ann_router.detect module

Hardware probing — fill in the hardware criterion automatically.

The router’s decision changes with the accelerator: a GPU unlocks FAISS batch search, Apple Silicon makes turbovec’s Rust/NEON path especially attractive, and a plain CPU box narrows the field. Rather than make the caller hand-classify their machine, this module probes it the same way the best-engine-ai-helper sibling’s detect.py does — cheap subprocess/library calls with graceful fallbacks, never raising, always returning one of the three HardwareName values.

Consumes: os_helper (worker count), optional psutil/torch if present. Produces: detect_hardware(), hardware_report().

Author: Warith Harchaoui <warith.harchaoui@deraison.ai>

ann_router.detect.detect_hardware()[source]

Classify the local accelerator into the router’s three-way taxonomy.

Returns:: "gpu" when a CUDA GPU is found (it dominates the batch regime), else "apple_silicon" on M-series Macs, else "cpu".
Return type:: {“gpu”, “apple_silicon”, “cpu”}

Examples

>>> detect_hardware() in {"gpu", "apple_silicon", "cpu"}
True

ann_router.detect.hardware_report()[source]

Return a small JSON-ready dict describing the host for the CLI/API.

Returns:: os, machine, cpu_count, workers and the classified hardware label.
Return type:: dict

Examples

>>> report = hardware_report()
>>> report["hardware"] in {"gpu", "apple_silicon", "cpu"}
True