[ao][fx] fixing public v private match_utils.py (#88396)

Summary: made _is_match, _find_matches, _MatchResult private also added
__all__ to lower_to_qnnpack.py

Test Plan: python test/test_public_bindings.py

Reviewers:

Subscribers:

Tasks:

Tags:

Differential Revision: [D41015540](https://our.internmc.facebook.com/intern/diff/D41015540)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/88396
Approved by: https://github.com/jcaip

test/quantization/ao_migration/test_quantization_fx.py

@@ -135,10 +135,10 @@ class TestAOMigrationQuantizationFx(AOMigrationTestCase):
 
     def test_function_import_fx_match_utils(self):
         function_list = [
-            'MatchResult',
+            '_MatchResult',
             'MatchAllNode',
-            'is_match',
+            '_is_match',
-            'find_matches'
+            '_find_matches'
         ]
         self._test_function_import('fx.match_utils', function_list)
 

test/quantization/fx/test_quantize_fx.py

@@ -27,7 +27,7 @@ from torch.ao.quantization.quantize_fx import (
 from torch.ao.quantization.fx.quantize_handler import DefaultNodeQuantizeHandler
 
 from torch.ao.quantization.fx.match_utils import (
-    is_match,
+    _is_match,
     MatchAllNode,
 )
 
@@ -711,7 +711,7 @@ class TestQuantizeFx(QuantizationTestCase):
         modules = dict(m.named_modules())
         for n in m.graph.nodes:
             if n.op == 'call_module' and type(modules[n.target]) == nn.ReLU:
-                self.assertTrue(is_match(modules, n, pattern))
+                self.assertTrue(_is_match(modules, n, pattern))
 
     def test_pattern_match_constant(self):
         class M(torch.nn.Module):
@@ -727,7 +727,7 @@ class TestQuantizeFx(QuantizationTestCase):
         modules = dict(m.named_modules())
         for n in m.graph.nodes:
             if n.op == "call_function" and n.target == operator.getitem:
-                self.assertTrue(is_match(modules, n, pattern))
+                self.assertTrue(_is_match(modules, n, pattern))
 
     def test_fused_module_qat_swap(self):
         class Tmp(torch.nn.Module):

torch/ao/ns/_numeric_suite_fx.py

@@ -121,7 +121,7 @@ from .fx.ns_types import (
 )
 from torch.ao.quantization.backend_config.utils import get_fusion_pattern_to_root_node_getter
 from torch.ao.quantization.backend_config import BackendConfig
-from torch.ao.quantization.fx.match_utils import find_matches
+from torch.ao.quantization.fx.match_utils import _find_matches
 from torch.ao.quantization.fx.qconfig_mapping_utils import generate_node_name_to_qconfig
 from torch.ao.quantization.fx.quantize_handler import _get_pattern_to_quantize_handlers
 from torch.ao.quantization.qconfig import QConfigAny
@@ -815,7 +815,7 @@ def prepare_n_shadows_model(
     standalone_module_names: List[str] = []
     standalone_module_classes: List[Type] = []
     custom_module_classes: List[Type] = []
-    matches = find_matches(
+    matches = _find_matches(
         mt.graph, modules, patterns, root_node_getter_mapping,
         standalone_module_names, standalone_module_classes, custom_module_classes)
     subgraphs_dedup: Dict[str, List[Node]] = \

torch/ao/ns/fx/n_shadows_utils.py

@@ -20,7 +20,7 @@ from torch.ao.quantization import QConfigMapping
 from torch.ao.quantization.fx.custom_config import PrepareCustomConfig
 from torch.ao.quantization.qconfig import QConfigAny
 from torch.ao.quantization.utils import getattr_from_fqn
-from torch.ao.quantization.fx.match_utils import MatchResult
+from torch.ao.quantization.fx.match_utils import _MatchResult
 
 import collections
 import copy
@@ -100,7 +100,7 @@ class OutputProp:
         return None
 
 def _get_dedup_subgraphs(
-    matches: Dict[str, MatchResult]
+    matches: Dict[str, _MatchResult]
 ) -> Dict[str, List[Node]]:
     # the original matches variable is unique by node, make it unique by subgraph
     # instead
@@ -110,7 +110,7 @@ def _get_dedup_subgraphs(
     # Dict items are not reversible until Python 3.8, so we hack it
     # to be compatible with previous Python versions
     # TODO(future PR): try reversed(list(matches.items()))
-    matches_items_reversed: List[Tuple[str, MatchResult]] = []
+    matches_items_reversed: List[Tuple[str, _MatchResult]] = []
     for name, cur_match in matches.items():
         matches_items_reversed.insert(0, (name, cur_match))
 
@@ -162,7 +162,7 @@ def _get_dedup_subgraphs(
             assert len(cur_match[1]) == 2
             # either (a, b), or ((a, b), c) or (c, (a, b))
             # cannot make any assumptions on order, not clear what the
-            # find_matches function is doing to populate this
+            # _find_matches function is doing to populate this
             # TODO(future PR): make this code less confusing,  see discussion
             # in https://github.com/pytorch/pytorch/pull/80521/files#r975918836
 

torch/ao/ns/fx/pattern_utils.py

@@ -72,7 +72,7 @@ def get_reversed_fusions() -> List[Tuple[NSFusionType, int]]:
     default_base_op_idx = 0
     for quant_pattern, _quant_handler in all_quant_patterns.items():
         # TODO: this is a temporary hack to flatten the patterns from quantization so
-        # that it works with the ns matcher function, maybe we should use `is_match`
+        # that it works with the ns matcher function, maybe we should use `_is_match`
         # in torch.ao.quantization.fx.match_utils to match the patterns
         if isinstance(quant_pattern, tuple) and len(quant_pattern) == 2 and \
            isinstance(quant_pattern[1], tuple) and len(quant_pattern[1]) == 2:

torch/ao/quantization/fx/fuse.py

@@ -8,7 +8,7 @@ from .graph_module import (
     FusedGraphModule
 )
 from .match_utils import (
-    is_match,
+    _is_match,
     MatchAllNode,
 )
 from .pattern_utils import (
@@ -157,7 +157,7 @@ def _find_matches(
         if node.name not in match_map:
             for pattern, value in patterns.items():
                 matched_node_pattern: List[Node] = []
-                if is_match(modules, node, pattern):
+                if _is_match(modules, node, pattern):
                     apply_match(pattern, node, (node, pattern, value(node)), matched_node_pattern, node_to_subpattern)
                     break
 

torch/ao/quantization/fx/lower_to_qnnpack.py

@@ -3,6 +3,10 @@ from .graph_module import QuantizedGraphModule
 from ..qconfig import QConfigAny
 from typing import Dict, Tuple
 
+__all__ = [
+    "lower_to_qnnpack"
+]
+
 def lower_to_qnnpack(
     model: QuantizedGraphModule,
     qconfig_map: Dict[str, QConfigAny],

torch/ao/quantization/fx/match_utils.py

@@ -21,15 +21,11 @@ from torch.nn.utils.parametrize import type_before_parametrizations
 from typing import Any, Dict, List, Callable, Optional, Tuple, Type, Set, Iterable
 
 
-# TODO: revisit this list. Many helper methods shouldn't be public
+__all__: List[str] = []
-__all__ = [
-    "is_match",
-    "find_matches",
-]
 
 # TODO(future PR): the 1st argument is typed as `List[Node]`, but a better type
 # would be a recursive `List[Union[Node, Tuple[Union[Node, ...]]]]`
-MatchResult = Tuple[Node, List[Node], Optional[Pattern], QuantizeHandler]
+_MatchResult = Tuple[Node, List[Node], Optional[Pattern], QuantizeHandler]
 
 _MatchResultWithQConfig = Tuple[Node, List[Node], Optional[Pattern], QuantizeHandler,
                                 QConfigAny]
@@ -38,7 +34,7 @@ _MatchResultWithQConfig = Tuple[Node, List[Node], Optional[Pattern], QuantizeHan
 # need to put the fusion patterns before single patterns. For example, add_relu should be registered come before relu.
 # decorators are applied in the reverse order we see. Also when we match the nodes in the graph with these patterns,
 # we'll start from the last node of the graph and traverse back.
-def is_match(modules, node, pattern, max_uses=sys.maxsize):
+def _is_match(modules, node, pattern, max_uses=sys.maxsize):
     """ Matches a node in fx against a pattern
     """
     if isinstance(pattern, tuple):
@@ -82,16 +78,16 @@ def is_match(modules, node, pattern, max_uses=sys.maxsize):
     if len(arg_matches) != len(node.args):
         return False
 
-    return all(is_match(modules, node, arg_match, max_uses=1) for node, arg_match in zip(node.args, arg_matches))
+    return all(_is_match(modules, node, arg_match, max_uses=1) for node, arg_match in zip(node.args, arg_matches))
 
-def find_matches(
+def _find_matches(
         graph: Graph,
         modules: Dict[str, torch.nn.Module],
         patterns: Dict[Pattern, QuantizeHandler],
         root_node_getter_mapping: Dict[Pattern, Callable],
         standalone_module_names: List[str] = None,
         standalone_module_classes: List[Type] = None,
-        custom_module_classes: List[Any] = None) -> Dict[str, MatchResult]:
+        custom_module_classes: List[Any] = None) -> Dict[str, _MatchResult]:
     """
     Matches the nodes in the input graph to quantization patterns, and
     outputs the information needed to quantize them in future steps.
@@ -123,7 +119,7 @@ def find_matches(
     if standalone_module_names is None:
         standalone_module_names = []
 
-    match_map: Dict[str, MatchResult] = {}
+    match_map: Dict[str, _MatchResult] = {}
     all_matched : Set[str] = set()
 
     def _recursive_record_node_in_match_map(
@@ -188,7 +184,7 @@ def find_matches(
         if node.name not in match_map and node.name not in all_matched:
             for pattern, quantize_handler_cls in patterns.items():
                 root_node_getter = root_node_getter_mapping.get(pattern, None)
-                if is_match(modules, node, pattern) and node.name not in match_map:
+                if _is_match(modules, node, pattern) and node.name not in match_map:
                     matched_node_pattern: List[Node] = []
                     record_match(
                         pattern,

torch/ao/quantization/fx/prepare.py

@@ -57,7 +57,7 @@ from .pattern_utils import (
 
 from .match_utils import (
     _MatchResultWithQConfig,
-    find_matches,
+    _find_matches,
 )
 
 from ..utils import _parent_name
@@ -1527,7 +1527,7 @@ def prepare(
     standalone_module_classes = list(prepare_custom_config.standalone_module_classes.keys())
 
     custom_module_classes = get_custom_module_class_keys(prepare_custom_config.float_to_observed_mapping)
-    matches_without_qconfig = find_matches(
+    matches_without_qconfig = _find_matches(
         model.graph, modules, pattern_to_quantize_handler, root_node_getter_mapping,
         standalone_module_names, standalone_module_classes, custom_module_classes)
 

torch/quantization/fx/match_utils.py

@@ -7,8 +7,8 @@ appropriate files under `torch/ao/quantization/fx/`, while adding an import stat
 here.
 """
 from torch.ao.quantization.fx.match_utils import (
-    MatchResult,
+    _MatchResult,
     MatchAllNode,
-    is_match,
+    _is_match,
-    find_matches
+    _find_matches
 )