Add interaction counts and update time to smzdm search results

[dengshu2]

Summary

Please enhance opencli smzdm search to include the structured interaction metrics and update time that are already visible on the SMZDM search result page.

Current output columns are:

rank, title, price, mall, comments, url

Suggested output columns:

rank, title, price, mall, updated_at, zhi_count, buzhi_count, favorite_count, comments, url

comments already exists; the request is to add the surrounding metrics and time so agents/users can filter search results without opening each detail page.

Why this is useful

For deal triage, these fields are high-signal:

• updated_at: tells whether the deal is fresh or stale.
• zhi_count: visible "值" count.
• buzhi_count: visible "不值" count.
• favorite_count: visible star/favorite count.
• comments: already present, useful for filtering out zero-discussion results.

Example use case: when searching thinkbook14+, results with comments = 0 can be skipped, while older results with high favorites/comments may still be worth checking.

DOM evidence

The current adapter already scrapes li.feed-row-wide from:

https://search.smzdm.com/?c=home&s=<keyword>&v=b

The requested fields are present in the same list item DOM and do not require opening each detail page.

Observed structure:

<div class="z-feed-foot-l">
  <span class="feed-btn-group price-btn-hover">
    <span class="J_zhi_like_fav price-btn-up" data-type="zhi" data-zhi-type="1">
      <span class="unvoted-wrap"><span>11</span></span>
    </span>
    <span class="J_zhi_like_fav price-btn-down" data-type="zhi" data-zhi-type="-1">
      <span class="unvoted-wrap"><span>3</span></span>
    </span>
  </span>
  <span class="J_zhi_like_fav z-group-data feed-btn-fav">
    <span>40</span>
  </span>
  <a class="z-group-data feed-btn-comment" title="评论数 24">24</a>
</div>

<span class="feed-block-extras">
  05-23 00:28
  <span>天猫精选</span>
</span>

This corresponds to:

{
  "updated_at": "05-23 00:28",
  "zhi_count": 11,
  "buzhi_count": 3,
  "favorite_count": 40,
  "comments": 24,
  "mall": "天猫精选"
}

Suggested implementation

In clis/smzdm/search.js, inside the existing page.evaluate block:

• Extract up count from .price-btn-up .unvoted-wrap span.
• Extract down count from .price-btn-down .unvoted-wrap span.
• Extract favorite count from .feed-btn-fav span.
• Extract update time from .z-feed-foot-r .feed-block-extras by removing the nested mall span text or reading child text nodes before the nested span.
• Keep the existing comments extraction.

This should be a small extension of the current DOM scraping approach rather than a new detail-page request.

Example desired row

{
  "rank": 6,
  "title": "ThinkPad 联想ThinkBook14+ 锐龙R7 H 255 14.5英寸轻薄商务办公学生游戏笔记本",
  "price": "4015.44元（需用券）",
  "mall": "天猫精选",
  "updated_at": "05-23 00:28",
  "zhi_count": 11,
  "buzhi_count": 3,
  "favorite_count": 40,
  "comments": 24,
  "url": "https://www.smzdm.com/p/174854494/"
}

[LeoLin990405]

Claiming this — implemented + tested, PR incoming.

Added updated_at, zhi_count, buzhi_count and favorite_count to the li.feed-row-wide scrape using the exact selectors from your DOM evidence, with jsdom tests that parse your sample markup end to end (zhi 11 / buzhi 3 / fav 40 / comments 24 / updated_at 05-23 00:28 / mall 天猫精选). Missing metrics default to 0 / '' so no column is dropped.