epub converter: update css processing

src/util/css_reader.py

@@ -1,26 +1,23 @@
 import re
-
-from itertools import takewhile
-
 import cssutils
+
 from bs4 import BeautifulSoup
 from ebooklib import epub
 from premailer import transform
+from itertools import takewhile
+from logging import CRITICAL
 
 from src.config import LawCartaConfig
 
-
+cssutils.log.setLevel(CRITICAL)
-def convert_font_property(property):
-    return ''
-
 
 sizes_pr = [-1, 0.5, 0.56, 0.63, 0.69, 0.75, 0.81, 0.88, 0.94, 1.0, 1.06, 1.13, 1.19, 1.25, 1.31, 1.38, 1.44, 1.5, 1.56,
             1.63, 1.69, 1.75, 1.81, 1.88, 1.94, 2.0, 2.06, 2.13, 2.19, 2.25, 2.31, 2.38, 2.44, 2.5, 2.56, 2.63, 2.69,
             2.75, 2.81, 2.88, 2.94, 3.0, 4.0, 5.0]
 
 sizes_px = ['10px', '10px', '11px', '12px', '13px', '14px', '15px', '16px', '17px', '18px', '19px', '20px', '21px',
-            '22px',
+            '22px', '23px', '24px', '25px', '26px', '27px', '28px', '29px', '30px', '31px', '32px', '33px', '34px',
-            '23px', '24px', '25px', '26px', '27px', '28px', '29px', '30px', '31px', '32px', '33px', '34px', '35px',
+            '35px',
             '36px', '37px', '38px', '39px', '40px', '41px', '42px', '43px', '44px', '45px', '46px', '47px', '48px',
             '49px', '50px', '64px', '72px']
 
@@ -75,28 +72,38 @@ LIVECARTA_STYLE_ATTRS_MAPPING = {
     'text-indent': lambda x: LawCartaConfig.INDENT,
     'font-variant': lambda x: x,
     'text-align': lambda x: x,
-    'font': convert_font_property,
+    'font': lambda x: '',
     'font-family': lambda x: LawCartaConfig.font_correspondence_table.get(x.capitalize()),
     'font-size': convert_font_size,
 }
 
 LIVECARTA_STYLE_ATTRS_SHOULD_BE_TAG = {
-    'font-weight': ['bold', '600', '700', '800', '900'],  # <strong>
+    ('font-weight', 'bold'): 'strong',
-    'font-style': ['italic'],  # <i>
+    ('font-weight', '600'): 'strong',
-    'text-decoration': ['underline', 'line-through'],  # <u> , <s>
+    ('font-weight', '700'): 'strong',
-    'text-decoration-line': ['underline', 'line-through'],  # <u> , <s>
+    ('font-weight', '800'): 'strong',
-    'vertical-align': ['super'],  # <sup>
+    ('font-weight', '900'): 'strong',
+    ('font-style', 'italic'): 'i',
+    ('text-decoration', 'underline'): 'u',
+    ('text-decoration', 'line-through'): 's',
+    ('text-decoration-line', 'underline'): 'u',
+    ('text-decoration-line', 'line-through'): 's',
+    ('vertical-align', 'super'): 'sup',
 }
-'''
-FONT  -> <span>
-font-size:14pt; pt->px
 
+
+def check_style_to_be_tag(style):
+    to_remove = []
+    for k in LIVECARTA_STYLE_ATTRS_SHOULD_BE_TAG:
+        if f'{k[0]}:{k[1]}' in style:
+            to_remove.append(k)
+    return to_remove
+
+
+'''
 LATER:
-vertical-align: sub;  <span style="font-size:10px">o</span>
 text-transform: uppercase;
 text-decoration-color: red;
-
-em, in, pt -> px
 '''
 
 
@@ -116,7 +123,6 @@ def clean_css(css):
                             func = LIVECARTA_STYLE_ATTRS_MAPPING[property_.name]
                             tmp = property_.value.replace('\"', '')
                             rule.style[property_.name] = func(tmp)
-                            print(property_.name, rule.style[property_.name], )
                     else:
                         rule.style[property_.name] = ''
                 else:
@@ -124,24 +130,21 @@ def clean_css(css):
                         func = LIVECARTA_STYLE_ATTRS_MAPPING[property_.name]
                         tmp = property_.value.replace('\"', '')
                         rule.style[property_.name] = func(tmp)
-                        print(property_.name, rule.style[property_.name], )
 
     css_text = sheet._getCssText().decode()
     return css_text
 
 
-def style_property2livecarta_convention(style_str):
-    return style_str
-
-
 def add_inline_style_to_html_soup(soup1, css_text):
     livecarta_p_ids = []
-    h_regex = f'^h[{LawCartaConfig.SUPPORTED_LEVELS + 1}-9]$'
+    h_regex = f'(^h[{LawCartaConfig.SUPPORTED_LEVELS + 1}-9]$)'
-    for i, x in enumerate(soup1.find_all(re.compile('(^p$)|(^span$)|(^li$)|(^ul$)'))):
+    for i, x in enumerate(soup1.find_all(re.compile('(^p$)|(^span$)|(^li$)|(^ul$)' + h_regex))):
         x.attrs['livecarta_id'] = i
         livecarta_p_ids.append(i)
 
-    html_with_inline_style = transform(str(soup1), css_text=css_text, remove_classes=False, external_styles=False,
+    html_with_inline_style = transform(str(soup1), css_text=css_text,
+                                       remove_classes=False,
+                                       external_styles=False,
                                        disable_validation=True)
     soup2 = BeautifulSoup(html_with_inline_style, features='lxml')
 
@@ -150,8 +153,36 @@ def add_inline_style_to_html_soup(soup1, css_text):
         tag_with_style = soup2.find(attrs={'livecarta_id': i})
         if tag_with_style.attrs.get('style'):
             style = tag_with_style.attrs.get('style') + ';'
-            tag.attrs['style'] = style_property2livecarta_convention(style)
+            to_remove = check_style_to_be_tag(style)
-        del tag.attrs['livecarta_id']
+
+            for i, (p, v) in enumerate(to_remove):
+                s = f'{p}:{v};'
+                style = style.replace(s, '')
+                if i == 0:
+                    tag.name = LIVECARTA_STYLE_ATTRS_SHOULD_BE_TAG[(p, v)]
+                    tmp_attrs = tag.attrs.copy()
+                    tag.attrs = {}
+
+                    new_tag = BeautifulSoup(features='lxml').new_tag('span')
+                    new_tag.attrs = tmp_attrs
+                    tag.wrap(new_tag)
+                    print(new_tag)
+                else:
+                    name = LIVECARTA_STYLE_ATTRS_SHOULD_BE_TAG[(p, v)]
+                    new_tag = BeautifulSoup(features='lxml').new_tag(name)
+                    tag.wrap(new_tag)
+
+            if to_remove:
+                new_tag = BeautifulSoup(features='lxml').new_tag('span')
+                new_tag.attrs['style'] = style
+                tag.wrap(new_tag)
+                print(tag)
+                print(list(tag.parent))
+                print()
+                print('---')
+            else:
+                tag.attrs['style'] = style
+                del tag.attrs['livecarta_id']
 
     return soup1