@rivi-software/agentic-seo-check 0.0.0-stage → 0.1.0
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/LICENSE +21 -0
- package/README.md +160 -2
- package/dist/cli.d.ts +2 -0
- package/dist/cli.js +149 -0
- package/dist/cli.js.map +1 -0
- package/dist/core/audit.d.ts +72 -0
- package/dist/core/audit.js +337 -0
- package/dist/core/audit.js.map +1 -0
- package/dist/core/concurrency.d.ts +4 -0
- package/dist/core/concurrency.js +29 -0
- package/dist/core/concurrency.js.map +1 -0
- package/dist/core/crawl.d.ts +40 -0
- package/dist/core/crawl.js +195 -0
- package/dist/core/crawl.js.map +1 -0
- package/dist/core/fetch.d.ts +27 -0
- package/dist/core/fetch.js +135 -0
- package/dist/core/fetch.js.map +1 -0
- package/dist/core/html/extract.d.ts +10 -0
- package/dist/core/html/extract.js +150 -0
- package/dist/core/html/extract.js.map +1 -0
- package/dist/core/html/structured-data.d.ts +11 -0
- package/dist/core/html/structured-data.js +185 -0
- package/dist/core/html/structured-data.js.map +1 -0
- package/dist/core/html/text.d.ts +23 -0
- package/dist/core/html/text.js +149 -0
- package/dist/core/html/text.js.map +1 -0
- package/dist/core/lighthouse.d.ts +49 -0
- package/dist/core/lighthouse.js +140 -0
- package/dist/core/lighthouse.js.map +1 -0
- package/dist/core/performance-rules.d.ts +3 -0
- package/dist/core/performance-rules.js +53 -0
- package/dist/core/performance-rules.js.map +1 -0
- package/dist/core/render-rules.d.ts +7 -0
- package/dist/core/render-rules.js +77 -0
- package/dist/core/render-rules.js.map +1 -0
- package/dist/core/render.d.ts +22 -0
- package/dist/core/render.js +73 -0
- package/dist/core/render.js.map +1 -0
- package/dist/core/report/markdown.d.ts +2 -0
- package/dist/core/report/markdown.js +152 -0
- package/dist/core/report/markdown.js.map +1 -0
- package/dist/core/robots.d.ts +11 -0
- package/dist/core/robots.js +33 -0
- package/dist/core/robots.js.map +1 -0
- package/dist/core/rules/canonical.d.ts +2 -0
- package/dist/core/rules/canonical.js +105 -0
- package/dist/core/rules/canonical.js.map +1 -0
- package/dist/core/rules/content.d.ts +1 -0
- package/dist/core/rules/content.js +41 -0
- package/dist/core/rules/content.js.map +1 -0
- package/dist/core/rules/document.d.ts +1 -0
- package/dist/core/rules/document.js +60 -0
- package/dist/core/rules/document.js.map +1 -0
- package/dist/core/rules/headings.d.ts +1 -0
- package/dist/core/rules/headings.js +61 -0
- package/dist/core/rules/headings.js.map +1 -0
- package/dist/core/rules/hreflang.d.ts +1 -0
- package/dist/core/rules/hreflang.js +68 -0
- package/dist/core/rules/hreflang.js.map +1 -0
- package/dist/core/rules/images.d.ts +1 -0
- package/dist/core/rules/images.js +59 -0
- package/dist/core/rules/images.js.map +1 -0
- package/dist/core/rules/index.d.ts +14 -0
- package/dist/core/rules/index.js +76 -0
- package/dist/core/rules/index.js.map +1 -0
- package/dist/core/rules/indexing.d.ts +10 -0
- package/dist/core/rules/indexing.js +53 -0
- package/dist/core/rules/indexing.js.map +1 -0
- package/dist/core/rules/keyword.d.ts +1 -0
- package/dist/core/rules/keyword.js +74 -0
- package/dist/core/rules/keyword.js.map +1 -0
- package/dist/core/rules/links.d.ts +1 -0
- package/dist/core/rules/links.js +72 -0
- package/dist/core/rules/links.js.map +1 -0
- package/dist/core/rules/meta.d.ts +1 -0
- package/dist/core/rules/meta.js +60 -0
- package/dist/core/rules/meta.js.map +1 -0
- package/dist/core/rules/response.d.ts +1 -0
- package/dist/core/rules/response.js +59 -0
- package/dist/core/rules/response.js.map +1 -0
- package/dist/core/rules/rule.d.ts +42 -0
- package/dist/core/rules/rule.js +51 -0
- package/dist/core/rules/rule.js.map +1 -0
- package/dist/core/rules/social.d.ts +1 -0
- package/dist/core/rules/social.js +74 -0
- package/dist/core/rules/social.js.map +1 -0
- package/dist/core/rules/structured-data.d.ts +23 -0
- package/dist/core/rules/structured-data.js +169 -0
- package/dist/core/rules/structured-data.js.map +1 -0
- package/dist/core/rules/title.d.ts +1 -0
- package/dist/core/rules/title.js +59 -0
- package/dist/core/rules/title.js.map +1 -0
- package/dist/core/score.d.ts +3 -0
- package/dist/core/score.js +15 -0
- package/dist/core/score.js.map +1 -0
- package/dist/core/site-rules.d.ts +64 -0
- package/dist/core/site-rules.js +251 -0
- package/dist/core/site-rules.js.map +1 -0
- package/dist/core/sitemap.d.ts +13 -0
- package/dist/core/sitemap.js +95 -0
- package/dist/core/sitemap.js.map +1 -0
- package/dist/core/static.d.ts +9 -0
- package/dist/core/static.js +184 -0
- package/dist/core/static.js.map +1 -0
- package/dist/core/summary.d.ts +2 -0
- package/dist/core/summary.js +28 -0
- package/dist/core/summary.js.map +1 -0
- package/dist/core/types.d.ts +296 -0
- package/dist/core/types.js +26 -0
- package/dist/core/types.js.map +1 -0
- package/dist/core/url.d.ts +20 -0
- package/dist/core/url.js +139 -0
- package/dist/core/url.js.map +1 -0
- package/dist/index.d.ts +11 -0
- package/dist/index.js +10 -0
- package/dist/index.js.map +1 -0
- package/dist/mcp-bin.d.ts +2 -0
- package/dist/mcp-bin.js +7 -0
- package/dist/mcp-bin.js.map +1 -0
- package/dist/mcp.d.ts +5 -0
- package/dist/mcp.js +190 -0
- package/dist/mcp.js.map +1 -0
- package/dist/version.d.ts +2 -0
- package/dist/version.js +4 -0
- package/dist/version.js.map +1 -0
- package/package.json +89 -3
- package/skills/seo-audit/SKILL.md +53 -0
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"crawl.js","sourceRoot":"","sources":["../../src/core/crawl.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,EAAE,kBAAkB,EAAE,SAAS,EAAE,cAAc,EAAE,QAAQ,EAAE,cAAc,EAAqB,MAAM,YAAY,CAAC;AACxH,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAC5C,OAAO,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAGnD,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,aAAa,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAC;AACrF,OAAO,EAAE,aAAa,EAAE,KAAK,EAAE,MAAM,kBAAkB,CAAC;AA8CxD,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,UAAkB,EAAE,OAAqB;IACvE,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;IAClC,MAAM,QAAQ,GAAG,OAAO,CAAC,UAAU,IAAI,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;IACzD,MAAM,QAAQ,GAAG,YAAY,CAAC,UAAU,CAAC,IAAI,UAAU,CAAC;IACxD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC;IACxC,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;IACzC,MAAM,SAAS,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,IAAI,kBAAkB,CAAC;IAEhE,QAAQ,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,YAAY,MAAM,aAAa,EAAE,CAAC,CAAC;IACxE,MAAM,MAAM,GAAG,MAAM,aAAa,CAAC,MAAM,EAAE,SAAS,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC;IACrE,MAAM,MAAM,GAAG,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC;IAElE,QAAQ,CAAC,EAAE,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,mBAAmB,EAAE,CAAC,CAAC;IAC7D,MAAM,iBAAiB,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,QAAQ,EAAE,IAAI,GAAG,CAAC,cAAc,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACxG,MAAM,QAAQ,GAAG,MAAM,YAAY,CAAC,iBAAiB,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC;IAEtE,MAAM,OAAO,GAAG,IAAI,GAAG,CAAS,CAAC,QAAQ,CAAC,CAAC,CAAC;IAC5C,MAAM,UAAU,GAAG,IAAI,GAAG,CAAS,CAAC,QAAQ,CAAC,CAAC,CAAC;IAC/C,MAAM,UAAU,GAAG,IAAI,GAAG,EAAqB,CAAC;IAChD,MAAM,WAAW,GAAG,IAAI,GAAG,EAAuB,CAAC;IACnD,MAAM,eAAe,GAAG,IAAI,GAAG,EAAuB,CAAC;IACvD,MAAM,OAAO,GAAiB,EAAE,CAAC;IACjC,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,MAAM,OAAO,GAAG,CAAC,GAAW,EAAW,EAAE;QACvC,IAAI,OAAO,CAAC,eAAe,CAAC,MAAM,IAAI,CAAC,OAAO,CAAC,eAAe,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QACxG,IAAI,OAAO,CAAC,eAAe,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QACrE,OAAO,IAAI,CAAC;IACd,CAAC,CAAC;IAEF,MAAM,KAAK,GAAgB,CAAC,EAAE,GAAG,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAC;IAEzD,MAAM,WAAW,GAAG,KAAK,EAAE,IAAe,EAAiB,EAAE;QAC3D,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YACtB,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACvB,OAAO;QACT,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC;QACrD,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC;QAC9B,MAAM,QAAQ,GAAG,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC,QAAQ,CAAC;QAC5D,UAAU,CAAC,GAAG,CAAC,YAAY,EAAE,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,aAAa,EAAE,GAAG,CAAC,aAAa,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC;QAChJ,IAAI,QAAQ,KAAK,YAAY,EAAE,CAAC;YAC9B,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,QAAQ,CAAC;gBAAE,UAAU,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,QAAQ,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,aAAa,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC;YAChK,IAAI,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,QAAQ,CAAC;gBAAE,OAAO;YAC7E,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;YACtB,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,SAAS,EAAE,OAAO,CAAC,iBAAiB,CAAC;gBAAE,OAAO,CAAC,sBAAsB;QACjG,CAAC;QACD,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,GAAG,GAAG,CAAC,MAAM,IAAI,KAAK,IAAI,GAAG,CAAC,QAAQ,EAAE,EAAE,OAAO,EAAE,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,UAAU,EAAE,UAAU,CAAC,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,QAAQ,EAAE,CAAC,CAAC;QAC7J,IAAI,GAAG,CAAC,KAAK,IAAI,OAAO,GAAG,CAAC,IAAI,KAAK,QAAQ,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,EAAE,CAAC;YACtE,IAAI,GAAG,CAAC,MAAM,KAAK,IAAI,IAAI,GAAG,CAAC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,IAAI,OAAO,GAAG,CAAC,IAAI,KAAK,QAAQ,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC;gBAAE,OAAO,CAAC,iBAAiB;YAC5I,8FAA8F;YAC9F,IAAI,IAAI,CAAC,KAAK,KAAK,CAAC,IAAI,GAAG,CAAC,MAAM,KAAK,IAAI,EAAE,CAAC;gBAC5C,MAAM,IAAI,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,IAAI,EAAE,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,EAAE,CAAC,CAAC;gBACnI,OAAO,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,QAAQ,EAAE,YAAY,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,QAAQ,EAAE,cAAc,CAAC,GAAG,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;YAClH,CAAC;YACD,OAAO;QACT,CAAC;QACD,MAAM,IAAI,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,EAAE,CAAC,CAAC;QAC7H,OAAO,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,QAAQ,EAAE,YAAY,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,QAAQ,EAAE,cAAc,CAAC,GAAG,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;QAEhH,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,IAAI,CAAC,IAAI,CAAC,GAAG;gBAAE,SAAS;YACxB,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACtC,IAAI,CAAC,MAAM;gBAAE,SAAS;YACtB,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACnB,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC;oBAAE,eAAe,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC,CAAC;gBACzE,eAAe,CAAC,GAAG,CAAC,MAAM,CAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;gBAC3C,SAAS;YACX,CAAC;YACD,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC;gBAAE,WAAW,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC,CAAC;YACjE,WAAW,CAAC,GAAG,CAAC,MAAM,CAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;YACvC,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;gBAAE,SAAS;YACzD,IAAI,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC;gBAAE,SAAS;YACrC,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YACvB,IAAI,IAAI,CAAC,KAAK,GAAG,CAAC,GAAG,OAAO,CAAC,QAAQ;gBAAE,SAAS;YAChD,IAAI,OAAO,CAAC,IAAI,IAAI,OAAO,CAAC,QAAQ,EAAE,CAAC;gBACrC,UAAU,EAAE,CAAC;gBACb,SAAS;YACX,CAAC;YACD,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YACpB,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC,CAAC;QACrD,CAAC;IACH,CAAC,CAAC;IAEF,MAAM,QAAQ,CAAC,KAAK,EAAE,OAAO,CAAC,WAAW,EAAE,OAAO,CAAC,OAAO,EAAE,WAAW,CAAC,CAAC;IAEzE,iGAAiG;IACjG,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,gCAAgC,EAAE,CAAC,CAAC;IACxE,MAAM,OAAO,GAAG,CAAC,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC;IACvH,MAAM,YAAY,GAAG,QAAQ,CAAC,IAAI;SAC/B,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;SAChC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,iBAAiB,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC;SAC7H,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,aAAa,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC;IACjE,KAAK,MAAM,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC;YAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC7G,MAAM,aAAa,CAAC,CAAC,GAAG,OAAO,EAAE,GAAG,YAAY,CAAC,EAAE,OAAO,CAAC,WAAW,EAAE,KAAK,EAAE,CAAC,EAAE,EAAE;QAClF,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC;QAC7C,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,aAAa,EAAE,GAAG,CAAC,aAAa,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC;QAC9H,IAAI,OAAO,CAAC,OAAO;YAAE,MAAM,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IACpD,CAAC,CAAC,CAAC;IAEH,IAAI,cAAkD,CAAC;IACvD,IAAI,OAAO,CAAC,kBAAkB,EAAE,CAAC;QAC/B,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,YAAY,IAAI,CAAC,GAAG,CAAC,eAAe,CAAC,IAAI,EAAE,OAAO,CAAC,aAAa,CAAC,iBAAiB,EAAE,CAAC,CAAC;QAC1H,cAAc,GAAG,IAAI,GAAG,EAAE,CAAC;QAC3B,MAAM,OAAO,GAAG,CAAC,GAAG,eAAe,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC;QAC5E,MAAM,aAAa,CAAC,OAAO,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,EAAE;YACzE,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,CAAC,EAAE,EAAE,GAAG,OAAO,CAAC,KAAK,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,SAAS,IAAI,MAAM,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;YACpH,cAAe,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,aAAa,EAAE,GAAG,CAAC,aAAa,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC;QACrI,CAAC,CAAC,CAAC;IACL,CAAC;IAED,QAAQ,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,qCAAqC,EAAE,CAAC,CAAC;IAC9E,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,QAAQ,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC;IAExD,MAAM,KAAK,GAAe;QACxB,YAAY,EAAE,OAAO,CAAC,MAAM;QAC5B,eAAe,EAAE,UAAU,CAAC,IAAI;QAChC,YAAY,EAAE,UAAU,CAAC,IAAI,GAAG,CAAC,cAAc,EAAE,IAAI,IAAI,CAAC,CAAC;QAC3D,QAAQ,EAAE,OAAO,CAAC,QAAQ;QAC1B,QAAQ,EAAE,OAAO,CAAC,QAAQ;QAC1B,YAAY,EAAE,UAAU,GAAG,CAAC;QAC5B,UAAU,EAAE,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,GAAG,EAAE,GAAG,OAAO,CAAC;QACnD,eAAe,EAAE,OAAO,CAAC,aAAa;KACvC,CAAC;IAEF,OAAO;QACL,QAAQ;QACR,OAAO;QACP,UAAU;QACV,WAAW;QACX,cAAc;QACd,eAAe,EAAE,OAAO,CAAC,kBAAkB,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,SAAS;QACzE,MAAM,EAAE,MAAM,CAAC,IAAI;QACnB,YAAY,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI;QACtD,eAAe,EAAE,OAAO;QACxB,QAAQ;QACR,MAAM;QACN,KAAK;KACN,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,QAAQ,CAAC,KAAkB,EAAE,WAAmB,EAAE,OAAe,EAAE,IAAwC;IACxH,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE;QAClC,MAAM,IAAI,GAAG,GAAS,EAAE;YACtB,OAAO,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,WAAW,CAAC,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;gBACzD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,EAAG,CAAC;gBAC5B,MAAM,EAAE,CAAC;gBACT,IAAI,CAAC,IAAI,CAAC;qBACP,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC;qBACtB,OAAO,CAAC,GAAG,EAAE;oBACZ,MAAM,EAAE,CAAC;oBACT,IAAI,OAAO,GAAG,CAAC;wBAAE,UAAU,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;;wBACtC,IAAI,EAAE,CAAC;gBACd,CAAC,CAAC,CAAC;YACP,CAAC;YACD,IAAI,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;gBAAE,OAAO,EAAE,CAAC;QACpD,CAAC,CAAC;QACF,IAAI,EAAE,CAAC;IACT,CAAC,CAAC,CAAC;AACL,CAAC;AAED,KAAK,UAAU,SAAS,CAAC,QAAgB,EAAE,SAAuB;IAChE,MAAM,MAAM,GAAiB,EAAE,CAAC;IAChC,MAAM,GAAG,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC9B,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,sBAAsB,IAAI,CAAC,GAAG,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC;IACvI,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,SAAS,CAAC,CAAC;IACnD,MAAM,CAAC,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;IAE9E,IAAI,WAAW,CAAC,QAAQ,CAAC;QAAE,OAAO,MAAM,CAAC;IAEzC,IAAI,GAAG,CAAC,QAAQ,KAAK,QAAQ,EAAE,CAAC;QAC9B,MAAM,OAAO,GAAG,UAAU,GAAG,CAAC,IAAI,GAAG,CAAC;QACtC,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,OAAO,EAAE,EAAE,GAAG,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC;QACzE,MAAM,CAAC,YAAY,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,gBAAgB,EAAE,GAAG,CAAC,QAAQ,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;IAC/J,CAAC;IAED,MAAM,IAAI,GAAG,GAAG,CAAC,QAAQ,CAAC;IAC1B,MAAM,WAAW,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,OAAO,IAAI,EAAE,CAAC;IAC7E,MAAM,UAAU,GAAG,GAAG,GAAG,CAAC,QAAQ,KAAK,WAAW,GAAG,CAAC;IACtD,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,EAAE,GAAG,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC;IAC5E,MAAM,CAAC,WAAW,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,eAAe,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC;IAChK,OAAO,MAAM,CAAC;AAChB,CAAC"}
|
|
@@ -0,0 +1,27 @@
|
|
|
1
|
+
import type { RedirectHop, ResponseInfo } from './types.js';
|
|
2
|
+
export declare const DEFAULT_USER_AGENT: string;
|
|
3
|
+
export interface FetchOptions {
|
|
4
|
+
timeoutMs?: number;
|
|
5
|
+
userAgent?: string;
|
|
6
|
+
headers?: Record<string, string>;
|
|
7
|
+
maxRedirects?: number;
|
|
8
|
+
/** Stop reading the body after this many bytes. */
|
|
9
|
+
maxBodyBytes?: number;
|
|
10
|
+
}
|
|
11
|
+
export interface FetchResult {
|
|
12
|
+
requestedUrl: string;
|
|
13
|
+
finalUrl: string;
|
|
14
|
+
status: number | null;
|
|
15
|
+
headers: Record<string, string>;
|
|
16
|
+
body: string | null;
|
|
17
|
+
redirectChain: RedirectHop[];
|
|
18
|
+
timeMs: number;
|
|
19
|
+
contentType: string | null;
|
|
20
|
+
error?: string;
|
|
21
|
+
}
|
|
22
|
+
/** GET a URL, following redirects manually so the chain is recorded. */
|
|
23
|
+
export declare function fetchPage(url: string, options?: FetchOptions, readBody?: boolean): Promise<FetchResult>;
|
|
24
|
+
/** Status/redirect check without downloading the body. */
|
|
25
|
+
export declare function probeUrl(url: string, options?: FetchOptions): Promise<FetchResult>;
|
|
26
|
+
export declare function toResponseInfo(r: FetchResult): ResponseInfo;
|
|
27
|
+
export declare function isHtmlResponse(r: FetchResult): boolean;
|
|
@@ -0,0 +1,135 @@
|
|
|
1
|
+
import { VERSION } from '../version.js';
|
|
2
|
+
export const DEFAULT_USER_AGENT = `Mozilla/5.0 (compatible; agentic-seo-check/${VERSION}; +https://github.com/rivi-software/agentic-seo-check)`;
|
|
3
|
+
const KEPT_HEADERS = ['content-type', 'x-robots-tag', 'link', 'content-language', 'cache-control', 'last-modified', 'location', 'content-length', 'vary'];
|
|
4
|
+
const REDIRECT_STATUSES = new Set([301, 302, 303, 307, 308]);
|
|
5
|
+
/** GET a URL, following redirects manually so the chain is recorded. */
|
|
6
|
+
export async function fetchPage(url, options = {}, readBody = true) {
|
|
7
|
+
const timeoutMs = options.timeoutMs ?? 15_000;
|
|
8
|
+
const maxRedirects = options.maxRedirects ?? 10;
|
|
9
|
+
const started = performance.now();
|
|
10
|
+
const chain = [];
|
|
11
|
+
let current = url;
|
|
12
|
+
for (let hop = 0; hop <= maxRedirects; hop++) {
|
|
13
|
+
const controller = new AbortController();
|
|
14
|
+
const timer = setTimeout(() => controller.abort(), timeoutMs);
|
|
15
|
+
try {
|
|
16
|
+
const res = await fetch(current, {
|
|
17
|
+
method: 'GET',
|
|
18
|
+
redirect: 'manual',
|
|
19
|
+
signal: controller.signal,
|
|
20
|
+
headers: {
|
|
21
|
+
'user-agent': options.userAgent ?? DEFAULT_USER_AGENT,
|
|
22
|
+
accept: 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
|
|
23
|
+
'accept-language': 'en,*;q=0.5',
|
|
24
|
+
...(options.headers ?? {}),
|
|
25
|
+
},
|
|
26
|
+
});
|
|
27
|
+
const headers = pickHeaders(res.headers);
|
|
28
|
+
if (REDIRECT_STATUSES.has(res.status) && headers['location']) {
|
|
29
|
+
chain.push({ url: current, status: res.status });
|
|
30
|
+
let nextUrl;
|
|
31
|
+
try {
|
|
32
|
+
nextUrl = new URL(headers['location'], current).href;
|
|
33
|
+
}
|
|
34
|
+
catch {
|
|
35
|
+
return result(url, current, res.status, headers, null, chain, started, `Invalid Location header: ${headers['location']}`);
|
|
36
|
+
}
|
|
37
|
+
await res.body?.cancel().catch(() => undefined);
|
|
38
|
+
if (chain.some((h) => h.url === nextUrl) || nextUrl === current) {
|
|
39
|
+
return result(url, nextUrl, res.status, headers, null, chain, started, 'Redirect loop');
|
|
40
|
+
}
|
|
41
|
+
current = nextUrl;
|
|
42
|
+
continue;
|
|
43
|
+
}
|
|
44
|
+
let body = null;
|
|
45
|
+
if (readBody)
|
|
46
|
+
body = await readLimited(res, options.maxBodyBytes ?? 5_000_000);
|
|
47
|
+
else
|
|
48
|
+
await res.body?.cancel().catch(() => undefined);
|
|
49
|
+
return result(url, current, res.status, headers, body, chain, started);
|
|
50
|
+
}
|
|
51
|
+
catch (err) {
|
|
52
|
+
const message = err instanceof Error ? (err.name === 'AbortError' ? `Timed out after ${timeoutMs} ms` : err.message) : String(err);
|
|
53
|
+
return result(url, current, null, {}, null, chain, started, message);
|
|
54
|
+
}
|
|
55
|
+
finally {
|
|
56
|
+
clearTimeout(timer);
|
|
57
|
+
}
|
|
58
|
+
}
|
|
59
|
+
return result(url, current, null, {}, null, chain, started, `More than ${maxRedirects} redirects`);
|
|
60
|
+
}
|
|
61
|
+
/** Status/redirect check without downloading the body. */
|
|
62
|
+
export function probeUrl(url, options = {}) {
|
|
63
|
+
return fetchPage(url, options, false);
|
|
64
|
+
}
|
|
65
|
+
async function readLimited(res, maxBytes) {
|
|
66
|
+
if (!res.body)
|
|
67
|
+
return '';
|
|
68
|
+
const reader = res.body.getReader();
|
|
69
|
+
const chunks = [];
|
|
70
|
+
let total = 0;
|
|
71
|
+
for (;;) {
|
|
72
|
+
const { done, value } = await reader.read();
|
|
73
|
+
if (done)
|
|
74
|
+
break;
|
|
75
|
+
if (value) {
|
|
76
|
+
chunks.push(value);
|
|
77
|
+
total += value.byteLength;
|
|
78
|
+
if (total >= maxBytes) {
|
|
79
|
+
await reader.cancel().catch(() => undefined);
|
|
80
|
+
break;
|
|
81
|
+
}
|
|
82
|
+
}
|
|
83
|
+
}
|
|
84
|
+
