Rework fuzzing tests (#3695)

### Changes

Rework fuzzing tests to check nncf.quantize and nncf.compress_weights
with different parameters

## Reason of change

- Fail on expected error
- Check only quantize api

Author: AlexanderDokuchaev

tests/cross_fw/sdl/fuzz/README.md

@@ -0,0 +1,13 @@
+# Fuzzing tests
+
+1. Install the required dependencies
+
+    ```bash
+    pip install -e ../../../.. -r requirements.txt
+    ```
+
+2. Run the fuzz test
+
+    ```bash
+    python fuzz_target.py -runs=10000
+    ```

tests/cross_fw/sdl/fuzz/fuzz_target.py

@@ -0,0 +1,120 @@
+# Copyright (c) 2025 Intel Corporation
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import contextlib
+import os
+import sys
+
+import atheris
+import numpy as np
+import openvino as ov
+from atheris import FuzzedDataProvider
+
+with atheris.instrument_imports(include=["nncf"]):
+    import nncf
+
+# Disable logging for cleaner fuzzing output
+nncf.set_log_level(40)
+
+# To disable telemetry during fuzzing
+os.environ["NNCF_CI"] = "1"
+
+
+class MockDataset:
+    def __init__(self):
+        self.n = 0
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        if self.n < 300:
+            self.n += 1
+            return np.ones((1, 1, 1, 1), dtype=np.float32)
+        raise StopIteration
+
+
+def check_quantize_api(fdp: FuzzedDataProvider) -> None:
+    model = ov.Model([], [])
+    dataset = nncf.Dataset(MockDataset())
+    r_mode = fdp.PickValueInList(list(nncf.QuantizationMode) + [None])
+    r_preset = fdp.PickValueInList(list(nncf.QuantizationPreset) + [None])
+    r_device = fdp.PickValueInList(list(nncf.TargetDevice))
+    r_subset_size = fdp.ConsumeIntInRange(-10, 500)
+    r_fbc = fdp.ConsumeBool()
+    r_model_type = fdp.PickValueInList(list(nncf.ModelType) + [None])  # Keeping it None for simplicity
+
+    with contextlib.suppress(nncf.ParameterNotSupportedError, nncf.ValidationError):
+        nncf.quantize(
+            model,
+            calibration_dataset=dataset,
+            mode=r_mode,
+            preset=r_preset,
+            target_device=r_device,
+            subset_size=r_subset_size,
+            fast_bias_correction=r_fbc,
+            model_type=r_model_type,
+        )
+
+
+def check_compress_weights_api(fdp: FuzzedDataProvider) -> None:
+    model = ov.Model([], [])
+    dataset = nncf.Dataset(MockDataset())
+
+    r_mode = fdp.PickValueInList(list(nncf.CompressWeightsMode))
+    r_ratio = fdp.ConsumeFloatInRange(-0.1, 1.1) if fdp.ConsumeBool() else None
+    r_group_size = fdp.ConsumeIntInRange(-10, 1000) if fdp.ConsumeBool() else None
+    r_all_layers = fdp.PickValueInList([True, False, None])
+    r_dataset = dataset if fdp.ConsumeBool() else None
+    r_sensitivity_metric = fdp.PickValueInList(list(nncf.SensitivityMetric) + [None])
+    r_subset_size = fdp.ConsumeIntInRange(-10, 400)
+    r_awq = fdp.PickValueInList([True, False, None])
+    r_scale_estimation = fdp.PickValueInList([True, False, None])
+    r_gptq = fdp.PickValueInList([True, False, None])
+    r_lora_correction = fdp.PickValueInList([True, False, None])
+    r_backup_mode = fdp.PickValueInList(list(nncf.BackupMode) + [None])
+    r_compression_format = fdp.PickValueInList(list(nncf.CompressionFormat))
+    with contextlib.suppress(nncf.ParameterNotSupportedError, nncf.ValidationError):
+        nncf.compress_weights(
+            model,
+            mode=r_mode,
+            ratio=r_ratio,
+            group_size=r_group_size,
+            ignored_scope=None,
+            dataset=r_dataset,
+            sensitivity_metric=r_sensitivity_metric,
+            all_layers=r_all_layers,
+            subset_size=r_subset_size,
+            awq=r_awq,
+            scale_estimation=r_scale_estimation,
+            gptq=r_gptq,
+            lora_correction=r_lora_correction,
+            backup_mode=r_backup_mode,
+            compression_format=r_compression_format,
+        )
+
+
+def TestOneInput(data: bytes) -> None:
+    fdp = FuzzedDataProvider(data)
+    algo = fdp.PickValueInList(["ptq", "wc"])
+    if algo == "ptq":
+        check_quantize_api(fdp)
+    elif algo == "wc":
+        check_compress_weights_api(fdp)
+
+
+def main() -> None:
+    atheris.Setup(sys.argv, TestOneInput)
+    atheris.Fuzz()
+
+
+if __name__ == "__main__":
+    main()