@fffangchenghao/dsh-web-search 0.24.1 → 0.24.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,25 @@
1
1
  # Changelog
2
2
 
3
+ ## 0.24.2
4
+
5
+ Documentation and image patch; no runtime behavior changed.
6
+
7
+ ### Fixed
8
+ - **The benchmark chart no longer claims to be v0.15.0.** `docs/images/bench.png`
9
+ was regenerated from the current offline replay and labelled v0.24.x. The
10
+ numbers are unchanged (P@8 0.4375, R@8 0.7917, MRR@8 0.8071) because the
11
+ ranking path did not change; the chart now says so without presenting an old
12
+ release as the current one.
13
+ - **Baseline labels now say "current" everywhere.** README, Chinese README and
14
+ `docs/rrf-ranking.md` still explain that the tier-aware replay protocol was
15
+ introduced in 0.15.0, but no longer call the current baseline a "0.15.0
16
+ baseline".
17
+ - **The chart has an editable source.** `docs/images/bench.svg` is checked in
18
+ next to the PNG. Both files are shipped in the npm package so the relative
19
+ image link in `docs/rrf-ranking.md` resolves there too.
20
+
21
+ **Test suite: 690 tests across 56 files** (689 pass, 1 skipped).
22
+
3
23
  ## 0.24.1
4
24
 
5
25
  Documentation and metadata patch; no search or fetch behavior changed.
package/README.md CHANGED
@@ -65,7 +65,7 @@ read is:
65
65
  common in the other plugins.
66
66
  - **Adoption and maturity:** bottom third of this sample. At the 2026-10-25
67
67
  snapshot the repository was four days old, had no stars and little
68
- independent adoption; the current npm release is 0.24.1.
68
+ independent adoption; the current npm release is 0.24.2.
69
69
  - **Tool-surface economy:** now competitive. 0.20.x exposes one `web_tools`
70
70
  entry with eight actions, close to `dsh-web-search-pro`'s two tools with 20
71
71
  actions; the legacy eight-name surface remains opt-in for old prompts.
@@ -340,7 +340,7 @@ Each line is the plain version first, the technical note second. The technical n
340
340
 
341
341
  21. **Search quality has a repeatable benchmark.**
342
342
  Ranking changes are no longer judged by feel: a fixed set of 20 queries produces the same scores every time.
343
- *Technical: `bench/` holds 20 queries (10 Chinese, 10 English), five in each of docs / release / package / news, each annotated with 1–3 expected domains. `node scripts/bench.mjs` replays recorded fixtures offline, never touches the network, and prints precision@8, recall@8, MRR@8 plus each engine's contribution; `--record` re-records live (1.5 s between queries) to refresh the fixtures; `--json` gives machine-readable output for version comparisons, and past results live in `bench/results/`. Baseline for 0.15.0 (tier-aware replay, so the recorded fallback engines only count when production would run them): P@8 `0.4375`, R@8 `0.7917`, MRR@8 `0.8071`; Chinese `0.4000 / 0.8167 / 0.7976`, English `0.4750 / 0.7667 / 0.8167`. A separate five-query long-Chinese diagnostic (`bench/queries-long.json`, recorded with `--queries ... --fixtures ...`) scores `0.4500 / 1.0000 / 0.7667`, with every query hitting an expected domain; it is kept separate so the 20-query baseline stays comparable. The controlled experiments and known gaps are in the appendix.*
343
+ *Technical: `bench/` holds 20 queries (10 Chinese, 10 English), five in each of docs / release / package / news, each annotated with 1–3 expected domains. `node scripts/bench.mjs` replays recorded fixtures offline, never touches the network, and prints precision@8, recall@8, MRR@8 plus each engine's contribution; `--record` re-records live (1.5 s between queries) to refresh the fixtures; `--json` gives machine-readable output for version comparisons, and past results live in `bench/results/`. Current baseline (re-run on the 0.24.x line; tier-aware replay, so the recorded fallback engines only count when production would run them): P@8 `0.4375`, R@8 `0.7917`, MRR@8 `0.8071`; Chinese `0.4000 / 0.8167 / 0.7976`, English `0.4750 / 0.7667 / 0.8167`. A separate five-query long-Chinese diagnostic (`bench/queries-long.json`, recorded with `--queries ... --fixtures ...`) scores `0.4500 / 1.0000 / 0.7667`, with every query hitting an expected domain; it is kept separate so the 20-query baseline stays comparable. The controlled experiments and known gaps are in the appendix.*
344
344
 