const charset = /charset=([\w-]+)/i.exec(res.headers.get('content-type') ?? '')?.[1];
|
|
85
|
+
try {
|
|
86
|
+
return new TextDecoder(charset ?? 'utf-8').decode(Buffer.concat(chunks));
|
|
87
|
+
}
|
|
88
|
+
catch {
|
|
89
|
+
return new TextDecoder('utf-8').decode(Buffer.concat(chunks));
|
|
90
|
+
}
|
|
91
|
+
}
|
|
92
|
+
function pickHeaders(headers) {
|
|
93
|
+
const out = {};
|
|
94
|
+
for (const name of KEPT_HEADERS) {
|
|
95
|
+
const value = headers.get(name);
|
|
96
|
+
if (value !== null)
|
|
97
|
+
out[name] = value;
|
|
98
|
+
}
|
|
99
|
+
return out;
|
|
100
|
+
}
|
|
101
|
+
function result(requestedUrl, finalUrl, status, headers, body, chain, started, error) {
|
|
102
|
+
const r = {
|
|
103
|
+
requestedUrl,
|
|
104
|
+
finalUrl,
|
|
105
|
+
status,
|
|
106
|
+
headers,
|
|
107
|
+
body,
|
|
108
|
+
redirectChain: chain,
|
|
109
|
+
timeMs: Math.round(performance.now() - started),
|
|
110
|
+
contentType: headers['content-type'] ?? null,
|
|
111
|
+
};
|
|
112
|
+
if (error)
|
|
113
|
+
r.error = error;
|
|
114
|
+
return r;
|
|
115
|
+
}
|
|
116
|
+
export function toResponseInfo(r) {
|
|
117
|
+
const info = {
|
|
118
|
+
status: r.status,
|
|
119
|
+
finalUrl: r.finalUrl,
|
|
120
|
+
redirectChain: r.redirectChain,
|
|
121
|
+
timeMs: r.timeMs,
|
|
122
|
+
contentType: r.contentType,
|
|
123
|
+
headers: r.headers,
|
|
124
|
+
};
|
|
125
|
+
if (r.error)
|
|
126
|
+
info.error = r.error;
|
|
127
|
+
return info;
|
|
128
|
+
}
|
|
129
|
+
export function isHtmlResponse(r) {
|
|
130
|
+
const ct = r.contentType?.toLowerCase() ?? '';
|
|
131
|
+
if (ct)
|
|
132
|
+
return ct.includes('text/html') || ct.includes('application/xhtml');
|
|
133
|
+
return typeof r.body === 'string' && /<html|<!doctype/i.test(r.body.slice(0, 2000));
|
|
134
|
+
}
|
|
135
|
+
//# sourceMappingURL=fetch.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"fetch.js","sourceRoot":"","sources":["../../src/core/fetch.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,eAAe,CAAC;AAGxC,MAAM,CAAC,MAAM,kBAAkB,GAAG,8CAA8C,OAAO,wDAAwD,CAAC;AAEhJ,MAAM,YAAY,GAAG,CAAC,cAAc,EAAE,cAAc,EAAE,MAAM,EAAE,kBAAkB,EAAE,eAAe,EAAE,eAAe,EAAE,UAAU,EAAE,gBAAgB,EAAE,MAAM,CAAC,CAAC;AAuB1J,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7D,wEAAwE;AACxE,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,GAAW,EAAE,UAAwB,EAAE,EAAE,QAAQ,GAAG,IAAI;IACtF,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,IAAI,MAAM,CAAC;IAC9C,MAAM,YAAY,GAAG,OAAO,CAAC,YAAY,IAAI,EAAE,CAAC;IAChD,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;IAClC,MAAM,KAAK,GAAkB,EAAE,CAAC;IAChC,IAAI,OAAO,GAAG,GAAG,CAAC;IAElB,KAAK,IAAI,GAAG,GAAG,CAAC,EAAE,GAAG,IAAI,YAAY,EAAE,GAAG,EAAE,EAAE,CAAC;QAC7C,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAC;QACzC,MAAM,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,CAAC,CAAC;QAC9D,IAAI,CAAC;YACH,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,OAAO,EAAE;gBAC/B,MAAM,EAAE,KAAK;gBACb,QAAQ,EAAE,QAAQ;gBAClB,MAAM,EAAE,UAAU,CAAC,MAAM;gBACzB,OAAO,EAAE;oBACP,YAAY,EAAE,OAAO,CAAC,SAAS,IAAI,kBAAkB;oBACrD,MAAM,EAAE,iEAAiE;oBACzE,iBAAiB,EAAE,YAAY;oBAC/B,GAAG,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC;iBAC3B;aACF,CAAC,CAAC;YACH,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;YACzC,IAAI,iBAAiB,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC;gBAC7D,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC;gBACjD,IAAI,OAAe,CAAC;gBACpB,IAAI,CAAC;oBACH,OAAO,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC;gBACvD,CAAC;gBAAC,MAAM,CAAC;oBACP,OAAO,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,GAAG,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,4BAA4B,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC;gBAC5H,CAAC;gBACD,MAAM,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;gBAChD,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,OAAO,CAAC,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;oBAChE,OAAO,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,GAAG,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,eAAe,CAAC,CAAC;gBAC1F,CAAC;gBACD,OAAO,GAAG,OAAO,CAAC;gBAClB,SAAS;YACX,CAAC;YACD,IAAI,IAAI,GAAkB,IAAI,CAAC;YAC/B,IAAI,QAAQ;gBAAE,IAAI,GAAG,MAAM,WAAW,CAAC,GAAG,EAAE,OAAO,CAAC,YAAY,IAAI,SAAS,CAAC,CAAC;;gBAC1E,MAAM,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;YACrD,OAAO,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,GAAG,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC;QACzE,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,OAAO,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,KAAK,YAAY,CAAC,CAAC,CAAC,mBAAmB,SAAS,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YACnI,OAAO,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;QACvE,CAAC;gBAAS,CAAC;YACT,YAAY,CAAC,KAAK,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,aAAa,YAAY,YAAY,CAAC,CAAC;AACrG,CAAC;AAED,0DAA0D;AAC1D,MAAM,UAAU,QAAQ,CAAC,GAAW,EAAE,UAAwB,EAAE;IAC9D,OAAO,SAAS,CAAC,GAAG,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC;AACxC,CAAC;AAED,KAAK,UAAU,WAAW,CAAC,GAAa,EAAE,QAAgB;IACxD,IAAI,CAAC,GAAG,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IACzB,MAAM,MAAM,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;IACpC,MAAM,MAAM,GAAiB,EAAE,CAAC;IAChC,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,SAAS,CAAC;QACR,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QAC5C,IAAI,IAAI;YAAE,MAAM;QAChB,IAAI,KAAK,EAAE,CAAC;YACV,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,KAAK,IAAI,KAAK,CAAC,UAAU,CAAC;YAC1B,IAAI,KAAK,IAAI,QAAQ,EAAE,CAAC;gBACtB,MAAM,MAAM,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;gBAC7C,MAAM;YACR,CAAC;QACH,CAAC;IACH,CAAC;IACD,MAAM,OAAO,GAAG,mBAAmB,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACrF,IAAI,CAAC;QACH,OAAO,IAAI,WAAW,CAAC,OAAO,IAAI,OAAO,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;IAC3E,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;IAChE,CAAC;AACH,CAAC;AAED,SAAS,WAAW,CAAC,OAAgB;IACnC,MAAM,GAAG,GAA2B,EAAE,CAAC;IACvC,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;QAChC,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAChC,IAAI,KAAK,KAAK,IAAI;YAAE,GAAG,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC;IACxC,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,MAAM,CAAC,YAAoB,EAAE,QAAgB,EAAE,MAAqB,EAAE,OAA+B,EAAE,IAAmB,EAAE,KAAoB,EAAE,OAAe,EAAE,KAAc;IACxL,MAAM,CAAC,GAAgB;QACrB,YAAY;QACZ,QAAQ;QACR,MAAM;QACN,OAAO;QACP,IAAI;QACJ,aAAa,EAAE,KAAK;QACpB,MAAM,EAAE,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,GAAG,EAAE,GAAG,OAAO,CAAC;QAC/C,WAAW,EAAE,OAAO,CAAC,cAAc,CAAC,IAAI,IAAI;KAC7C,CAAC;IACF,IAAI,KAAK;QAAE,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC;IAC3B,OAAO,CAAC,CAAC;AACX,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,CAAc;IAC3C,MAAM,IAAI,GAAiB;QACzB,MAAM,EAAE,CAAC,CAAC,MAAM;QAChB,QAAQ,EAAE,CAAC,CAAC,QAAQ;QACpB,aAAa,EAAE,CAAC,CAAC,aAAa;QAC9B,MAAM,EAAE,CAAC,CAAC,MAAM;QAChB,WAAW,EAAE,CAAC,CAAC,WAAW;QAC1B,OAAO,EAAE,CAAC,CAAC,OAAO;KACnB,CAAC;IACF,IAAI,CAAC,CAAC,KAAK;QAAE,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;IAClC,OAAO,IAAI,CAAC;AACd,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,CAAc;IAC3C,MAAM,EAAE,GAAG,CAAC,CAAC,WAAW,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;IAC9C,IAAI,EAAE;QAAE,OAAO,EAAE,CAAC,QAAQ,CAAC,WAAW,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,mBAAmB,CAAC,CAAC;IAC5E,OAAO,OAAO,CAAC,CAAC,IAAI,KAAK,QAAQ,IAAI,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC;AACtF,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
