search-web-api 1.0.201 → 1.0.203
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/package.json +1 -1
- package/CLAUDE.md +0 -31
package/package.json
CHANGED
package/CLAUDE.md
DELETED
|
@@ -1,31 +0,0 @@
|
|
|
1
|
-
# CLAUDE.md — `search-web-api`
|
|
2
|
-
|
|
3
|
-
**Read [`skills/ask-search-web-api`](../../skills/ask-search-web-api/SKILL.md)
|
|
4
|
-
and its `API.md` first.**
|
|
5
|
-
|
|
6
|
-
70+ search engines across 10 categories, plus dedupe and ranking. Published.
|
|
7
|
-
|
|
8
|
-
## Engine adapters are the bulk of this package
|
|
9
|
-
|
|
10
|
-
- **One adapter per engine, isolated.** An engine that changes its HTML or
|
|
11
|
-
starts rate-limiting must degrade to "this engine returned nothing", never
|
|
12
|
-
take down the query. Failure of one adapter is normal operation.
|
|
13
|
-
- **Never let an engine's markup reach a consumer unsanitized.** Results are
|
|
14
|
-
scraped HTML from arbitrary sites.
|
|
15
|
-
- Dedupe and ranking are what make 70 engines useful rather than noisy. A change
|
|
16
|
-
to either reorders results for every user — pin the behaviour in a test and
|
|
17
|
-
say so in the PR.
|
|
18
|
-
- `domain-rank` supplies source labels, rank and favicons; don't reimplement
|
|
19
|
-
them here.
|
|
20
|
-
|
|
21
|
-
## Rules
|
|
22
|
-
|
|
23
|
-
- Adding an engine: a new adapter plus its category, nothing else.
|
|
24
|
-
- Respect robots/ToS and rate limits — this is a polite client, not a scraper
|
|
25
|
-
race.
|
|
26
|
-
- Tests must not depend on a live engine returning particular results. Use
|
|
27
|
-
fixtures.
|
|
28
|
-
|
|
29
|
-
```bash
|
|
30
|
-
cd packages/search-web-api && bun run test
|
|
31
|
-
```
|