Index PhyslibAlpha and QuantumInfo, not just Physlib

[Gabrielebattimelli]

Why

PhysLib's lakefile declares three Lean libraries:

[[lean_lib]] name = "Physlib"
[[lean_lib]] name = "PhyslibAlpha"
[[lean_lib]] name = "QuantumInfo"

MODULE_NAMES only listed Physlib, so roughly 130 of ~620 source files (~47 in PhyslibAlpha, ~84 in QuantumInfo) were never indexed and have never appeared in search.

Note: the recent runs are not failing — the last 7 scheduled runs all succeeded. The pipeline was healthy; it was just indexing a subset of the library.

What

Two changes, because jixia analyses .olean files rather than sources:

1. Index all three namespaces — MODULE_NAMES: Physlib,PhyslibAlpha,QuantumInfo
2. Build all three. PhyslibAlpha is absent from the lakefile's defaultTargets, so plain lake build never compiled it — jixia cannot analyse what lake never built. The build step now passes MODULE_NAMES to lake explicitly. Timeout raised 90 → 120 min for the extra libraries on a cold cache.
3. Assert per-namespace module counts after the load. A library that fails to build yields no modules rather than an error, so a namespace can vanish from the index silently — exactly how PhyslibAlpha stayed missing while every run reported success.

Verified

• YAML and the embedded Python parse
• parse_name('PhyslibAlpha.Foo') → ['PhyslibAlpha', 'Foo'], confirming name->>0 extracts the namespace as the new check assumes

Expect

The first run will be long — it builds two additional libraries and indexes ~130 new files through informalize and embedding.