345
345
  22. **Research, in one call.**
346
346
  For "what do the official docs and two other sources actually say about X", `research` runs up to three searches, opens the leading pages through the same fetch ladder, and returns one evidence pack in which every excerpt sits under its URL. A page that will not load is reported by URL and reason instead of aborting the pack; bot challenges and login redirects are rejected rather than quoted as evidence.
@@ -792,7 +792,7 @@ No. `package.json` does not list Playwright, and the plugin never downloads or i
792
792
 
793
793
  ### How do you know search quality improved?
794
794
 
795
- From the fixed query set in `bench/`, not from intuition. There are 20 queries (10 Chinese, 10 English), five in each of docs / release / package / news, each annotated with 1–3 expected domains. `node scripts/bench.mjs` replays recorded fixtures offline and prints P@8 / R@8 / MRR@8. The 0.15.0 baseline (tier-aware replay) is **0.4375 / 0.7917 / 0.8071**; Chinese 0.4000 / 0.8167 / 0.7976, English 0.4750 / 0.7667 / 0.8167. Run it before and after a ranking change and the direction is a number. Twenty queries are not statistical significance; news recall is only 0.467 and the per-query table still prints every miss — the appendix says so.
795
+ From the fixed query set in `bench/`, not from intuition. There are 20 queries (10 Chinese, 10 English), five in each of docs / release / package / news, each annotated with 1–3 expected domains. `node scripts/bench.mjs` replays recorded fixtures offline and prints P@8 / R@8 / MRR@8. The current 20-query baseline (tier-aware replay, protocol introduced in 0.15.0) is **0.4375 / 0.7917 / 0.8071**; Chinese 0.4000 / 0.8167 / 0.7976, English 0.4750 / 0.7667 / 0.8167. Run it before and after a ranking change and the direction is a number. Twenty queries are not statistical significance; news recall is only 0.467 and the per-query table still prints every miss — the appendix says so.
796
796
 
797
797
  ### Why do the results differ from what I get searching by hand?
798
798
 
@@ -1040,7 +1040,7 @@ Measured on `https://docs.python.org/3/library/asyncio-task.html`: starting at `
1040
1040
 
1041
1041
  Metrics: P@8 = matched rows in the top 8 ÷ 8 (fixed denominator; a short list is not excused); R@8 = distinct expected domains found in the top 8 ÷ all expected domains; MRR@8 = the mean reciprocal rank of the first match over 20 queries, with a miss scoring 0. Domain matching is one-way from subdomain to parent: `docs.python.org` matches expected `python.org`, but the reverse does not.
1042
1042
 
1043
- **0.15.0 baseline** (tier-aware replay):
1043
+ **Current baseline** (v0.24.x replay; tier-aware protocol introduced in 0.15.0):
1044
1044
 
1045
1045
  | Scope | P@8 | R@8 | MRR@8 |
1046
1046
  | --- | --- | --- | --- |
package/README.zh-CN.md CHANGED
@@ -58,7 +58,7 @@ DSH 插件生态里的「联网搜索」已经很挤。下面是一张 2026-10-2
58
58
  明确的安全边界、只有 1 个运行时依赖,同时体积远小于 Playwright 重型方案。
59
59
  - **免 key + 官方源纠错:最强卖点。** Chrome、PostgreSQL、AI Act 的官方路径
60
60
  和 34 个垂直源,不是其他插件普遍具备的。
61
- - **采用度与成熟度:这个样本里的后段。** 2026-10-25 快照时 GitHub 仓库刚创建 4 天、0 stars,还缺少独立采用;当前 npm 版本是 0.24.1。
61
+ - **采用度与成熟度:这个样本里的后段。** 2026-10-25 快照时 GitHub 仓库刚创建 4 天、0 stars,还缺少独立采用;当前 npm 版本是 0.24.2。
62
62
  - **工具面的经济性:现在有竞争力。** 0.20.x 只暴露一个 `web_tools` + 8 个
63
63
  action,接近 `dsh-web-search-pro` 的 2 工具 + 20 动作;旧的八个工具名保留为
64
64
  可选兼容模式。
@@ -331,7 +331,7 @@ Harness 里有个叫 `web` 的**服务**,它自己不会搜索,只负责「
331
331
 
