From 6cd4f9401d6098310bece17eef765008808ca134 Mon Sep 17 00:00:00 2001
From: shirshasa <katerinagorbac@gmail.com>
Date: Fri, 23 Apr 2021 15:21:24 +0300
Subject: [PATCH] epub converter: update

---
 requirements.txt          | 6 +++++-
 src/epub_postprocessor.py | 9 +++++++--
 src/util/color_reader.py  | 2 ++
 src/util/css_reader.py    | 4 ++--
 4 files changed, 16 insertions(+), 5 deletions(-)

diff --git a/requirements.txt b/requirements.txt
index 54ef82a..04ceb99 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,4 +1,8 @@
 bs4>=0.0.1
 pika>=1.1.0
 requests>=2.22.0
-lxml>=4.5.0
\ No newline at end of file
+lxml>=4.5.0
+cssutils~=2.2.0
+ebooklib~=0.17.1
+premailer~=3.8.0
+webcolors==1.3
\ No newline at end of file
diff --git a/src/epub_postprocessor.py b/src/epub_postprocessor.py
index 52a6744..8c23edd 100644
--- a/src/epub_postprocessor.py
+++ b/src/epub_postprocessor.py
@@ -1,5 +1,6 @@
 import codecs
 import json
+from os.path import dirname, normpath, join
 from collections import defaultdict
 from typing import Dict, Union
 
@@ -89,8 +90,12 @@ class EpubPostprocessor:
                 css_href = tag.attrs.get('href')
                 self.html_href2css_href[item.file_name] = css_href
                 if css_href not in self.css_href2content:
-                    print(css_href)
-                    css_content: str = self.ebooklib_book.get_item_with_href(css_href).get_content().decode()
+                    path_to_css_from_html = css_href
+                    html_folder = dirname(item.file_name)
+                    path_to_css_from_root = normpath(join(html_folder, path_to_css_from_html))
+                    css_obj = self.ebooklib_book.get_item_with_href(path_to_css_from_root)
+                    assert css_obj, f'Css style {css_href} was not in manifest.'
+                    css_content: str = css_obj.get_content().decode()
                     self.css_href2content[css_href] = clean_css(css_content)
 
             for i, tag in enumerate(soup.find_all('style')):
diff --git a/src/util/color_reader.py b/src/util/color_reader.py
index 740e170..c15a581 100644
--- a/src/util/color_reader.py
+++ b/src/util/color_reader.py
@@ -59,6 +59,8 @@ def str2color_name(s: str):
         name = get_hex_colour_name(s)
         return name
 
+    elif s in html4_hex_to_names.items():
+        return s
     else:
         return ''
 
diff --git a/src/util/css_reader.py b/src/util/css_reader.py
index 1e1781e..b7977b3 100644
--- a/src/util/css_reader.py
+++ b/src/util/css_reader.py
@@ -76,8 +76,8 @@ LIVECARTA_STYLE_ATTRS_MAPPING = {
     'font': lambda x: '',
     'font-family': lambda x: LawCartaConfig.font_correspondence_table.get(x.capitalize()),
     'font-size': convert_font_size,
-    'color': str2color_name,
-    'background-color': str2color_name,
+    'color': lambda x: LawCartaConfig.HTML42LIVECARTA_COLORS.get(str2color_name(x), ''),
+    'background-color': lambda x: LawCartaConfig.HTML42LIVECARTA_COLORS.get(str2color_name(x), ''),
 }
 
 LIVECARTA_STYLE_ATTRS_SHOULD_BE_TAG = {