…dboxInferenceConfig (#2747)

Follow-up to #3678. That PR added a Python-side conditional in
`generate-openclaw-config.py` that toggled `supportsUsageInStreaming`
when `NEMOCLAW_PROVIDER_KEY` was `"ollama"` or `"ollama-local"`. However,
in real onboard runs `getSandboxInferenceConfig("model", "ollama-local")`
hits the default switch arm and returns `providerKey = "inference"` —
the same value used by all managed-inference providers — so the
build-time `NEMOCLAW_PROVIDER_KEY` is never `"ollama"` and the
conditional was dead code. The earlier regression tests in
`test/generate-openclaw-config.test.ts` passed only because they set
`NEMOCLAW_PROVIDER_KEY=ollama` directly, which never happens in
production.

Verified locally with the python script and the real env values
NemoClaw passes through at docker build time
(`NEMOCLAW_PROVIDER_KEY=inference`,
`NEMOCLAW_INFERENCE_BASE_URL=https://inference.local/v1`,
`NEMOCLAW_INFERENCE_COMPAT_B64=e30=` for an empty compat): the
generated `openclaw.json` has `model.compat = null`, confirming the
flag never landed.

Move the wiring to the proper layer: add an explicit `"ollama-local"`
case in `getSandboxInferenceConfig` that sets
`inferenceCompat = { supportsUsageInStreaming: true }`. The existing
encode-into-`NEMOCLAW_INFERENCE_COMPAT_B64` path then carries the flag
into the python script's `inference_compat` dict, which writes it onto
`model.compat` exactly as before — except now it actually fires for
ollama-local builds. Cloud providers and other local backends keep
their existing `inferenceCompat` (or absence thereof).

Drop the now-redundant python conditional, drop the three
unrealistic-fixture vitest cases from `test/generate-openclaw-config.test.ts`,
and add two new ones in `src/lib/inference/config.test.ts` that
exercise `getSandboxInferenceConfig` with realistic provider names.

Test plan
- `npx vitest run src/lib/inference/config.test.ts` — 24/24 pass,
  including the two new ones
  (`forces supportsUsageInStreaming for ollama-local`,
  `does not force supportsUsageInStreaming for non-ollama providers`)
- `npx vitest run test/generate-openclaw-config.test.ts` — 67/67 pass
  (was 70/70 with the three deleted unrealistic cases)
- Direct script run with prod-shaped env values:
  `NEMOCLAW_INFERENCE_COMPAT_B64={"supportsUsageInStreaming":true}`
  (the shape TypeScript will emit for ollama-local) →
  `model.compat = {'supportsUsageInStreaming': True}` in generated
  openclaw.json ✓

Signed-off-by: Shawn Xie <shaxie@nvidia.com>

Resolve conflicts from the generate-openclaw-config .py->.mts migration
(landed on main in 9641ce0):

- scripts/generate-openclaw-config.py: deleted on main. This PR had only
  removed the dead `provider_key in {ollama,ollama-local}` block from it,
  so accept the deletion — the file is gone.
- test/generate-openclaw-config.test.ts: rewritten on main. This PR only
  deleted the old .py-era tests here (+0/-64), which no longer exist in
  the rewritten file, so take main's version (keeps main's new tests).

The real fix is unchanged and lives in src/lib/inference/config.ts:
getSandboxInferenceConfig("model","ollama-local") sets
inferenceCompat.supportsUsageInStreaming=true, which flows to
NEMOCLAW_INFERENCE_COMPAT_B64 (dockerfile-patch.ts) and into
openclaw.json. Note main's migration re-ported the same provider-key
-gated block into the .mts script; that block is inert in production
(NEMOCLAW_PROVIDER_KEY resolves to "inference" for ollama-local), so the
two coexist and config.ts is the layer that actually fires.

Verified: build:cli clean; 130/130 tests pass across
src/lib/inference/config.test.ts and test/generate-openclaw-config.test.ts.

Signed-off-by: Shawn Xie <shaxie@nvidia.com>

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>