332
332
  21. **搜索质量有了可重复的基准。**
333
333
  改排序不再靠感觉:仓库里有一组固定的 20 条查询,每次改完都能跑出同一套分数。
334
- *技术补充:`bench/` 下 20 条查询(中文 10、英文 10),四类各 5 条(文档 / 版本发布 / 包与库 / 新闻),每条标注 1–3 个期望域名。`node scripts/bench.mjs` 离线读夹具重放,不联网,输出 precision@8、recall@8、MRR@8 与每个引擎的贡献;`--record` 联网重新录制(每条限速 1.5 秒),用于更新夹具;`--json` 输出机器可读结果,便于版本间对比,历史结果存在 `bench/results/`。当前基线(0.15.0,按 tier 回放):P@8 `0.4375`、R@8 `0.7917`、MRR@8 `0.8071`;中文 `0.4000 / 0.8167 / 0.7976`,英文 `0.4750 / 0.7667 / 0.8167`。另有五条中文长查询的独立诊断集(`bench/queries-long.json`,用 `--queries ... --fixtures ...` 录制):`0.4500 / 1.0000 / 0.7667`,五条都命中期望域名;它与 20 条基线分开保存,避免改变基线的可比性。对照实验和已知短板见附录。*
334
+ *技术补充:`bench/` 下 20 条查询(中文 10、英文 10),四类各 5 条(文档 / 版本发布 / 包与库 / 新闻),每条标注 1–3 个期望域名。`node scripts/bench.mjs` 离线读夹具重放,不联网,输出 precision@8、recall@8、MRR@8 与每个引擎的贡献;`--record` 联网重新录制(每条限速 1.5 秒),用于更新夹具;`--json` 输出机器可读结果,便于版本间对比,历史结果存在 `bench/results/`。当前基线(0.24.x 重放,沿用 0.15.0 的 tier 回放协议):P@8 `0.4375`、R@8 `0.7917`、MRR@8 `0.8071`;中文 `0.4000 / 0.8167 / 0.7976`,英文 `0.4750 / 0.7667 / 0.8167`。另有五条中文长查询的独立诊断集(`bench/queries-long.json`,用 `--queries ... --fixtures ...` 录制):`0.4500 / 1.0000 / 0.7667`,五条都命中期望域名;它与 20 条基线分开保存,避免改变基线的可比性。对照实验和已知短板见附录。*
335
335
 
336
336
  22. **一次调用完成小调研。**
337
337
  问「官方文档和另外两个来源到底怎么说」时,`research` 先跑最多三条查询,再用与 `web_fetch` 相同的抓取阶梯打开靠前的页面,最后返回一份证据包:每段摘录下面就是它的 URL。打不开的页面按 URL 和原因列出,不会让整份证据包失败;人机验证页和登录跳转页会被拒绝,不会被当作证据引用。
@@ -730,7 +730,7 @@ pnpm install
730
730
 
