[analytics] Made suppression fixes

[rahul-tuli]

This PR makes the requested suppression fixes, the work is two fold:

• Disable all logs from request and urllib3 loggers: Added a context manager for that
• Disable all stderr from is_gdpr_country: Accomplished using contextlib.redirect_stderr()

After this change me and @KSGulin verified locally that output was still produced by utilities such as deepsparse.check_hardware:

deepsparse.check_hardware 
/home/rahul/.venvs/sparsezoo/lib/python3.10/site-packages/requests/__init__.py:102: RequestsDependencyWarning: urllib3 (1.26.16) or chardet (5.1.0)/charset_normalizer (2.0.12) doesn't match a supported version!
  warnings.warn("urllib3 ({}) or chardet ({})/charset_normalizer ({}) doesn't match a supported "
GenuineIntel CPU detected with 10 cores. (1 sockets with 10 cores each)
DeepSparse FP32 model performance supported: True.
DeepSparse INT8 (quantized) model performance supported: TRUE (emulated).

Non VNNI system detected. Performance speedups for INT8 (quantized) models is available, but will be slower compared with a VNNI system. Set NM_FAST_VNNI_EMULATION=True in the environment to enable faster emulated inference which may have a minor effect on accuracy.

Additional CPU info: {'L1_data_cache_size': 32768, 'L1_instruction_cache_size': 32768, 'L2_cache_size': 1048576, 'L3_cache_size': 14417920, 'architecture': 'x86_64', 'available_cores_per_socket': 10, 'available_num_cores': 10, 'available_num_hw_threads': 20, 'available_num_numa': 1, 'available_num_sockets': 1, 'available_sockets': 1, 'available_threads_per_core': 2, 'bf16': False, 'cores_per_socket': 10, 'dotprod': False, 'i8mm': False, 'isa': 'avx512', 'num_cores': 10, 'num_hw_threads': 20, 'num_numa': 1, 'num_sockets': 1, 'threads_per_core': 2, 'vbmi': False, 'vbmi2': False, 'vendor': 'GenuineIntel', 'vendor_id': 'Intel', 'vendor_model': 'Intel(R) Core(TM) i9-7900X CPU @ 3.30GHz', 'vnni': False, 'zen1': False}

And also that integration tests for deepsparse are green ✅

(sparsezoo) 🥃 deepsparse (main) 💍 make test_integrations 
Running package integrations tests
=========================================================================== test session starts ===========================================================================
platform linux -- Python 3.10.11, pytest-7.3.1, pluggy-1.0.0
rootdir: /home/rahul/github_projects/deepsparse
configfile: pyproject.toml
plugins: flaky-3.7.0, anyio-3.7.0
collected 2 items                                                                                                                                                         

integrations/haystack/tests/test_smoke.py ..                                                                                                                        [100%]

Noting: It will take time for local-install-checks here in deepsparse to be green as they install sparsezoo nightly directly from pypi

Also ran the following to verify:

from multiprocessing import Pool
import subprocess


def task(*args, **kwargs):
    return subprocess.check_output(["deepsparse.check_hardware"])

with Pool(100) as p:
    p.map(task, list(range(1000)))

out = task()
print(out)

Output:

b"GenuineIntel CPU detected with 10 cores. (1 sockets with 10 cores each)\nDeepSparse FP32 model performance supported: True.\nDeepSparse INT8 (quantized) model performance supported: TRUE (emulated).\n\nNon VNNI system detected. Performance speedups for INT8 (quantized) models is available, but will be slower compared with a VNNI system. Set NM_FAST_VNNI_EMULATION=True in the environment to enable faster emulated inference which may have a minor effect on accuracy.\n\nAdditional CPU info: {'L1_data_cache_size': 32768, 'L1_instruction_cache_size': 32768, 'L2_cache_size': 1048576, 'L3_cache_size': 14417920, 'architecture': 'x86_64', 'available_cores_per_socket': 10, 'available_num_cores': 10, 'available_num_hw_threads': 20, 'available_num_numa': 1, 'available_num_sockets': 1, 'available_sockets': 1, 'available_threads_per_core': 2, 'bf16': False, 'cores_per_socket': 10, 'dotprod': False, 'i8mm': False, 'isa': 'avx512', 'num_cores': 10, 'num_hw_threads': 20, 'num_numa': 1, 'num_sockets': 1, 'threads_per_core': 2, 'vbmi': False, 'vbmi2': False, 'vendor': 'GenuineIntel', 'vendor_id': 'Intel', 'vendor_model': 'Intel(R) Core(TM) i9-7900X CPU @ 3.30GHz', 'vnni': False, 'zen1': False}\n"