Target py310 and modernize codebase with ruff (#23401)

Change `target-version = "py310"` and modernize the code base with ruff.

Author: justinchuby

csharp/tools/MauiModelTester/create_test_data.py

@@ -2,7 +2,6 @@
 import shutil
 import sys
 from pathlib import Path
-from typing import Dict, List, Optional
 
 import numpy as np
 
@@ -84,7 +83,7 @@ def parse_args():
     return args
 
 
-def create_existing_data_map(pb_files: List[Path]):
+def create_existing_data_map(pb_files: list[Path]):
     import onnx_test_data_utils as data_utils
 
     data_map = {}
@@ -98,9 +97,9 @@ def create_existing_data_map(pb_files: List[Path]):
 
 def add_model_and_test_data_to_app(
     model_path: Path,
-    symbolic_dims: Optional[Dict[str, int]] = None,
-    input_map: Optional[Dict[str, np.ndarray]] = None,
-    output_map: Optional[Dict[str, np.ndarray]] = None,
+    symbolic_dims: dict[str, int] | None = None,
+    input_map: dict[str, np.ndarray] | None = None,
+    output_map: dict[str, np.ndarray] | None = None,
 ):
     import ort_test_dir_utils as utils
 

onnxruntime/python/backend/backend_rep.py

@@ -6,8 +6,6 @@
 Implements ONNX's backend API.
 """
 
-from typing import Any, Tuple  # noqa: F401
-
 from onnx.backend.base import BackendRep
 
 from onnxruntime import RunOptions

onnxruntime/python/onnxruntime_inference_collection.py

@@ -9,7 +9,8 @@
 import os
 import typing
 import warnings
-from typing import Any, Sequence
+from collections.abc import Sequence
+from typing import Any
 
 from onnxruntime.capi import _pybind_state as C
 
@@ -143,7 +144,7 @@ def set_provider_options(name, options):
         if not all([isinstance(options_for_provider, dict) for options_for_provider in provider_options]):
             raise ValueError("'provider_options' values must be dicts.")
 
-        for name, options in zip(providers, provider_options):
+        for name, options in zip(providers, provider_options, strict=False):
             set_provider_options(name, options)
 
     else:

onnxruntime/python/tools/custom_op_wrapper/create_custom_op_wrapper.py

@@ -22,7 +22,6 @@
 import os
 import sys
 from dataclasses import dataclass
-from typing import List, Optional, Union
 
 import onnx
 from onnx import TensorProto, helper
@@ -65,7 +64,7 @@ class IOInfo:
     index: int
     name: str
     elem_type: TensorProto.DataType
-    shape: Optional[List[Union[int, str]]]
+    shape: list[int | str] | None
 
 
 def str_is_int(string: str) -> bool:
@@ -76,7 +75,7 @@ def str_is_int(string: str) -> bool:
         return False
 
 
-def parse_shape(shape_str: str) -> Optional[List[Union[int, str]]]:
+def parse_shape(shape_str: str) -> list[int | str] | None:
     try:
         shape = [int(s) if str_is_int(s) else s for s in shape_str.split(",")]
     except ValueError:
@@ -204,7 +203,7 @@ def parse_arguments() -> argparse.Namespace:
     return parser.parse_args()
 
 
-def get_attributes(attr_data_info: List[List[str]]):
+def get_attributes(attr_data_info: list[list[str]]):
     if not attr_data_info:
         return {}
 

onnxruntime/python/tools/kernel_explorer/kernels/kernel_explorer.py

@@ -13,11 +13,11 @@
 import sys
 from abc import abstractmethod
 from argparse import Action, ArgumentParser
+from collections.abc import Callable
 from contextlib import contextmanager
 from dataclasses import dataclass
 from fnmatch import fnmatch
 from functools import wraps
-from typing import Callable
 
 build_dir = os.environ.get("KERNEL_EXPLORER_BUILD_DIR", None)
 if build_dir is None:

onnxruntime/python/tools/offline_tuning.py

@@ -7,11 +7,11 @@
 import sys
 from collections import OrderedDict
 from pprint import pprint
-from typing import Any, Dict, List
+from typing import Any
 
 import onnx
 
-TuningResults = Dict[str, Any]
+TuningResults = dict[str, Any]
 
 _TUNING_RESULTS_KEY = "tuning_results"
 
@@ -32,7 +32,7 @@ def extract(model: onnx.ModelProto):
     return json.loads(tuning_results_prop.value)
 
 
-def embed(model: onnx.ModelProto, tuning_results: List[TuningResults], overwrite=False):
+def embed(model: onnx.ModelProto, tuning_results: list[TuningResults], overwrite=False):
     idx = _find_tuning_results_in_props(model.metadata_props)
     assert overwrite or idx <= 0, "the supplied onnx file already have tuning results embedded!"
 