731
731
  ### 你们怎么知道搜索质量变好了?
732
732
 
733
- 靠 `bench/` 里的固定查询集,而不是拍脑袋。20 条查询(中文 10、英文 10),文档 / 版本发布 / 包与库 / 新闻四类各 5 条,每条标注 1–3 个期望域名。`node scripts/bench.mjs` 离线重放录好的夹具,输出 P@8 / R@8 / MRR@8。0.15.0 的基线(按 tier 回放)是 **0.4375 / 0.7917 / 0.8071**;中文 0.4000 / 0.8167 / 0.7976,英文 0.4750 / 0.7667 / 0.8167。改排序前后各跑一次,数字是变好还是变坏一目了然。样本只有 20 条,不能当统计显著性;新闻类 recall 只有 0.467,逐条输出会把每次没命中都打出来——这些短板都写在附录里。
733
+ 靠 `bench/` 里的固定查询集,而不是拍脑袋。20 条查询(中文 10、英文 10),文档 / 版本发布 / 包与库 / 新闻四类各 5 条,每条标注 1–3 个期望域名。`node scripts/bench.mjs` 离线重放录好的夹具,输出 P@8 / R@8 / MRR@8。当前基线(沿用 0.15.0 的 tier 回放协议)是 **0.4375 / 0.7917 / 0.8071**;中文 0.4000 / 0.8167 / 0.7976,英文 0.4750 / 0.7667 / 0.8167。改排序前后各跑一次,数字是变好还是变坏一目了然。样本只有 20 条,不能当统计显著性;新闻类 recall 只有 0.467,逐条输出会把每次没命中都打出来——这些短板都写在附录里。
734
734
 
735
735
  ### 为什么搜索结果和我手动搜的不一样?
736
736
 
@@ -978,7 +978,7 @@ read_page 调用
978
978
 
979
979
  指标:P@8 = 前 8 条里命中期望域名的行数 ÷ 8(固定分母,返回不足 8 条也不放宽);R@8 = 前 8 条覆盖的期望域名 ÷ 全部期望域名;MRR@8 = 第一条命中的排名倒数,20 条取平均,没命中记 0。域名匹配允许「子域命中父域」单向成立:`docs.python.org` 命中期望的 `python.org`,反过来不算。
980
980
 
981
- **0.15.0 基线**(按 tier 回放):
981
+ **当前基线**(0.24.x 重放;沿用 0.15.0 的 tier 回放协议):
982
982
 
983
983
  | 范围 | P@8 | R@8 | MRR@8 |
984
984
  | --- | --- | --- | --- |
package/client.js CHANGED
@@ -74,7 +74,7 @@
74
74
  var ENTRY = 'web-search-bing';
75
75
  var PACKAGE_NAME = 'dsh-web-search-bing';
76
76
  /** Keep in step with package.json; compared with the npm registry. */
77
- var VERSION = '0.24.1';
77
+ var VERSION = '0.24.2';
78
78
  /** Snapshot used while no configuration service is reachable. */
79
79
  var EMPTY_STATE = { status: 'unavailable', value: undefined, base: undefined, user: undefined, revision: undefined, writable: false };
80
80
 
@@ -6,7 +6,7 @@ whether a provider answers at all, how long it takes, how many distinct domains
6
6
  it returns, and whether a page on a known official domain appears in the top 8.
7
7
  The failures are part of the result.
8
8
 
9
- > **Version note (2026-11-02):** 0.24.0 added Windows system-proxy discovery, and 0.24.1 is a documentation/metadata patch. Neither changes the search, ranking or source-adapter path measured below, so the 0.23.0 numbers remain the latest comparison.
9
+ > **Version note (2026-11-02):** 0.24.0 added Windows system-proxy discovery, and 0.24.1 and 0.24.2 are documentation/metadata patches. These patches do not change the search, ranking or source-adapter path measured below, so the 0.23.0 numbers remain the latest comparison.
10
10
 
