Chinese normalization is not very useful

Currently Chinese segmentation + normalization is applied. In normalization process simplified Chinese are converted to traditional Chinese.

However, the normalization actually does not help with querying with Simplified / Traditional Chinese, because the segmentation is Simplified Chinese only.

Consider the source text:

中国—东盟合作树立“全球南方”联合自强重要范本（国际论坛）

After segmentation + normalization:

中國/東盟/合作/樹立/全球/南方/聯合/自强/重要/范本/國際/論坛

So far that's good.

However, consider a Traditional Chinese query:

國際論坛

After segmentation + normalization it becomes:

國/際/論/坛

As you can see, for Traditional Chinese the text processing is not working well, and does not match what's produced from Simplified Chinese.

Considering that the segmentation is based on dicts, I'm not sure whether it would be a good idea to do some conversion first (for example, unify Simplified and Traditional) and then do the segmentation :)

Start by tracing the Chinese segmentation and normalization paths, paying particular attention to the dictionary-based segmentation described in the issue. Compare processing of the supplied Simplified source and Traditional query examples. Done means equivalent Simplified and Traditional forms are normalized and segmented consistently enough for the query to match the document.