@@ -47,7 +47,7 @@ def embed(model: onnx.ModelProto, tuning_results: List[TuningResults], overwrite
 
 class Merger:
     class EpAndValidators:
-        def __init__(self, ep: str, validators: Dict[str, str]):
+        def __init__(self, ep: str, validators: dict[str, str]):
             self.ep = ep
             self.validators = copy.deepcopy(validators)
             self.key = (ep, tuple(sorted(validators.items())))
@@ -61,7 +61,7 @@ def __eq__(self, other):
     def __init__(self):
         self.ev_to_results = OrderedDict()
 
-    def merge(self, tuning_results: List[TuningResults]):
+    def merge(self, tuning_results: list[TuningResults]):
         for trs in tuning_results:
             self._merge_one(trs)
 

onnxruntime/python/tools/quantization/base_quantizer.py

@@ -4,7 +4,7 @@
 # license information.
 # --------------------------------------------------------------------------
 import logging
-from typing import Any, Dict
+from typing import Any
 
 import numpy as np
 import onnx
@@ -36,7 +36,7 @@
 
 
 class QuantizationParams:
-    def __init__(self, **data: Dict[str, Any]):
+    def __init__(self, **data: dict[str, Any]):
         self.data = {}
         for k, v in data.items():
             if not isinstance(k, str):

onnxruntime/python/tools/quantization/calibrate.py

@@ -9,9 +9,9 @@
 import itertools
 import os
 import uuid
+from collections.abc import Sequence
 from enum import Enum
 from pathlib import Path
-from typing import Dict, Optional, Sequence, Tuple, Union
 
 import numpy as np
 import onnx
@@ -39,7 +39,7 @@ def rel_entr(pk: np.ndarray, qk: np.ndarray) -> np.ndarray:
 def entropy(
     pk: np.ndarray,
     qk: np.ndarray,
-    base: Optional[float] = None,
+    base: float | None = None,
     axis: int = 0,
 ) -> np.ndarray:
     """
@@ -100,7 +100,7 @@ def to_dict(self):
 
 
 class TensorsData:
-    def __init__(self, calibration_method, data: Dict[str, Union[TensorData, Tuple]]):
+    def __init__(self, calibration_method, data: dict[str, TensorData | tuple]):
         self.calibration_method = calibration_method
         self.data = {}
         for k, v in data.items():
@@ -187,8 +187,8 @@ def set_range(self, start_index: int, end_index: int):
 class CalibraterBase:
     def __init__(
         self,
-        model_path: Union[str, Path],
-        op_types_to_calibrate: Optional[Sequence[str]] = None,
+        model_path: str | Path,
+        op_types_to_calibrate: Sequence[str] | None = None,
         augmented_model_path="augmented_model.onnx",
         symmetric=False,
         use_external_data_format=False,
@@ -297,8 +297,8 @@ def compute_data(self) -> TensorsData:
 class MinMaxCalibrater(CalibraterBase):
     def __init__(
         self,
-        model_path: Union[str, Path],
-        op_types_to_calibrate: Optional[Sequence[str]] = None,
+        model_path: str | Path,
+        op_types_to_calibrate: Sequence[str] | None = None,
         augmented_model_path="augmented_model.onnx",
         symmetric=False,
         use_external_data_format=False,
@@ -476,7 +476,8 @@ def compute_data(self) -> TensorsData:
 
         output_names = [self.infer_session.get_outputs()[i].name for i in range(len(self.intermediate_outputs[0]))]
         output_dicts_list = [
-            dict(zip(output_names, intermediate_output)) for intermediate_output in self.intermediate_outputs
+            dict(zip(output_names, intermediate_output, strict=False))
+            for intermediate_output in self.intermediate_outputs
         ]
 
         merged_output_dict = {}
@@ -507,7 +508,9 @@ def compute_data(self) -> TensorsData:
             else:
                 pairs.append(tuple([min_value_array, max_value_array]))
 