11
11
  ## Method
12
12
 
Binary file
@@ -0,0 +1,84 @@
1
+ <svg xmlns="http://www.w3.org/2000/svg" width="1400" height="760" viewBox="0 0 1400 760" role="img" aria-labelledby="title desc">
2
+ <title id="title">Ranking quality, measured offline: baseline and parameter sweep</title>
3
+ <desc id="desc">Baseline precision, recall and MRR at 8 for the current 0.24.x replay over 20 fixed queries, plus the 81-configuration marginal sweep.</desc>
4
+ <defs>
5
+ <linearGradient id="bg" x1="0" y1="0" x2="1" y2="1">
6
+ <stop offset="0" stop-color="#0a1120"/><stop offset="0.55" stop-color="#0d1626"/><stop offset="1" stop-color="#0a1322"/>
7
+ </linearGradient>
8
+ <linearGradient id="card" x1="0" y1="0" x2="0" y2="1">
9
+ <stop offset="0" stop-color="#122138"/><stop offset="1" stop-color="#0d1829"/>
10
+ </linearGradient>
11
+ <filter id="glow" x="-40%" y="-40%" width="180%" height="180%">
12
+ <feGaussianBlur stdDeviation="8" result="blur"/><feMerge><feMergeNode in="blur"/><feMergeNode in="SourceGraphic"/></feMerge>
13
+ </filter>
14
+ <style>
15
+ .sans { font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", "Helvetica Neue", Arial, sans-serif; }
16
+ .title { fill: #f4f7fb; font-size: 38px; font-weight: 700; letter-spacing: -0.6px; }
17
+ .sub { fill: #8ea3bf; font-size: 17px; }
18
+ .h2 { fill: #f4f7fb; font-size: 24px; font-weight: 700; }
19
+ .h3 { fill: #9db1cc; font-size: 15px; font-weight: 600; letter-spacing: 0.2px; }
20
+ .muted { fill: #8195b0; font-size: 14px; }
21
+ .tiny { fill: #6f839d; font-size: 12.5px; }
22
+ .val { fill: #dbe8f8; font-size: 17px; font-weight: 600; }
23
+ .axis { fill: #8ba0bb; font-size: 15px; font-weight: 600; }
24
+ .blue { fill: #11243d; stroke: #5b8def; }
25
+ .green { fill: #112a2a; stroke: #37c996; }
26
+ .orange { fill: #2b2418; stroke: #f0a43a; }
27
+ </style>
28
+ </defs>
29
+ <rect width="1400" height="760" fill="url(#bg)"/>
30
+ <text x="60" y="62" class="sans title">Ranking quality, measured offline</text>
31
+ <text x="60" y="98" class="sans sub">20 fixed queries, tier-aware replay of recorded fixtures; no network. v0.24.x.</text>
32
+ <rect x="60" y="135" width="600" height="565" rx="16" fill="url(#card)" stroke="#294563" stroke-width="1.5"/>
33
+ <rect x="700" y="135" width="640" height="565" rx="16" fill="url(#card)" stroke="#294563" stroke-width="1.5"/>
34
+ <text x="90" y="184" class="sans h2">Baseline (v0.24.x replay)</text>
35
+ <text x="730" y="184" class="sans h2">Parameter sweep: marginal MRR@8</text>
36
+ <text x="730" y="211" class="sans muted">81 configurations, paired bootstrap vs defaults</text>
37
+ <rect x="190" y="465.6" width="82" height="144.4" rx="5" fill="none" stroke="#5b8def" stroke-width="2.5" class="blue" opacity="0.95"/>
38
+ <text x="231" y="495.6" text-anchor="middle" class="sans val" fill="#5b8def">0.4375</text>
39
+ <text x="231" y="652" text-anchor="middle" class="sans axis">P@8</text>
40
+ <rect x="395" y="348.8" width="82" height="261.2" rx="5" fill="none" stroke="#37c996" stroke-width="2.5" class="green" opacity="0.95"/>
41
+ <text x="436" y="378.8" text-anchor="middle" class="sans val" fill="#37c996">0.7917</text>
42
+ <text x="436" y="652" text-anchor="middle" class="sans axis">R@8</text>
43
+ <rect x="515" y="343.6" width="82" height="266.4" rx="5" fill="none" stroke="#f0a43a" stroke-width="2.5" class="orange" opacity="0.95"/>
44
+ <text x="556" y="373.6" text-anchor="middle" class="sans val" fill="#f0a43a">0.8071</text>
45
+ <text x="556" y="652" text-anchor="middle" class="sans axis">MRR@8</text>
46
+ <line x1="88" y1="610" x2="632" y2="610" stroke="#39536f" stroke-width="1"/>
47
+ <text x="90" y="681" class="sans tiny">n=20, k=8; 8 engines (7 primary + 1 fallback), zero-hit queries: 0</text>
48
+ <text x="745" y="258" class="sans h3">rrfConstant</text>
49