import type { ExtractedPage } from '../types.js';
|
|
2
|
+
export interface ExtractOptions {
|
|
3
|
+
/** Absolute URL of the document; relative links are resolved against it (or against <base href>). */
|
|
4
|
+
url: string;
|
|
5
|
+
/** Treat subdomains of the site host as internal. */
|
|
6
|
+
includeSubdomains?: boolean;
|
|
7
|
+
/** Host used to classify links as internal. Defaults to the host of `url`. */
|
|
8
|
+
siteHost?: string;
|
|
9
|
+
}
|
|
10
|
+
export declare function extractPage(html: string, options: ExtractOptions): ExtractedPage;
|
|
@@ -0,0 +1,150 @@
|
|
|
1
|
+
import { load } from 'cheerio';
|
|
2
|
+
import { isSameSite, resolveUrl, hostOf } from '../url.js';
|
|
3
|
+
import { collapseWhitespace, wordCount } from './text.js';
|
|
4
|
+
import { extractStructuredData } from './structured-data.js';
|
|
5
|
+
export function extractPage(html, options) {
|
|
6
|
+
const $ = load(html);
|
|
7
|
+
const hasDoctype = /^\s*<!doctype\s+html/i.test(html);
|
|
8
|
+
const baseHref = $('base[href]').first().attr('href');
|
|
9
|
+
const base = (baseHref && resolveUrl(baseHref, options.url)) || options.url;
|
|
10
|
+
const siteHost = options.siteHost ?? hostOf(options.url) ?? '';
|
|
11
|
+
const titles = $('title')
|
|
12
|
+
.toArray()
|
|
13
|
+
.map((el) => collapseWhitespace($(el).text()));
|
|
14
|
+
const metaDescriptions = $('meta[name]')
|
|
15
|
+
.toArray()
|
|
16
|
+
.filter((el) => ($(el).attr('name') ?? '').trim().toLowerCase() === 'description')
|
|
17
|
+
.map((el) => collapseWhitespace($(el).attr('content') ?? ''));
|
|
18
|
+
const metaRobots = $('meta[name]')
|
|
19
|
+
.toArray()
|
|
20
|
+
.filter((el) => ['robots', 'googlebot', 'googlebot-news', 'bingbot'].includes(($(el).attr('name') ?? '').trim().toLowerCase()))
|
|
21
|
+
.map((el) => collapseWhitespace($(el).attr('content') ?? '').toLowerCase())
|
|
22
|
+
.filter(Boolean);
|
|
23
|
+
const canonicals = $('link[rel]')
|
|
24
|
+
.toArray()
|
|
25
|
+
.filter((el) => relList($(el).attr('rel')).includes('canonical'))
|
|
26
|
+
.map((el) => ($(el).attr('href') ?? '').trim());
|
|
27
|
+
const headings = $('h1, h2, h3, h4, h5, h6')
|
|
28
|
+
.toArray()
|
|
29
|
+
.map((el) => ({ level: Number(el.tagName.slice(1)), text: collapseWhitespace($(el).text()) }));
|
|
30
|
+
const og = {};
|
|
31
|
+
const twitter = {};
|
|
32
|
+
$('meta[property], meta[name]').each((_, el) => {
|
|
33
|
+
const key = ($(el).attr('property') ?? $(el).attr('name') ?? '').trim().toLowerCase();
|
|
34
|
+
const content = collapseWhitespace($(el).attr('content') ?? '');
|
|
35
|
+
if (!key)
|
|
36
|
+
return;
|
|
37
|
+
if (key.startsWith('og:') && !(key in og))
|
|
38
|
+
og[key.slice(3)] = content;
|
|
39
|
+
else if (key.startsWith('twitter:') && !(key in twitter))
|
|
40
|
+
twitter[key.slice(8)] = content;
|
|
41
|
+
});
|
|
42
|
+
const images = $('img')
|
|
43
|
+
.toArray()
|
|
44
|
+
.map((el) => ({
|
|
45
|
+
src: $(el).attr('src') ?? $(el).attr('data-src') ?? null,
|
|
46
|
+
alt: $(el).attr('alt') ?? null,
|
|
47
|
+
width: $(el).attr('width') ?? null,
|
|
48
|
+
height: $(el).attr('height') ?? null,
|
|
49
|
+
loading: $(el).attr('loading') ?? null,
|
|
50
|
+
}));
|
|
51
|
+
const links = $('a[href]')
|
|
52
|
+
.toArray()
|
|
53
|
+
.map((el) => {
|
|
54
|
+
const href = ($(el).attr('href') ?? '').trim();
|
|
55
|
+
// Fragment-only and empty hrefs are not links to other documents.
|
|
56
|
+
const url = href === '' || href.startsWith('#') ? null : resolveUrl(href, base);
|
|
57
|
+
const rel = $(el).attr('rel') ?? null;
|
|
58
|
+
const relTokens = relList(rel);
|
|
59
|
+
const text = collapseWhitespace($(el).text());
|
|
60
|
+
const ariaLabel = collapseWhitespace($(el).attr('aria-label') ?? '');
|
|
61
|
+
const title = collapseWhitespace($(el).attr('title') ?? '');
|
|
62
|
+
const imgAlt = $(el)
|
|
63
|
+
.find('img[alt]')
|
|
64
|
+
.toArray()
|
|
65
|
+
.map((img) => collapseWhitespace($(img).attr('alt') ?? ''))
|
|
66
|
+
.find(Boolean);
|
|
67
|
+
return {
|
|
68
|
+
href,
|
|
69
|
+
url,
|
|
70
|
+
text,
|
|
71
|
+
rel,
|
|
72
|
+
internal: url ? isSameSite(url, siteHost, options.includeSubdomains) : false,
|
|
73
|
+
nofollow: relTokens.includes('nofollow'),
|
|
74
|
+
hasAccessibleName: Boolean(text || ariaLabel || title || imgAlt),
|
|
75
|
+
};
|
|
76
|
+
});
|
|
77
|
+
const hreflang = $('link[rel]')
|
|
78
|
+
.toArray()
|
|
79
|
+
.filter((el) => relList($(el).attr('rel')).includes('alternate') && $(el).attr('hreflang'))
|
|
80
|
+
.map((el) => {
|
|
81
|
+
const href = ($(el).attr('href') ?? '').trim();
|
|
82
|
+
return { lang: ($(el).attr('hreflang') ?? '').trim(), href, url: resolveUrl(href, base) };
|
|
83
|
+
});
|
|
84
|
+
const structured = extractStructuredData($);
|
|
85
|
+
const text = visibleText($);
|
|
86
|
+
return {
|
|
87
|
+
url: options.url,
|
|
88
|
+
hasDoctype,
|
|
89
|
+
lang: attrOrNull($('html').attr('lang')),
|
|
90
|
+
charset: detectCharset($),
|
|
91
|
+
viewport: attrOrNull($('meta[name="viewport" i]').first().attr('content')),
|
|
92
|
+
titles,
|
|
93
|
+
title: firstNonEmpty(titles),
|
|
94
|
+
metaDescriptions,
|
|
95
|
+
metaDescription: firstNonEmpty(metaDescriptions),
|
|
96
|
+
metaRobots,
|
|
97
|
+
canonicals,
|
|
98
|
+
canonical: canonicals[0] ? resolveUrl(canonicals[0], base) : null,
|
|
99
|
+
headings,
|
|
100
|
+
h1s: headings.filter((h) => h.level === 1).map((h) => h.text),
|
|
101
|
+
og,
|
|
102
|
+
twitter,
|
|
103
|
+
images,
|
|
104
|
+
links,
|
|
105
|
+
hreflang,
|
|
106
|
+
structuredData: structured.items,
|
|
107
|
+
structuredDataErrors: structured.errors,
|
|
108
|
+
text,
|
|
109
|
+
wordCount: wordCount(text),
|
|
110
|
+
htmlBytes: Buffer.byteLength(html, 'utf8'),
|
|
111
|
+
hasBaseHref: Boolean(baseHref),
|
|
112
|
+
};
|
|
113
|
+
}
|
|
114
|
+
function relList(rel) {
|
|
115
|
+
return (rel ?? '')
|
|
116
|
+
.toLowerCase()
|
|
117
|
+
.split(/\s+/)
|
|
118
|
+
.filter(Boolean);
|
|
119
|
+
}
|
|
120
|
+
function attrOrNull(value) {
|
|
121
|
+
const v = value?.trim();
|
|
122
|
+
return v ? v : null;
|
|
123
|
+
}
|
|
124
|
+
function firstNonEmpty(values) {
|
|
125
|
+
for (const v of values)
|
|
126
|
+
if (v)
|
|
127
|
+
return v;
|
|
128
|
+
return null;
|
|
129
|
+
}
|
|
130
|
+
function detectCharset($) {
|
|
131
|
+
const direct = $('meta[charset]').first().attr('charset');
|
|
132
|
+
if (direct)
|
|
133
|
+
return direct.trim();
|
|
134
|
+
const httpEquiv = $('meta[http-equiv]')
|
|
135
|
+
.toArray()
|
|
136
|
+
.find((el) => ($(el).attr('http-equiv') ?? '').toLowerCase() === 'content-type');
|
|
137
|
+
const content = httpEquiv ? $(httpEquiv).attr('content') : undefined;
|
|
138
|
+
const match = content?.match(/charset=([\w-]+)/i);
|
|
139
|
+
return match?.[1] ?? null;
|
|
140
|
+
}
|
|
141
|
+
function visibleText($) {
|
|
142
|
+
const body = $('body').clone();
|
|
143
|
+
body.find('script, style, noscript, template, svg, iframe, head').remove();
|
|
144
|
+
// Insert spaces at block boundaries so words don't fuse.
