Normalizer. What features needed out of box?

[puzrin]

Linkification process expect 2 additional steps:

1. Links normalization
2. Text normalization

Adding too many features tend to api clobbering. Instead i'd prefer to simplify modifications. Some things i know about (with unknown usefulness):

• text
  • decode encoded texts
  • drop http:// (but don't drop for links with user/pass)
  • drop www
  • cut length (leave domain & tail, cut the middle & query)
• url
  • usually, nothing needed, but it would be nice to correctly encode host + path + query

---

1. It would be interesting to get feedback from big deployments, how do they care about links.
2. Do we need to enchance normalization in this pkg, or move to plugins?

[hemanth]

Something like:

• urls-md
• ytmd

Would be nice to have out of the box.

[puzrin]

Closed - no longer actual. Methods are rewriteable.

[hemanth]

👍