Start adding blog post authors and licenses.

Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>

1 files changed, 22 insertions, 0 deletions

diff --git a/convert.py b/convert.py
index a2eeb20..f5f4e62 100755
--- a/convert.py
+++ b/convert.py
@@ -145,6 +145,21 @@ def get_metadata(html_file_path):
 
         return ', '.join(results)
 
+    def get_author(soup):
+        results = []
+        author_vcard = soup.article.find_all('span', class_='author vcard')
+        assert(len(author_vcard) == 1)
+        link = author_vcard[0].find_all('a')
+        assert(len(link) == 1)
+        return link[0].string
+
+    # Returns SPDX license declaration.
+    def get_licenses(soup):
+        if get_author(soup) == 'GNUtoo':
+            return 'CC-BY-3.0 OR CC-BY-4.0'
+        else:
+            return None
+
     def get_title(soup):
         title = soup.title.string
         title = title.replace(os.linesep, '')
@@ -164,10 +179,17 @@ def get_metadata(html_file_path):
     # assert(title_metadata != None)
     # metadata += "title: {}".format(title_metadata)
 
+    authors_metadata = _get_metadata(html_file_path, get_author)
+    metadata += "title: {}".format(authors_metadata) + os.linesep
+
     tags_metadata = _get_metadata(html_file_path, get_tags)
     if tags_metadata:
         print("tags: {}".format(tags_metadata))
 
+    licenses_metadata = _get_metadata(html_file_path, get_licenses)
+    if licenses_metadata:
+        print("licenses: {}".format(licenses_metadata))
+
     metadata += "---" + os.linesep
 
     return metadata