|
|
145
|
+
body.find('p, div, li, h1, h2, h3, h4, h5, h6, br, tr, td, th, section, article, header, footer, nav, blockquote, pre, dd, dt, figcaption').each((_, el) => {
|
|
146
|
+
$(el).prepend(' ').append(' ');
|
|
147
|
+
});
|
|
148
|
+
return collapseWhitespace(body.text());
|
|
149
|
+
}
|
|
150
|
+
//# sourceMappingURL=extract.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"extract.js","sourceRoot":"","sources":["../../../src/core/html/extract.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,IAAI,EAAmB,MAAM,SAAS,CAAC;AAEhD,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM,WAAW,CAAC;AAC3D,OAAO,EAAE,kBAAkB,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AAC1D,OAAO,EAAE,qBAAqB,EAAE,MAAM,sBAAsB,CAAC;AAW7D,MAAM,UAAU,WAAW,CAAC,IAAY,EAAE,OAAuB;IAC/D,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;IACrB,MAAM,UAAU,GAAG,uBAAuB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACtD,MAAM,QAAQ,GAAG,CAAC,CAAC,YAAY,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IACtD,MAAM,IAAI,GAAG,CAAC,QAAQ,IAAI,UAAU,CAAC,QAAQ,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC;IAC5E,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC;IAE/D,MAAM,MAAM,GAAG,CAAC,CAAC,OAAO,CAAC;SACtB,OAAO,EAAE;SACT,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAEjD,MAAM,gBAAgB,GAAG,CAAC,CAAC,YAAY,CAAC;SACrC,OAAO,EAAE;SACT,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,KAAK,aAAa,CAAC;SACjF,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAEhE,MAAM,UAAU,GAAG,CAAC,CAAC,YAAY,CAAC;SAC/B,OAAO,EAAE;SACT,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,QAAQ,EAAE,WAAW,EAAE,gBAAgB,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,CAAC;SAC9H,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;SAC1E,MAAM,CAAC,OAAO,CAAC,CAAC;IAEnB,MAAM,UAAU,GAAG,CAAC,CAAC,WAAW,CAAC;SAC9B,OAAO,EAAE;SACT,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC;SAChE,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IAElD,MAAM,QAAQ,GAAkB,CAAC,CAAC,wBAAwB,CAAC;SACxD,OAAO,EAAE;SACT,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;IAEjG,MAAM,EAAE,GAA2B,EAAE,CAAC;IACtC,MAAM,OAAO,GAA2B,EAAE,CAAC;IAC3C,CAAC,CAAC,4BAA4B,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;QAC7C,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QACtF,MAAM,OAAO,GAAG,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC;QAChE,IAAI,CAAC,GAAG;YAAE,OAAO;QACjB,IAAI,GAAG,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,EAAE,CAAC;YAAE,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC;aACjE,IAAI,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,OAAO,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC;IAC5F,CAAC,CAAC,CAAC;IAEH,MAAM,MAAM,GAAgB,CAAC,CAAC,KAAK,CAAC;SACjC,OAAO,EAAE;SACT,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;QACZ,GAAG,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,IAAI;QACxD,GAAG,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI;QAC9B,KAAK,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,IAAI;QAClC,MAAM,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,IAAI;QACpC,OAAO,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,IAAI;KACvC,CAAC,CAAC,CAAC;IAEN,MAAM,KAAK,GAAe,CAAC,CAAC,SAAS,CAAC;SACnC,OAAO,EAAE;SACT,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE;QACV,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAC/C,kEAAkE;QAClE,MAAM,GAAG,GAAG,IAAI,KAAK,EAAE,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAChF,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC;QACtC,MAAM,SAAS,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;QAC/B,MAAM,IAAI,GAAG,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAC9C,MAAM,SAAS,GAAG,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,CAAC;QACrE,MAAM,KAAK,GAAG,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;QAC5D,MAAM,MAAM,GAAG,CAAC,CAAC,EAAE,CAAC;aACjB,IAAI,CAAC,UAAU,CAAC;aAChB,OAAO,EAAE;aACT,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,kBAAkB,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC;aAC1D,IAAI,CAAC,OAAO,CAAC,CAAC;QACjB,OAAO;YACL,IAAI;YACJ,GAAG;YACH,IAAI;YACJ,GAAG;YACH,QAAQ,EAAE,GAAG,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,EAAE,QAAQ,EAAE,OAAO,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAC,KAAK;YAC5E,QAAQ,EAAE,SAAS,CAAC,QAAQ,CAAC,UAAU,CAAC;YACxC,iBAAiB,EAAE,OAAO,CAAC,IAAI,IAAI,SAAS,IAAI,KAAK,IAAI,MAAM,CAAC;SACjE,CAAC;IACJ,CAAC,CAAC,CAAC;IAEL,MAAM,QAAQ,GAAmB,CAAC,CAAC,WAAW,CAAC;SAC5C,OAAO,EAAE;SACT,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,QAAQ,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;SAC1F,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE;QACV,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAC/C,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,UAAU,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC;IAC5F,CAAC,CAAC,CAAC;IAEL,MAAM,UAAU,GAAG,qBAAqB,CAAC,CAAC,CAAC,CAAC;IAE5C,MAAM,IAAI,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;IAE5B,OAAO;QACL,GAAG,EAAE,OAAO,CAAC,GAAG;QAChB,UAAU;QACV,IAAI,EAAE,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACxC,OAAO,EAAE,aAAa,CAAC,CAAC,CAAC;QACzB,QAAQ,EAAE,UAAU,CAAC,CAAC,CAAC,yBAAyB,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAC1E,MAAM;QACN,KAAK,EAAE,aAAa,CAAC,MAAM,CAAC;QAC5B,gBAAgB;QAChB,eAAe,EAAE,aAAa,CAAC,gBAAgB,CAAC;QAChD,UAAU;QACV,UAAU;QACV,SAAS,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI;QACjE,QAAQ;QACR,GAAG,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QAC7D,EAAE;QACF,OAAO;QACP,MAAM;QACN,KAAK;QACL,QAAQ;QACR,cAAc,EAAE,UAAU,CAAC,KAAK;QAChC,oBAAoB,EAAE,UAAU,CAAC,MAAM;QACvC,IAAI;QACJ,SAAS,EAAE,SAAS,CAAC,IAAI,CAAC;QAC1B,SAAS,EAAE,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,MAAM,CAAC;QAC1C,WAAW,EAAE,OAAO,CAAC,QAAQ,CAAC;KAC/B,CAAC;AACJ,CAAC;AAED,SAAS,OAAO,CAAC,GAA8B;IAC7C,OAAO,CAAC,GAAG,IAAI,EAAE,CAAC;SACf,WAAW,EAAE;SACb,KAAK,CAAC,KAAK,CAAC;SACZ,MAAM,CAAC,OAAO,CAAC,CAAC;AACrB,CAAC;AAED,SAAS,UAAU,CAAC,KAAyB;IAC3C,MAAM,CAAC,GAAG,KAAK,EAAE,IAAI,EAAE,CAAC;IACxB,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AACtB,CAAC;AAED,SAAS,aAAa,CAAC,MAAgB;IACrC,KAAK,MAAM,CAAC,IAAI,MAAM;QAAE,IAAI,CAAC;YAAE,OAAO,CAAC,CAAC;IACxC,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,aAAa,CAAC,CAAa;IAClC,MAAM,MAAM,GAAG,CAAC,CAAC,eAAe,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IAC1D,IAAI,MAAM;QAAE,OAAO,MAAM,CAAC,IAAI,EAAE,CAAC;IACjC,MAAM,SAAS,GAAG,CAAC,CAAC,kBAAkB,CAAC;SACpC,OAAO,EAAE;SACT,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,KAAK,cAAc,CAAC,CAAC;IACnF,MAAM,OAAO,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACrE,MAAM,KAAK,GAAG,OAAO,EAAE,KAAK,CAAC,mBAAmB,CAAC,CAAC;IAClD,OAAO,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC;AAC5B,CAAC;AAED,SAAS,WAAW,CAAC,CAAa;IAChC,MAAM,IAAI,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,CAAC;IAC/B,IAAI,CAAC,IAAI,CAAC,sDAAsD,CAAC,CAAC,MAAM,EAAE,CAAC;IAC3E,yDAAyD;IACzD,IAAI,CAAC,IAAI,CAAC,gIAAgI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;QACzJ,CAAC,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;IACjC,CAAC,CAAC,CAAC;IACH,OAAO,kBAAkB,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;AACzC,CAAC"}
|
|
@@ -0,0 +1,11 @@
|
|
|
1
|
+
import type { CheerioAPI } from 'cheerio';
|
|
2
|
+
import type { StructuredDataError, StructuredDataItem } from '../types.js';
|
|
3
|
+
export interface StructuredDataExtraction {
|
|
4
|
+
items: StructuredDataItem[];
|
|
5
|
+
errors: StructuredDataError[];
|
|
6
|
+
}
|
|
7
|
+
export declare function extractStructuredData($: CheerioAPI): StructuredDataExtraction;
|
|
8
|
+
export declare function contextString(value: unknown): string | null;
|
|
9
|
+
export declare function typesOf(value: unknown): string[];
|
|
10
|
+
/** `https://schema.org/Article` -> `Article`, `schema:Article` -> `Article`. */
|
|
11
|
+
export declare function shortType(type: string): string;
|
|
@@ -0,0 +1,185 @@
|
|
|
1
|
+
import { collapseWhitespace } from './text.js';
|
|
2
|
+
export function extractStructuredData($) {
