return 1
 
@@ -422,6 +463,7 @@ if __name__ == '__main__':
     #     r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\550ce7e7-747b-495f-b122-acdc4d0b8e54\New_Club_Spring_2018_Training_Gold.pptx"))
     # print(evaluate_presentation_fill_to_rgb_distance(r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\3b27600c-3668-4abd-8f84-7bcdebbccbdb\lec17-gui-events.pptx", {"rgb": (0, 0, 255)}))
     # print(check_auto_saving_time(r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\2cd43775-7085-45d8-89fa-9e35c0a915cf\registrymodifications.xcu", {"minutes": 3}))
-    print(compare_pptx_files(r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\a669ef01-ded5-4099-9ea9-25e99b569840\Writing-Outlines.pptx",
-                             r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\a669ef01-ded5-4099-9ea9-25e99b569840\Writing-Outlines_Gold.pptx",
-                             examine_shape=False))
+    print(compare_pptx_files(
+        r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\a669ef01-ded5-4099-9ea9-25e99b569840\Writing-Outlines.pptx",
+        r"C:\Users\tianbaox\Desktop\DesktopEnv\cache\a669ef01-ded5-4099-9ea9-25e99b569840\Writing-Outlines_Gold.pptx",
+        examine_shape=False))