docs: cleanup documentation for function-based rewriting

Author: KarelZe

README.md

@@ -207,10 +207,6 @@ return model_with_rewrite_applied
 
 For a detailed tutorial on how to create target_pattern, replacement_pattern and match_condition blocks in order to utilize the pattern-based rewriter, refer to the tutorial [Pattern-based Rewrite Using Rules](https://onnxscript.ai/tutorial/rewriter/rewrite_patterns.html)
 
-### Function-based rewriting
-
-This style of rewriting matches a `FUNCTION_KEYWORD` and `PACKAGE_NAME` provided by the user to an existing function within the graph and replaces it with a new function provided by the user.
-
 ## Development Guidelines
 
 Every change impacting the converter or the eager evaluation must be

docs/tutorial/rewriter/simple_example.md

@@ -49,8 +49,8 @@ rule = pattern.RewriteRule(
 Now that the rewrite rule has been created, the next step is to apply these pattern-based rewrite rules. The `rewriter.rewrite` call consists of three main components:
 
 1. `model` : The original model on which the pattern rewrite rules are to be applied. This is of type `onnx.ModelProto`.
-2. `function_rewrite_rules` : `(Optional)` This parameter is used to pass rewrite rules based on function names. Steps on how to use this parameter will be covered in a different tutorial. This parameter is of type `Sequence[type[FunctionRewriteRule]]`
-3. `pattern_rewrite_rules` : `(Optional)` This parameter is used to pass rewrite rules based on a provided replacement pattern. For the purpose of this tutorial, we will be using only this parameter in conjunction with `model`. This parameter is of either one of these types:
+
+2. `pattern_rewrite_rules` : `(Optional)` This parameter is used to pass rewrite rules based on a provided replacement pattern. This parameter is of either one of these types:
     - `Sequence[PatternRewriteRule]`
     - `RewriteRuleSet`
 

onnxscript/rewriter/onnxruntime/__init__.py

@@ -5,7 +5,7 @@
 
 from __future__ import annotations
 
-from typing import Any
+from typing import Any, Sequence
 
 import onnx
 
@@ -26,14 +26,13 @@ def rewrite(
     model_proto: onnx.ModelProto,
     /,
     function_rules=None,
-    pattern_rules: list[pattern.RewriteRule] | None = None,
+    pattern_rules: Sequence[pattern.RewriteRule] | None = None,
 ) -> onnx.ModelProto:
     """Rewrite the model using the given rules.
 
     Args:
         model_proto: The model to rewrite.
-        function_rules: The function rewrite rules to apply. If None, the default rules
-            for onnxruntime are used.
+        function_rules: The function rewrite rules to apply. Deprecated.
         pattern_rules: The pattern rewrite rules to apply. If None, the default rules
             for onnxruntime are used.
 

tools/ort_rewriter_profiling/README.md

@@ -127,14 +127,13 @@
 5. Develop optimization code.
     - `onnx-script/onnxscript/optimizer`: Optimizations such as constant folding, inlining, dead code elimination etc.
     - `onnx-script/onnxscript/rewriter`: Pattern based fusions.
-    - `onnx-script/onnxscript/rewriter/onnxruntime`: Onnxruntime specific pattern based fusions.
-    - `onnx-script/onnxscript/rewriter/onnxruntime/transformers`: Onnxruntime specific function based fusions.
+    - `onnx-script/onnxscript/rewriter/ort_fusions`: Onnxruntime specific pattern based fusions.
         - Use function unittest producer tool to create function fusion unittest. Example command to distill 4 unittests for function `LlamaSdpaAttention` from `llama_v2_7b` `dynamo` model. The unittest models are named with prefix `sdpa_llama2`:
             ```
-            # Under onnx-script/onnxscript/rewriter/transformers
-            CUDA_VISIBLE_DEVICES="3" python tools/function_unittest_producer.py --model-path ../../../tools/onnx_models/llama_v2_7b_16h/dynamo_ort_rewritten/llama_v2_7b_16h_dynamo_ort_rewritten.onnx --function LlamaSdpaAttention --output-dir ../../testing/rewriter/transformers/unittest_models/ --max-outputs 4 --name sdpa_llama2
+            # Under onnx-script/onnxscript/rewriter
+            CUDA_VISIBLE_DEVICES="3" python tools/function_unittest_producer.py --model-path ../../../tools/onnx_models/llama_v2_7b_16h/dynamo_ort_rewritten/llama_v2_7b_16h_dynamo_ort_rewritten.onnx --function LlamaSdpaAttention --output-dir ../../testing/rewriter/unittest_models/ --max-outputs 4 --name sdpa_llama2
             ```
-        - Create new testcase under `onnx-script/onnxscript/rewriter/transformers` with the generated unittest models.
+        - Create new testcase under `onnx-script/onnxscript/rewriter/ort_fusions` with the generated unittest models.
             ```python
                 def test_sdpa_llama2(self):
                     common.test_function_rewrite("sdpa_llama2", 4)