|
|
3
|
+
const items = [];
|
|
4
|
+
const errors = [];
|
|
5
|
+
$('script[type]').each((_, el) => {
|
|
6
|
+
const type = ($(el).attr('type') ?? '').trim().toLowerCase();
|
|
7
|
+
if (type !== 'application/ld+json')
|
|
8
|
+
return;
|
|
9
|
+
const raw = $(el).text().trim();
|
|
10
|
+
if (!raw) {
|
|
11
|
+
errors.push({ source: 'json-ld', message: 'Empty JSON-LD script block', snippet: '' });
|
|
12
|
+
return;
|
|
13
|
+
}
|
|
14
|
+
let parsed;
|
|
15
|
+
try {
|
|
16
|
+
parsed = JSON.parse(stripCdata(raw));
|
|
17
|
+
}
|
|
18
|
+
catch (err) {
|
|
19
|
+
errors.push({
|
|
20
|
+
source: 'json-ld',
|
|
21
|
+
message: `Invalid JSON: ${err.message}`,
|
|
22
|
+
snippet: raw.slice(0, 200),
|
|
23
|
+
});
|
|
24
|
+
return;
|
|
25
|
+
}
|
|
26
|
+
for (const node of flattenJsonLd(parsed, null))
|
|
27
|
+
items.push(node);
|
|
28
|
+
});
|
|
29
|
+
for (const item of extractMicrodata($))
|
|
30
|
+
items.push(item);
|
|
31
|
+
return { items, errors };
|
|
32
|
+
}
|
|
33
|
+
function stripCdata(text) {
|
|
34
|
+
return text
|
|
35
|
+
.replace(/^\s*<!--/, '')
|
|
36
|
+
.replace(/-->\s*$/, '')
|
|
37
|
+
.replace(/^\s*\/\*\s*<!\[CDATA\[\s*\*\//, '')
|
|
38
|
+
.replace(/\/\*\s*\]\]>\s*\*\/\s*$/, '')
|
|
39
|
+
.replace(/^\s*<!\[CDATA\[/, '')
|
|
40
|
+
.replace(/\]\]>\s*$/, '');
|
|
41
|
+
}
|
|
42
|
+
function flattenJsonLd(node, inheritedContext) {
|
|
43
|
+
if (Array.isArray(node))
|
|
44
|
+
return node.flatMap((n) => flattenJsonLd(n, inheritedContext));
|
|
45
|
+
if (!node || typeof node !== 'object')
|
|
46
|
+
return [];
|
|
47
|
+
const obj = node;
|
|
48
|
+
const context = contextString(obj['@context']) ?? inheritedContext;
|
|
49
|
+
if (Array.isArray(obj['@graph'])) {
|
|
50
|
+
return obj['@graph'].flatMap((n) => flattenJsonLd(n, context));
|
|
51
|
+
}
|
|
52
|
+
return [{ source: 'json-ld', types: typesOf(obj['@type']), context, data: obj }];
|
|
53
|
+
}
|
|
54
|
+
export function contextString(value) {
|
|
55
|
+
if (typeof value === 'string')
|
|
56
|
+
return value;
|
|
57
|
+
if (Array.isArray(value)) {
|
|
58
|
+
for (const v of value) {
|
|
59
|
+
const s = contextString(v);
|
|
60
|
+
if (s)
|
|
61
|
+
return s;
|
|
62
|
+
}
|
|
63
|
+
return null;
|
|
64
|
+
}
|
|
65
|
+
if (value && typeof value === 'object') {
|
|
66
|
+
const vocab = value['@vocab'];
|
|
67
|
+
if (typeof vocab === 'string')
|
|
68
|
+
return vocab;
|
|
69
|
+
}
|
|
70
|
+
return null;
|
|
71
|
+
}
|
|
72
|
+
export function typesOf(value) {
|
|
73
|
+
if (typeof value === 'string')
|
|
74
|
+
return [shortType(value)];
|
|
75
|
+
if (Array.isArray(value))
|
|
76
|
+
return value.filter((v) => typeof v === 'string').map(shortType);
|
|
77
|
+
return [];
|
|
78
|
+
}
|
|
79
|
+
/** `https://schema.org/Article` -> `Article`, `schema:Article` -> `Article`. */
|
|
80
|
+
export function shortType(type) {
|
|
81
|
+
const t = type.trim();
|
|
82
|
+
const slash = t.lastIndexOf('/');
|
|
83
|
+
const colon = t.lastIndexOf(':');
|
|
84
|
+
const cut = Math.max(slash, t.startsWith('http') ? -1 : colon);
|
|
85
|
+
return cut >= 0 ? t.slice(cut + 1) : t;
|
|
86
|
+
}
|
|
87
|
+
/* ---------------------------------------------------------------------------
|
|
88
|
+
* Microdata
|
|
89
|
+
* ------------------------------------------------------------------------- */
|
|
90
|
+
function extractMicrodata($) {
|
|
91
|
+
const items = [];
|
|
92
|
+
$('[itemscope]').each((_, el) => {
|
|
93
|
+
// Top-level scopes only: nested scopes become property values.
|
|
94
|
+
if ($(el).parents('[itemscope]').length > 0)
|
|
95
|
+
return;
|
|
96
|
+
const data = microdataObject($, el);
|
|
97
|
+
items.push({
|
|
98
|
+
source: 'microdata',
|
|
99
|
+
types: typesOf(data['@type']),
|
|
100
|
+
context: contextString(data['@context']),
|
|
101
|
+
data,
|
|
102
|
+
});
|
|
103
|
+
});
|
|
104
|
+
return items;
|
|
105
|
+
}
|
|
106
|
+
function microdataObject($, scope) {
|
|
107
|
+
const obj = {};
|
|
108
|
+
const itemtype = $(scope).attr('itemtype');
|
|
109
|
+
if (itemtype) {
|
|
110
|
+
const types = itemtype.split(/\s+/).filter(Boolean);
|
|
111
|
+
obj['@type'] = types.length === 1 ? types[0] : types;
|
|
112
|
+
const first = types[0];
|
|
113
|
+
if (first) {
|
|
114
|
+
try {
|
|
115
|
+
const u = new URL(first);
|
|
116
|
+
obj['@context'] = `${u.protocol}//${u.host}`;
|
|
117
|
+
}
|
|
118
|
+
catch {
|
|
119
|
+
/* not a URL */
|
|
120
|
+
}
|
|
121
|
+
}
|
|
122
|
+
}
|
|
123
|
+
const itemid = $(scope).attr('itemid');
|
|
124
|
+
if (itemid)
|
|
125
|
+
obj['@id'] = itemid;
|
|
126
|
+
const props = [];
|
|
127
|
+
collectProps($, scope, props);
|
|
128
|
+
for (const el of props) {
|
|
129
|
+
const names = ($(el).attr('itemprop') ?? '').split(/\s+/).filter(Boolean);
|
|
130
|
+
const value = $(el).is('[itemscope]') ? microdataObject($, el) : microdataValue($, el);
|
|
131
|
+
for (const name of names) {
|
|
132
|
+
const existing = obj[name];
|
|
133
|
+
if (existing === undefined)
|
|
134
|
+
obj[name] = value;
|
|
135
|
+
else if (Array.isArray(existing))
|
|
136
|
+
existing.push(value);
|
|
137
|
+
else
|
|
138
|
+
obj[name] = [existing, value];
|
|
139
|
+
}
|
|
140
|
+
}
|
|
141
|
+
return obj;
|
|
142
|
+
}
|
|
143
|
+
/** Walk descendants, stopping at nested itemscopes (they are added as a single property). */
|
|
144
|
+
function collectProps($, root, out) {
|
|
145
|
+
const children = $(root).children().toArray();
|
|
146
|
+
for (const child of children) {
|
|
147
|
+
const hasProp = $(child).is('[itemprop]');
|
|
148
|
+
if (hasProp)
|
|
149
|
+
out.push(child);
|
|
150
|
+
if (!$(child).is('[itemscope]'))
|
|
151
|
+
collectProps($, child, out);
|
|
152
|
+
}
|
|
153
|
+
}
|
|
154
|
+
function microdataValue($, el) {
|
|
155
|
+
const $el = $(el);
|
|
156
|
+
const tag = (el.tagName ?? '').toLowerCase();
|
|
157
|
+
switch (tag) {
|
|
158
|
+
case 'meta':
|
|
159
|
+
return $el.attr('content') ?? '';
|
|
160
|
+
case 'img':
|
|
161
|
+
case 'audio':
|
|
162
|
+
case 'embed':
|
|
163
|
+
case 'iframe':
|
|
164
|
+
case 'source':
|
|
165
|
+
case 'track':
|
|
166
|
+
case 'video':
|
|
167
|
+
return $el.attr('src') ?? '';
|
|
168
|
+
case 'a':
|
|
169
|
+
case 'area':
|
|
170
|
+
case 'link':
|
|
171
|
+
return $el.attr('href') ?? '';
|
|
172
|
+
case 'object':
|
|
173
|
+
return $el.attr('data') ?? '';
|
|
174
|
+
case 'data':
|
|
175
|
+
case 'meter':
|
|
176
|
+
return $el.attr('value') ?? collapseWhitespace($el.text());
|
|
177
|
+
case 'time':
|
|
178
|
+
return $el.attr('datetime') ?? collapseWhitespace($el.text());
|
|
179
|
+
default: {
|
|
180
|
+
const content = $el.attr('content');
|
|
181
|
+
return content !== undefined ? content : collapseWhitespace($el.text());
|
|
182
|
+
}
|
|
183
|
+
}
|
|
184
|
+
}
|
|
185
|
+
