Feature: Add Display Detection & Audio Detection support for Windows platform && Refactor: move systems to use C++ wmi query function, much better performance

[kernel-dev]

This PR implements display detection support for the Windows platform; more specifically:

• Helper C++ method GetGPUForDisplay(...), GetWmiInfo(...), GetNetworkHardwareInfo(...), GetAudioHardwareInfo(...)
  • Gets compiled to a DLL, inside of src/pysysinfo/interops/win/dll
  • Setuptools now bundles DLLs / libraries from interops/{platform}/dll/*.dll
• Implements models for Display, DisplayModule and Resolution info
• Implements fetching various display information through enumerating monitor PROCs and fetching their respective EDIDs
• __init__.py now loads bundled DLLs for Windows platform properly - example of loading currently included DLL module (this DLL path can be bundled, exact partitioning / structure TBD):

with resources.path("pysysinfo.interops.win.dll", "gpu_helper.dll") as dll_path:
   gpu_helper = ctypes.CDLL(str(dll_path))

gpu_helper.GetGPUForDisplay.argtypes = [ctypes.c_char_p, ctypes.c_char_p, ctypes.c_int]
gpu_helper.GetGPUForDisplay.restype = ctypes.c_uint32

...

• display.py may need some cleaning up, it's very messy at the moment, so that's why this PR is marked as draft for now
  • Also will attempt to implement support for detecting size in inches, aspect ratio & mode (landscape/portrait) COMPLETED
• Implements Audio Detection support for Windows platform
  • Groups audio devices (microphone, speakers, AMP/DAC, USB device, etc.) by their parent audio controllers (Realtek, NVidia, etc.)
• Unit tests for Display; improvements to Network UTs

Performance

Tested the performance of fetch_display_info(), on average it takes about ~11ms to fetch display data for 2 displays.
The below logs are pre-network implementation and ACPI/PCI path fetching improvements.

Time for additional details: 0.0002758502960205078
CPU: 0.50830078125 ms
Memory: 6.171630859375 ms
Storage: 39.002685546875 ms
Display: 20.0087890625 ms
Network: 25.7490234375 ms
Audio: 55.676025390625 ms
Graphics: 125.243408203125 ms
Total: 272.35986328125 ms

Reference `main.py` timing code

start_times = [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
# CPU, Memory, Storage, Display, Network, Audio, Graphics, Total
end_times = [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

start_times[0] = start_times[-1] = time.time() * 1000

hm.fetch_cpu_info()
end_times[0] = start_times[1] = time.time() * 1000

hm.fetch_memory_info()
end_times[1] = start_times[2] = time.time() * 1000

hm.fetch_storage_info()
end_times[2] = start_times[3] = time.time() * 1000

hm.fetch_display_info()
end_times[3] = start_times[4] = time.time() * 1000

hm.fetch_network_info()
end_times[4] = start_times[5] = time.time() * 1000

hm.fetch_audio_info()
end_times[5] = start_times[6] = time.time() * 1000

hm.fetch_graphics_info()
end_times[-2] = end_times[-1] = time.time() * 1000


print("CPU:", end_times[0] - start_times[0], "ms")
print("Memory:", end_times[1] - start_times[1], "ms")
print("Storage:", end_times[2] - start_times[2], "ms")
print("Display:", end_times[3] - start_times[3], "ms")
print("Network:", end_times[4] - start_times[4], "ms")
print("Audio:", end_times[5] - start_times[5], "ms")
print("Graphics:", end_times[-2] - start_times[-2], "ms")
print("Total:", end_times[-1] - start_times[-1], "ms")

[kernel-dev]

Below you can see that display UTs have full coverage over the module

[kernel-dev]

Total time of fetching all (currently implemented) hardware implementation post-optimizations averages anywhere between 200-250ms. Executed using pyinstrument