Fix incorrect tokenization of 和习 in Xi-related phrases

[Copilot]

The default segmentation dictionary was missing several high-value political terms, which caused 和习 to be merged and split the following phrase incorrectly in text like 马克思主义和习总书记新时代中国特色社会主义....

• Dictionary updates

  • Add explicit default user-dictionary entries for:
    • 习总书记
    • 新时代
    • 中国特色社会主义
  • Assign high frequencies so these terms win against undesirable fallback splits in mixed-context sentences.

• Regression coverage

  • Add a focused POS-tagging regression test for:
    • 在马克思主义和习总书记新时代中国特色社会主义的思想指引下
  • Lock in the expected boundaries and tags for the corrected phrase.

• Expected output

  在 / 马克思主义 / 和 / 习总书记 / 新时代 / 中国特色社会主义 / 的 / 思想 / 指引 / 下
  

• Code snippet

  习总书记 100000 nr
  新时代 100000 n
  中国特色社会主义 100000 nz
  

[frankslin]

有点好奇用户词典里的「韩玉鉴赏」是什么意思？好像整个中文互联网上这个词出处全部指向这个文件。

[yanyiwu]

Closing this in favor of #225.

#225 keeps the fix narrower by adding only the required main-dictionary entry for 习总书记 and a focused MixSegment regression test for the reported 和习总书记 split. This avoids expanding the default user dictionary with additional high-frequency terms while still addressing #223.

Thanks for the alternative implementation.

[yanyiwu]

Thanks. 韩玉鉴赏 doesn't appear to be an intentional dictionary entry with semantic meaning here; it was just a stray line in the default user dictionary. I cleaned it up separately in #226 so the actual segmentation fix in this area could stay focused.

The narrower fix is now in #227: it only adds the required 习总书记 main-dictionary entry plus a regression test for the reported 和习总书记 case.