//# sourceMappingURL=structured-data.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"structured-data.js","sourceRoot":"","sources":["../../../src/core/html/structured-data.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,kBAAkB,EAAE,MAAM,WAAW,CAAC;AAO/C,MAAM,UAAU,qBAAqB,CAAC,CAAa;IACjD,MAAM,KAAK,GAAyB,EAAE,CAAC;IACvC,MAAM,MAAM,GAA0B,EAAE,CAAC;IAEzC,CAAC,CAAC,cAAc,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;QAC/B,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QAC7D,IAAI,IAAI,KAAK,qBAAqB;YAAE,OAAO;QAC3C,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;QAChC,IAAI,CAAC,GAAG,EAAE,CAAC;YACT,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,4BAA4B,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC,CAAC;YACvF,OAAO;QACT,CAAC;QACD,IAAI,MAAe,CAAC;QACpB,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC;QACvC,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,CAAC,IAAI,CAAC;gBACV,MAAM,EAAE,SAAS;gBACjB,OAAO,EAAE,iBAAkB,GAAa,CAAC,OAAO,EAAE;gBAClD,OAAO,EAAE,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC;aAC3B,CAAC,CAAC;YACH,OAAO;QACT,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC;YAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnE,CAAC,CAAC,CAAC;IAEH,KAAK,MAAM,IAAI,IAAI,gBAAgB,CAAC,CAAC,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAEzD,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;AAC3B,CAAC;AAED,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,IAAI;SACR,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC;SACvB,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC;SACtB,OAAO,CAAC,+BAA+B,EAAE,EAAE,CAAC;SAC5C,OAAO,CAAC,yBAAyB,EAAE,EAAE,CAAC;SACtC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC;SAC9B,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;AAC9B,CAAC;AAED,SAAS,aAAa,CAAC,IAAa,EAAE,gBAA+B;IACnE,IAAI,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,aAAa,CAAC,CAAC,EAAE,gBAAgB,CAAC,CAAC,CAAC;IACxF,IAAI,CAAC,IAAI,IAAI,OAAO,IAAI,KAAK,QAAQ;QAAE,OAAO,EAAE,CAAC;IACjD,MAAM,GAAG,GAAG,IAA+B,CAAC;IAC5C,MAAM,OAAO,GAAG,aAAa,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,IAAI,gBAAgB,CAAC;IACnE,IAAI,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC;QACjC,OAAQ,GAAG,CAAC,QAAQ,CAAe,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,aAAa,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;IAChF,CAAC;IACD,OAAO,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC,CAAC;AACnF,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,KAAc;IAC1C,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC;IAC5C,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QACzB,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;YACtB,MAAM,CAAC,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;YAC3B,IAAI,CAAC;gBAAE,OAAO,CAAC,CAAC;QAClB,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IACD,IAAI,KAAK,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QACvC,MAAM,KAAK,GAAI,KAAiC,CAAC,QAAQ,CAAC,CAAC;QAC3D,IAAI,OAAO,KAAK,KAAK,QAAQ;YAAE,OAAO,KAAK,CAAC;IAC9C,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,MAAM,UAAU,OAAO,CAAC,KAAc;IACpC,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC;IACzD,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;IACxG,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,gFAAgF;AAChF,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,MAAM,KAAK,GAAG,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IACjC,MAAM,KAAK,GAAG,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IACjC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;IAC/D,OAAO,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACzC,CAAC;AAED;;+EAE+E;AAE/E,SAAS,gBAAgB,CAAC,CAAa;IACrC,MAAM,KAAK,GAAyB,EAAE,CAAC;IACvC,CAAC,CAAC,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;QAC9B,+DAA+D;QAC/D,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO;QACpD,MAAM,IAAI,GAAG,eAAe,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QACpC,KAAK,CAAC,IAAI,CAAC;YACT,MAAM,EAAE,WAAW;YACnB,KAAK,EAAE,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YAC7B,OAAO,EAAE,aAAa,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YACxC,IAAI;SACL,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IACH,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,eAAe,CAAC,CAAa,EAAE,KAAc;IACpD,MAAM,GAAG,GAA4B,EAAE,CAAC;IACxC,MAAM,QAAQ,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IAC3C,IAAI,QAAQ,EAAE,CAAC;QACb,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QACpD,GAAG,CAAC,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;QACrD,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACvB,IAAI,KAAK,EAAE,CAAC;YACV,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;gBACzB,GAAG,CAAC,UAAU,CAAC,GAAG,GAAG,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;YAC/C,CAAC;YAAC,MAAM,CAAC;gBACP,eAAe;YACjB,CAAC;QACH,CAAC;IACH,CAAC;IACD,MAAM,MAAM,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IACvC,IAAI,MAAM;QAAE,GAAG,CAAC,KAAK,CAAC,GAAG,MAAM,CAAC;IAEhC,MAAM,KAAK,GAAc,EAAE,CAAC;IAC5B,YAAY,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;IAC9B,KAAK,MAAM,EAAE,IAAI,KAAK,EAAE,CAAC;QACvB,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC1E,MAAM,KAAK,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QACvF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,QAAQ,GAAG,GAAG,CAAC,IAAI,CAAC,CAAC;YAC3B,IAAI,QAAQ,KAAK,SAAS;gBAAE,GAAG,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC;iBACzC,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,CAAC;gBAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;;gBAClD,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;QACrC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,6FAA6F;AAC7F,SAAS,YAAY,CAAC,CAAa,EAAE,IAAa,EAAE,GAAc;IAChE,MAAM,QAAQ,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC,OAAO,EAAE,CAAC;IAC9C,KAAK,MAAM,KAAK,IAAI,QAAQ,EAAE,CAAC;QAC7B,MAAM,OAAO,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,YAAY,CAAC,CAAC;QAC1C,IAAI,OAAO;YAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC7B,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,aAAa,CAAC;YAAE,YAAY,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC;IAC/D,CAAC;AACH,CAAC;AAED,SAAS,cAAc,CAAC,CAAa,EAAE,EAAW;IAChD,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC;IAClB,MAAM,GAAG,GAAG,CAAE,EAAc,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IAC1D,QAAQ,GAAG,EAAE,CAAC;QACZ,KAAK,MAAM;YACT,OAAO,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;QACnC,KAAK,KAAK,CAAC;QACX,KAAK,OAAO,CAAC;QACb,KAAK,OAAO,CAAC;QACb,KAAK,QAAQ,CAAC;QACd,KAAK,QAAQ,CAAC;QACd,KAAK,OAAO,CAAC;QACb,KAAK,OAAO;YACV,OAAO,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAC/B,KAAK,GAAG,CAAC;QACT,KAAK,MAAM,CAAC;QACZ,KAAK,MAAM;YACT,OAAO,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QAChC,KAAK,QAAQ;YACX,OAAO,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QAChC,KAAK,MAAM,CAAC;QACZ,KAAK,OAAO;YACV,OAAO,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,kBAAkB,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;QAC7D,KAAK,MAAM;YACT,OAAO,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,kBAAkB,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;QAChE,OAAO,CAAC,CAAC,CAAC;YACR,MAAM,OAAO,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YACpC,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,kBAAkB,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;QAC1E,CAAC;IACH,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
/** Text analysis helpers: word counts, readability, placeholder detection. */
|
|
2
|
+
export declare function words(text: string): string[];
|
|
3
|
+
export declare function wordCount(text: string): number;
|
|
4
|
+
export declare function collapseWhitespace(text: string): string;
|
|
5
|
+
export declare function sentenceCount(text: string): number;
|
|
6
|
+
/** English syllable heuristic: vowel groups minus silent trailing e, min 1. */
|
|
7
|
+
export declare function syllables(word: string): number;
|
|
8
|
+
/**
|
|
9
|
+
* Flesch Reading Ease (English formula). Returns null when there is not enough text.
|
|
10
|
+
* 90-100 very easy, 60-70 plain English, 30-50 difficult, <30 very difficult.
|
|
11
|
+
*/
|
|
12
|
+
export declare function fleschReadingEase(text: string): number | null;
|
|
13
|
+
export declare function hasPlaceholderText(text: string): boolean;
|
|
14
|
+
export declare const GENERIC_TITLES: Set<string>;
|
|
15
|
+
export declare function isGenericTitle(title: string): boolean;
|
|
16
|
+
export declare const GENERIC_ANCHORS: Set<string>;
|
|
17
|
+
export declare function isGenericAnchor(text: string): boolean;
|
|
18
|
+
export declare function normalizeForComparison(text: string): string;
|
|
19
|
+
export declare function includesKeyword(haystack: string | null | undefined, keyword: string): boolean;
|
|
20
|
+
export declare function keywordOccurrences(text: string, keyword: string): number;
|
|
21
|
+
/** 5-word shingles for near-duplicate detection. */
|
|
22
|
+
export declare function shingles(text: string, size?: number): Set<string>;
|
|
23
|
+
export declare function jaccard(a: Set<string>, b: Set<string>): number;
|