Migration: `hypermeshdb` → `hypermesh`

The SDK was repackaged under the distribution name hypermesh with a curated facade. The change is additive and backward compatible.

Nothing breaks

• import hypermeshdb still works (it is the lower-level package).
• The hmdb CLI is unchanged.
• Every existing query, connection, and analytics call behaves identically.

Recommended new style

Old	New
import hypermeshdb	import hypermesh as hm
hypermeshdb.connect("/path")	hm.connect("/path")
hypermeshdb.Client("http://…")	hm.connect("http://…") or hm.Client(...)
from hypermeshdb import Analytics	from hypermesh import Analytics
hypermeshdb.HyperMeshError	hm.exceptions.HyperMeshError

What’s new

• hm.connect() factory (path → embedded, URL → remote).
• AsyncClient, plus client auth/TLS/retries/ping()/health().
• hm.exceptions taxonomy (the same classes hypermeshdb raises).
• hm.Config.from_env() and hm.enable_debug_logging().
• hm.ingest, hm.connectors, hm.analytics, hm.reports as first-class modules.

Install changes

• The compiled engine now ships inside the wheel — no make needed for a normal install. Use the [engine] extra to be explicit.
• Heavy dependencies moved behind extras ([analytics], [pandas], [arrow], [server], …). Add the extras your code uses.

Deprecation policy

Public API follows SemVer. Symbols are deprecated in a minor release (emitting DeprecationWarning), kept for at least one further minor, and removed only in a major release.