Skip to content

Configuration Classes

Configuration parsing and validation classes.

Overview

GPUX uses Pydantic models for configuration parsing and validation.

from gpux.config.parser import GPUXConfigParser

parser = GPUXConfigParser()
config = parser.parse_file("gpux.yml")

Class: GPUXConfigParser

Constructor

GPUXConfigParser()

Example: 

parser = GPUXConfigParser()

Methods

parse_file()

Parse configuration from file.

parse_file(config_path: str | Path) -> GPUXConfig

Parameters:

  • config_path (str | Path): Path to gpux.yml

Returns:

  • GPUXConfig: Parsed configuration

Raises:

  • FileNotFoundError: If config file not found
  • ValueError: If config invalid

Example: 

config = parser.parse_file("gpux.yml")
print(config.name)

parse_string()

Parse configuration from string.

parse_string(config_str: str) -> GPUXConfig

Parameters:

  • config_str (str): YAML configuration string

Returns:

  • GPUXConfig: Parsed configuration

Example: 

yaml_str = """
name: model
version: 1.0.0
model:
  source: ./model.onnx
inputs:
  - name: input
    type: float32
outputs:
  - name: output
    type: float32
"""
config = parser.parse_string(yaml_str)

get_model_path()

Get absolute model path.

get_model_path(base_path: str | Path | None = None) -> Path | None

Parameters:

  • base_path (str | Path, optional): Base path for relative paths

Returns:

  • Path | None: Absolute model path

Example: 

model_path = parser.get_model_path(Path("./project"))

Configuration Models

GPUXConfig

Main configuration model.

class GPUXConfig(BaseModel):
    name: str
    version: str = "1.0.0"
    description: str | None = None
    model: ModelConfig
    inputs: list[InputConfig]
    outputs: list[OutputConfig]
    runtime: RuntimeConfig = Field(default_factory=RuntimeConfig)
    serving: ServingConfig | None = None
    preprocessing: PreprocessingConfig | None = None
    metadata: dict[str, Any] = Field(default_factory=dict)

ModelConfig

Model configuration.

class ModelConfig(BaseModel):
    source: str | Path
    format: str = "onnx"
    version: str | None = None

RuntimeConfig

Runtime configuration.

class RuntimeConfig(BaseModel):
    gpu: GPUConfig = Field(default_factory=GPUConfig)
    timeout: int = 30
    batch_size: int = 1
    enable_profiling: bool = False

ServingConfig

Serving configuration.

class ServingConfig(BaseModel):
    port: int = 8080
    host: str = "0.0.0.0"
    batch_size: int = 1
    timeout: int = 5
    max_workers: int = 4

Complete Example

from gpux.config.parser import GPUXConfigParser
from pathlib import Path

# Create parser
parser = GPUXConfigParser()

# Parse configuration
config = parser.parse_file("gpux.yml")

# Access configuration
print(f"Model: {config.name} v{config.version}")
print(f"Source: {config.model.source}")

# Runtime settings
print(f"GPU Memory: {config.runtime.gpu.memory}")
print(f"Backend: {config.runtime.gpu.backend}")
print(f"Batch Size: {config.runtime.batch_size}")

# Inputs
for inp in config.inputs:
    print(f"Input: {inp.name} ({inp.type})")

# Outputs
for out in config.outputs:
    print(f"Output: {out.name} ({out.type})")

# Get model path
model_path = parser.get_model_path(Path("."))
print(f"Model path: {model_path}")

See Also