SDK and platform learning route¶
Follow a local request from submission through execution, result inspection and recovery. The five lessons use small circuits so you can concentrate on application behavior. Each has runnable source, expected observations and exercises.
| Lesson | Run from the repository root | What you will verify |
|---|---|---|
| Beginner: Backend and Job | python3 examples/user/tracks/sdk_platform_engineer/01_beginner_backend_job_en.py |
Reading status does not execute; repeated result reads use the cached result. |
| Results and diagnostics | python3 examples/user/tracks/sdk_platform_engineer/02_foundation_result_diagnostics_en.py |
A view and report retain the source result and do not rerun it. |
| Scan jobs | python3 examples/user/tracks/sdk_platform_engineer/03_applied_sweep_job_en.py |
Each point keeps its binding, status and measurements. |
| Retry, resume and history | python3 examples/user/tracks/sdk_platform_engineer/04_advanced_retry_resume_history_en.py |
Another Backend can resume queued work and reuse completed work from the same store. |
| Expert: Capability checks | python3 examples/user/tracks/sdk_platform_engineer/05_expert_capability_boundaries_en.py |
Unsupported device requests are rejected before a Job is returned. |
The scan uses a Digital-only Hybrid program and needs no Analog integrator. Both points must complete with eight counts each; missing results are not measured zeros. Persistent examples use local SQLite and result sidecars. The retry lesson configures a policy but does not induce a transient failure.
For interface details, see job lifecycle, local history and result diagnostics. Cloud and Hanyuan interfaces remain offline, dry-run or replay paths. These lessons do not implement a distributed scheduler, remote artifact service or live hardware access.