Add inference-compiler extension

[PawelPeczek-Roboflow]

What does this PR do?

Adding CLI to compile TRT packages in two modes:

• using inference-models package installed in env (if found)
• using inference server container and inference-cli command running inside, when client calls the command w/o environment

CLI command:

inference enterprise inference-compiler compile-model --model-id yolov8n-640 --api-key XXX

Usage:

 Usage: python -m inference_cli.main enterprise inference-compiler compile-model
            [OPTIONS]

╭─ Options ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ *  --model-id                -m                                 TEXT                     Model ID in format project/version. [required]                                                                                     │
│    --api-key                 -a                                 TEXT                     Roboflow API key for your workspace. If not given - env variable `ROBOFLOW_API_KEY` will be used                                   │
│    --debug-mode                  --no-debug-mode                                         Flag enabling errors stack traces to be displayed (helpful for debugging) [default: no-debug-mode]                                 │
│    --trt-forward-compatible      --no-trt-forward-compatible                             Flag to decide if forward-compatibility mode in TRT compilation should be enabled [default: no-trt-forward-compatible]             │
│    --trt-same-cc-compatible      --no-trt-same-cc-compatible                             Flag to decide if engine should be compiled to be compatible with devices sharing the same CUDA CC to the one running compilation  │
│                                                                                          procedure                                                                                                                          │
│                                                                                          [default: no-trt-same-cc-compatible]                                                                                               │
│    --compilation-mode                                           [auto|container|python]  Selection of compilation mode - `container` runs the procedure inside `inference` server, `python` runs in-process. `auto`         │
│                                                                                          (default) inspect environment dependencies to verify if the procedure can be run in-process, if not - offloading to the server.    │
│                                                                                          [default: auto]                                                                                                                    │
│    --image                                                      TEXT                     Point specific docker image you would like to run with command (useful for development of custom builds of inference server)       │
│    --use-local-images            --not-use-local-images                                  Flag to allow using local images (if set False image is always attempted to be pulled) [default: not-use-local-images]             │
│    --env-file-path                                              TEXT                     Path to key-value .env file to inject into compilation container (if you run in Python package, just export variables to env)      │
│    --help                                                                                Show this message and exit.                                                                                                        │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

Type of Change

• New feature (non-breaking change that adds functionality)

Testing

• e2e testing, complex behaviour hard to test w/o specifically created environment and introspection of the platform state
• I have tested this change locally
• I have added/updated tests for this change

Test details:

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code where necessary, particularly in hard-to-understand areas
• My changes generate no new warnings or errors
• I have updated the documentation accordingly (if applicable)

Additional Context