diff options
| -rw-r--r-- | plugins/tipue-search/tipue_search.py | 31 |
1 files changed, 28 insertions, 3 deletions
diff --git a/plugins/tipue-search/tipue_search.py b/plugins/tipue-search/tipue_search.py index 8a62a68..3ce71a8 100644 --- a/plugins/tipue-search/tipue_search.py +++ b/plugins/tipue-search/tipue_search.py @@ -13,6 +13,7 @@ from __future__ import unicode_literals import os.path import json +import re from bs4 import BeautifulSoup from codecs import open try: @@ -36,6 +37,23 @@ class Tipue_Search_JSON_Generator(object): self.output_path = output_path self.json_nodes = [] + def normalize(self, s): + replacements = ( + ("á", "a"), + ("é", "e"), + ("í", "i"), + ("ó", "o"), + ("ú", "u"), + (".", ""), + ) + s = s.lower() + for a, b in replacements: + s = s.replace(a, b).replace(a.lower(), b.lower()) + + s = re.sub(r"([a-z]) ([a-z])", r"\1-\2", s, 0, + re.IGNORECASE | re.DOTALL) + return s + def create_json_node(self, article): if getattr(article, 'status', 'published') != 'published': @@ -48,17 +66,18 @@ class Tipue_Search_JSON_Generator(object): video_text = soup_text.get_text(' ', strip=True).replace('“', '"').replace('”', '"').replace('’', "'").replace('¶', ' ').replace('^', '^') video_text = ' '.join(video_text.split()) + # base url if self.relative_urls: - image_url = '.' + base_url = '.' else: - image_url = self.siteurl + base_url = self.siteurl # thumbnail video_image = article.image if getattr( article, 'image', 'None') != 'None' else '' url_image = "%s/%s/../wp-content/uploads/article/poster/%s" % ( - image_url, self.tstatic, video_image + base_url, self.tstatic, video_image ) # publish @@ -69,6 +88,11 @@ class Tipue_Search_JSON_Generator(object): video_author = str(article.author) if getattr( article, 'author', 'None') != 'None' else '' + # author url + video_author_url = "%s/author/%s/" % ( + base_url, self.normalize(video_author) + ) + # time video_time = article.time if getattr( article, 'time', 'None') != 'None' else '' @@ -91,6 +115,7 @@ class Tipue_Search_JSON_Generator(object): 'url': video_src, }, 'author': video_author, + 'authorUrl': video_author_url, 'publishedText': video_publish, 'time': video_time, 'tags': video_category, |