feat: html loader support metadata extraction from URL

[Mini256]

part of #121

For example:

If there is a document URL as below:

https://docs.pingcap.com/tidb/v6.5/character-set-and-collation

The reader config:

{
  "reader": {
    "rag.loader.html": {
      "metadataExtraction": [
        {
          "extractors": [
            {
              "defaultMetadata": {
                "language": "en",
                "version": "stable"
              },
              "type": "url-metadata-extractor",
              "urlMetadataPattern": "/:language(zh|ja)?/:product(tidb|tidbcloud|tidb\\-data\\-migration|tidb\\-in\\-kubernetes)/:version(dev|stable|v\\d.\\d)(/.*)"
            }
          ],
          "urlPattern": "docs.pingcap.com/**"
        }
      ]
  }
}

The html loader will got the metadata as follows:

{
    "documentMetadata": {
      "language": "en",
      "product": "tidb",
      "version": "v6.5"
    }
}

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
rag	❌ Failed (Inspect)			Apr 19, 2024 0:56am
rag-docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Apr 19, 2024 0:56am
tidb-ai	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Apr 19, 2024 0:56am