-        new_calibrate_tensors_range = TensorsData(CalibrationMethod.MinMax, dict(zip(calibrate_tensor_names, pairs)))
+        new_calibrate_tensors_range = TensorsData(
+            CalibrationMethod.MinMax, dict(zip(calibrate_tensor_names, pairs, strict=False))
+        )
         if self.calibrate_tensors_range:
             self.calibrate_tensors_range = self.merge_range(self.calibrate_tensors_range, new_calibrate_tensors_range)
         else:
@@ -519,8 +522,8 @@ def compute_data(self) -> TensorsData:
 class HistogramCalibrater(CalibraterBase):
     def __init__(
         self,
-        model_path: Union[str, Path],
-        op_types_to_calibrate: Optional[Sequence[str]] = None,
+        model_path: str | Path,
+        op_types_to_calibrate: Sequence[str] | None = None,
         augmented_model_path="augmented_model.onnx",
         use_external_data_format=False,
         method="percentile",
@@ -608,7 +611,8 @@ def collect_data(self, data_reader: CalibrationDataReader):
             raise ValueError("No data is collected.")
 
         output_dicts_list = [
-            dict(zip(output_names, intermediate_output)) for intermediate_output in self.intermediate_outputs
+            dict(zip(output_names, intermediate_output, strict=False))
+            for intermediate_output in self.intermediate_outputs
         ]
 
         merged_dict = {}
@@ -653,8 +657,8 @@ def compute_data(self) -> TensorsData:
 class EntropyCalibrater(HistogramCalibrater):
     def __init__(
         self,
-        model_path: Union[str, Path],
-        op_types_to_calibrate: Optional[Sequence[str]] = None,
+        model_path: str | Path,
+        op_types_to_calibrate: Sequence[str] | None = None,
         augmented_model_path="augmented_model.onnx",
         use_external_data_format=False,
         method="entropy",
@@ -687,8 +691,8 @@ def __init__(
 class PercentileCalibrater(HistogramCalibrater):
     def __init__(
         self,
-        model_path: Union[str, Path],
-        op_types_to_calibrate: Optional[Sequence[str]] = None,
+        model_path: str | Path,
+        op_types_to_calibrate: Sequence[str] | None = None,
         augmented_model_path="augmented_model.onnx",
         use_external_data_format=False,
         method="percentile",
@@ -721,8 +725,8 @@ def __init__(
 class DistributionCalibrater(HistogramCalibrater):
     def __init__(
         self,
-        model_path: Union[str, Path],
-        op_types_to_calibrate: Optional[Sequence[str]] = None,
+        model_path: str | Path,
+        op_types_to_calibrate: Sequence[str] | None = None,
         augmented_model_path="augmented_model.onnx",
         use_external_data_format=False,
         method="distribution",
@@ -1168,8 +1172,8 @@ def get_entropy_threshold(self, histogram, num_quantized_bins):
 
 
 def create_calibrator(
-    model: Union[str, Path],
-    op_types_to_calibrate: Optional[Sequence[str]] = None,
+    model: str | Path,
+    op_types_to_calibrate: Sequence[str] | None = None,
     augmented_model_path="augmented_model.onnx",
     calibrate_method=CalibrationMethod.MinMax,
     use_external_data_format=False,

onnxruntime/python/tools/quantization/matmul_bnb4_quantizer.py

@@ -7,7 +7,6 @@
 import argparse
 import logging
 import os
-from typing import List, Tuple
 
 import numpy as np
 import numpy.typing as npt
@@ -44,7 +43,7 @@ def __init__(self, model: ModelProto, quant_type: int, block_size: int, nodes_to
         self.nodes_to_exclude = set(nodes_to_exclude)
 
     @staticmethod
-    def __get_initializer(name, graph_path: List[GraphProto]) -> Tuple[TensorProto, GraphProto]:
+    def __get_initializer(name, graph_path: list[GraphProto]) -> tuple[TensorProto, GraphProto]:
         for gid in range(len(graph_path) - 1, -1, -1):
             graph = graph_path[gid]
             for tensor in graph.initializer:
@@ -74,7 +73,7 @@ def bnb4_block_quant(self, fpweight: npt.ArrayLike) -> np.ndarray:
 
         return (packed, absmax)
 
-    def _bnb4_matmul_node_weight(self, node: NodeProto, graph_stack: List[GraphProto]) -> NodeProto:
+    def _bnb4_matmul_node_weight(self, node: NodeProto, graph_stack: list[GraphProto]) -> NodeProto:
         """If the node is MatMul with fp32 const weight, quantize the weight with int4, and return the new node"""
 
         if node.op_type != "MatMul":
@@ -129,7 +128,7 @@ def _bnb4_matmul_node_weight(self, node: NodeProto, graph_stack: List[GraphProto
 
         return matmul_bnb4_node
 
-    def _process_subgraph(self, graph_stack: List[GraphProto]):
+    def _process_subgraph(self, graph_stack: list[GraphProto]):
         new_nodes = []
         graph = graph_stack[-1]