fix: improve type map generation

[TusharThakur04]

Closes: #3

Summary
This PR improve the type map generation :

• use getFullName() to avoid name collisions
• filter out internal TypeDoc artifacts (__type, export=)
• filter out Reference kind duplicates
• clean export=/ from generated URLs

What kind of change does this PR introduce?
fix

Did you add tests for your changes?
no

Does this PR introduce a breaking change?
no

If relevant, what needs to be documented once your changes are merged or what have you already documented?

Use of AI
used ai for getting insight and discussion

[TusharThakur04]

rebased

[avivkeller]

@TusharThakur04 The issue isn't a matter of rebasing, the HEAD_COMMIT file needs to match the commit of webpack you generated the docs from.

Either:

1. Update HEAD_COMMIT to match
2. Use the current HEAD_COMMIT

[TusharThakur04]

thanks!!

[moshams272]

Hey @avivkeller and @TusharThakur04, great work on getting this merged!

However, I wanted to point out a couple of critical URL routing bugs in this implementation that I had actually addressed in my PR (#19):

1. Fragile .md Replacement: replace(".md", ".html") will break if a file or path has .md before the extension (e.g., utils.md5.md becomes utils.html5.md instead of utils.md5.html).
2. Multiple Hash Fragments: TypeDoc often generates multiple hashes (e.g., path.md#methods#myMethod). Browsers fail to anchor to these properly. We need to extract only the final meaningful hash fragment.

Since my PR (#19) is now conflicted, I’d be happy to open a quick follow-up PR to patch these two routing edge cases into the new implementation.

[avivkeller]

That won't happen.

That's an issue that existed before this PR, and is not a "critical routing bug". I've resolved it at the source in #26.

[avivkeller]

Please refrain from pointing out "bugs" in "this implementation" that aren't part of this PR.

[moshams272]

Please refrain from pointing out "bugs" in "this implementation" that aren't part of this PR.

My apologies for the poor phrasing, I definitely didn't mean to imply that the recent PR introduced these issues, but rather that I had noticed them during my previous testing.

Good to know the .md edge case isn't applicable to our setup, and I'm glad the hash routing was already properly addressed at the source in #26, Thanks for the clarification and for your patience. 🙏🤍