+ <rect x="745" y="442.7" width="38" height="147.3" rx="4" fill="#0f2036" stroke="#37c996" stroke-width="2" opacity="0.97"/>
50
+ <text x="764" y="432.7" text-anchor="middle" class="sans tiny" fill="#37c996">0.834</text>
51
+ <text x="764" y="615" text-anchor="middle" class="sans tiny">40</text>
52
+ <rect x="792" y="465.8" width="38" height="124.2" rx="4" fill="#0f2036" stroke="#5b8def" stroke-width="2" opacity="0.97"/>
53
+ <text x="811" y="455.8" text-anchor="middle" class="sans tiny" fill="#5b8def">0.828</text>
54
+ <text x="811" y="615" text-anchor="middle" class="sans tiny">60</text>
55
+ <rect x="839" y="496.9" width="38" height="93.1" rx="4" fill="#0f2036" stroke="#37c996" stroke-width="2" opacity="0.97"/>
56
+ <text x="858" y="486.9" text-anchor="middle" class="sans tiny" fill="#37c996">0.819</text>
57
+ <text x="858" y="615" text-anchor="middle" class="sans tiny">80</text>
58
+ <text x="935" y="258" class="sans h3">authorityBoost</text>
59
+ <rect x="935" y="466.9" width="38" height="123.1" rx="4" fill="#0f2036" stroke="#37c996" stroke-width="2" opacity="0.97"/>
60
+ <text x="954" y="456.9" text-anchor="middle" class="sans tiny" fill="#37c996">0.827</text>
61
+ <text x="954" y="615" text-anchor="middle" class="sans tiny">0.1</text>
62
+ <rect x="982" y="471.6" width="38" height="118.4" rx="4" fill="#0f2036" stroke="#5b8def" stroke-width="2" opacity="0.97"/>
63
+ <text x="1001" y="461.6" text-anchor="middle" class="sans tiny" fill="#5b8def">0.826</text>
64
+ <text x="1001" y="615" text-anchor="middle" class="sans tiny">0.25</text>
65
+ <rect x="1029" y="466.9" width="38" height="123.1" rx="4" fill="#0f2036" stroke="#37c996" stroke-width="2" opacity="0.97"/>
66
+ <text x="1048" y="456.9" text-anchor="middle" class="sans tiny" fill="#37c996">0.827</text>
67
+ <text x="1048" y="615" text-anchor="middle" class="sans tiny">0.4</text>
68
+ <text x="1125" y="258" class="sans h3">freshnessBoost</text>
69
+ <rect x="1125" y="483.6" width="38" height="106.4" rx="4" fill="#0f2036" stroke="#37c996" stroke-width="2" opacity="0.97"/>
70
+ <text x="1144" y="473.6" text-anchor="middle" class="sans tiny" fill="#37c996">0.823</text>
71
+ <text x="1144" y="615" text-anchor="middle" class="sans tiny">0.1</text>
72
+ <rect x="1172" y="478.0" width="38" height="112.0" rx="4" fill="#0f2036" stroke="#5b8def" stroke-width="2" opacity="0.97"/>
73
+ <text x="1191" y="468.0" text-anchor="middle" class="sans tiny" fill="#5b8def">0.825</text>
74
+ <text x="1191" y="615" text-anchor="middle" class="sans tiny">0.2</text>
75
+ <rect x="1219" y="443.8" width="38" height="146.2" rx="4" fill="#0f2036" stroke="#37c996" stroke-width="2" opacity="0.97"/>
76
+ <text x="1238" y="433.8" text-anchor="middle" class="sans tiny" fill="#37c996">0.834</text>
77
+ <text x="1238" y="615" text-anchor="middle" class="sans tiny">0.35</text>
78
+ <line x1="728" y1="590" x2="1310" y2="590" stroke="#39536f" stroke-width="1"/>
79
+ <rect x="730" y="632" width="580" height="62" rx="8" fill="#0b1728" stroke="#2f4b6a" stroke-width="1"/>
80
+ <text x="746" y="651" class="sans tiny" fill="#dbe8f8" font-weight="700">Best grid point: rrf=40, authority=0.25, freshness=0.35</text>
81
+ <text x="746" y="668" class="sans tiny" fill="#37c996">MRR 0.8583 vs 0.8071 default (+0.0512)</text>
82
+ <text x="746" y="685" class="sans tiny" fill="#f0a43a">95% CI [0.0000, +0.1250] touches zero: a direction, not a proven win</text>
83
+ <text x="730" y="725" class="sans tiny" fill="#6f839d">Blue = shipped default. Marginals average every configuration in the grid.</text>
84
+ </svg>
@@ -6,7 +6,7 @@ This note answers three questions from the public 30-query comparison:
6
6
  2. Can the AnySearch tail and the Chinese Rust miss be fixed?
7
7
  3. Does the fix require a vector database or a different retrieval model?
8
8
 
