Skip to content

api.md

Latest commit

 

History

History
89 lines (62 loc) · 2.01 KB

api.md

File metadata and controls

89 lines (62 loc) · 2.01 KB
Raw

API Reference

Core Classes

Profiler

The main profiling class that collects and manages profiling data.

from profiling_py import Profiler

profiler = Profiler(
    enable_time=True,     # Enable time profiling (default: True)
    enable_memory=False,  # Enable memory profiling (default: False)
    enable_gpu=False,     # Enable GPU profiling (default: False)
    gpu_device=0,         # GPU device index (default: 0)
    metadata=None         # Optional metadata dictionary
)

Methods

  • start_step(name: str) -> None: Start a new profiling step
  • end_step(name: str) -> Dict[str, Any]: End the current profiling step
  • profile(name: str) -> ContextManager: Context manager for profiling a code block
  • get_stats() -> Dict[str, Any]: Get current profiling statistics

Decorators

@profile

Decorator to profile a function.

from profiling_py import profile

@profile
def my_function():
    # Your code here
    pass

Context Managers

profile(name: str)

Context manager for profiling a code block.

from profiling_py import profile

with profile("my_operation"):
    # Your code here
    pass

Report Generation

generate_profiling_report()

Generate an HTML profiling report.

from profiling_py import generate_profiling_report

generate_profiling_report(
    profiler=None,           # Profiler instance (default: default_profiler)
    output_dir="./reports",  # Output directory
    file_name=None,          # Report filename (default: auto-generated)
    open_browser=True,       # Open report in browser
    measure_time=True,       # Include time measurements
    measure_ram=True,        # Include memory measurements
    measure_gpu=True         # Include GPU measurements
)

GPU Profiling

GPUProfiler

Low-level GPU profiler using NVML.

from profiling_py import GPUProfiler

with GPUProfiler(device_index=0) as gpu:
    metrics = gpu.get_gpu_metrics()
    # metrics contains GPU utilization, memory usage, etc.