In Chinese blog content, there is a link to its counterpart of English blog. The link contains invisible URL, but the URL is indexed and could be matched with a query.

For example, query http://localhost:8080/search/?query=time+management matches two blog articles:
http://diaryofsean.blogspot.com/2020/05/time-management.html and http://diaryofsean.blogspot.com/2020/05/blog-post.html

The first article is the correct match because its title is "Time Management". The second one is the Chinese blog that contains a link to the first article. The link contains "time" and "management", so it is matched with the query. However, it is not the intent of the query. Such invisible URLs and other HTML tags should not be indexed.

Chinese terms for Australian city name are incorrectly tokenized. For example, Chinese term "布里斯本" is the Australian city name "Brisbane". However, it is tokenized as "布", "里斯", "里斯本"

There is a special hyperlink (HTML tag) in the beginning of a Chinese article. The hyperlink is used for English readers to navigate to the corresponding English article. The following is an example of the special hyperlink in Chinese articles.

<div style="text-align: right;">
  <a href="http://diaryofsean.blogspot.com/2020/10/congratulations-your-application-has.html">English version</a>
</div>

The above hyperlink is not supposed to be indexed as document body. The consequence is when searching for word "English", almost all Chinese articles are matched. For example, the search results for https://search.seanlee.site/search/?query=English are all Chinese articles.

This special hyperlink may be useful for establishing the relationship between Chinese and English articles.

When a Chinese term is tokenized as multiple tokens. All tokens are returned when a search matches one of the terms matches. For example, search results for "布里斯本" (i.e. Australian city name "Brisbane") contain all tokens of "布里斯本", which are "布", "里斯", "里斯本" (This tokenization result is incorrect and to be resolved in #6)

As a result, the search result incorrectly display and highlight "布里斯里斯本", whereas the expected result is "布里斯本"

Chinese documents can be matched by English query. For example, Chinese documents contain "fish" in body are matched by query http://localhost:8080/search/?query=fish
However, Chinese documents cannot be matched by Chinese query. For example, no documents are matched by query http://localhost:8080/search/?query=%E9%AD%9A

English query words do not match title/body of English documents when the query words and title/body are in different forms. For example, a query word "reply" does not match "replied" in body.

This happens after adding Chinese tokenizer Jieba for fixing #1

There are non-breaking space ( ) in blog Atom feed. These non-breaking space are displayed as the raw format ( ) in search results. They should be unescaped and trimmed