9
- > **Version note (2026-11-02):** 0.24.0 added Windows system-proxy discovery, and 0.24.1 is a documentation/metadata patch. The search path is unchanged, so these 0.23.0 measurements still describe 0.24.x.
9
+ > **Version note (2026-11-02):** 0.24.0 added Windows system-proxy discovery, and 0.24.1 and 0.24.2 are documentation/metadata patches. The search path is unchanged, so these 0.23.0 measurements still describe 0.24.x.
10
10
 
11
11
  The short answer: Exa is a different architecture, not a faster RRF loop. This
12
12
  plugin can borrow tail-tolerance and source-specific facts, but copying Exa's
@@ -77,7 +77,7 @@ comparable across commits; `--record` is the only mode that touches the network.
77
77
 
78
78
  ![Baseline and sweep](images/bench.png)
79
79
 
80
- Baseline (v0.15.0, 20 queries, k=8, tier-aware replay):
80
+ Baseline (current v0.24.x replay, 20 queries, k=8; tier-aware protocol introduced in 0.15.0):
81
81
 
82
82
  | metric | overall | Chinese | English |
83
83
  |---|---:|---:|---:|
@@ -29,7 +29,7 @@ Only `0.2.1-rc.2` is declared, because that is the version actually exercised.
29
29
  `mktemp -d`; the real `~/.dsh` was never written to. Bundled pnpm 11.28.5.
30
30
  - **Windows CI:** GitHub Actions `windows-latest`, Node 24. Install, typecheck,
31
31
  the full test suite, build and the entry-point import all pass. Run:
32
- https://github.com/fangchenghao/dsh-web-search/actions/runs/41387433033
32
+ https://github.com/fangchenghao/dsh-web-search/actions/runs/41389877558
33
33
  (the matrix also runs on macOS and Ubuntu). This is a CI environment, not a
34
34
  claim about every Windows terminal or proxy client.
35
35
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@fffangchenghao/dsh-web-search",
3
- "version": "0.24.1",
3
+ "version": "0.24.2",
4
4
  "description": "DSH bundle: key-free web search, page fetch and research for DeepSeek Harness; eight engines fused with RRF, 34 vertical sources, arXiv, Windows/macOS system-proxy support and a dependency-free MCP server.",
5
5
  "type": "module",
6
6
  "main": "lib/index.js",
@@ -37,7 +37,9 @@
37
37
  "docs/plugin-settings-form.md",
38
38
  "docs/troubleshooting.md",
39
39
  "docs/architecture.md",
40
- "docs/engines.md"
40
+ "docs/engines.md",
41
+ "docs/images/bench.png",
42
+ "docs/images/bench.svg"
41
43
  ],
42
44
  "scripts": {
43
45
  "build": "tsc -p tsconfig.json",