[BE] fix typos in tools/ (#156082)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/156082
Approved by: https://github.com/soulitzer
ghstack dependencies: #156079

.lintrunner.toml

@@ -1164,7 +1164,6 @@ exclude_patterns = [
     'functorch/**',
     'scripts/**',
     'test/**',
-    'tools/**',
     'torch/**',
     'torchgen/**',
 ]

tools/autograd/derivatives.yaml

@@ -113,7 +113,7 @@
 #   - `wrap_opt_if`, is a 2-argument function that accepts a tensor
 #     variable and a boolean condition that dictates whether to save that
 #     variable in a graph. The result of this function is `std::optional<Tensor>`,
-#     and it is `::std::nullopt` when the condition evalutes to `false`,
+#     and it is `::std::nullopt` when the condition evaluates to `false`,
 #     otherwise it is the variable wrapped in `std::optional<Tensor>`.
 #     For example, wrap_opt_if(var_0, grad_input_mask[1] || grad_input_mask[2])
 #     would mean that `var_0` is saved as long as the second (grad_input_mask[1])
@@ -200,7 +200,7 @@
 #     Undefined Tensors are created with the default constructor `at::Tensor()`.
 #     It is an efficient way to represent a Tensor filled with zeros because
 #     the Tensor holds no sizing information and no Storage data is allocated.
-#     But consequentially, Tensor operations cannot be performed on them.
+#     But consequently, Tensor operations cannot be performed on them.
 #     Therefore, your backward function should treat an undefined output grad as
 #     a zero, and it needs to be a special case.
 #

tools/autograd/gen_inplace_or_view_type.py

@@ -2,7 +2,7 @@
 #
 # NOTE: If any changes are being made to the ADInplaceOrView codegen please also check
 # if updates are needed in torch/csrc/autograd/autograd_not_implemented_fallback.cpp
-# The fallback is expected to mimick this codegen, so we should keep the two in sync.
+# The fallback is expected to mimic this codegen, so we should keep the two in sync.
 
 from __future__ import annotations
 

tools/autograd/gen_python_functions.py

@@ -148,7 +148,7 @@ _SKIP_PYTHON_BINDINGS = [
     "mH",  # these need to be an attributes in Python, not functions
     "nonzero(_(out|numpy))?",
     "set_data",
-    ".*_overrideable",  # overrideable functions for backend extension
+    ".*_overrideable",  # overridable functions for backend extension
     "data",
     "is_leaf",
     "output_nr",
@@ -617,7 +617,7 @@ def load_deprecated_signatures(
         schema_args_by_name = {a.name: a for a in schema.arguments.flat_all}
         for name in call_args:
             assert name in schema_args_by_name or name in known_constants, (
-                f"deprecation definiton: Unrecognized value {name}"
+                f"deprecation definition: Unrecognized value {name}"
             )
 
         # Map deprecated signature arguments to their aten signature and test

tools/autograd/gen_trace_type.py

@@ -17,7 +17,7 @@ if TYPE_CHECKING:
 
 # Note [Manual Backend kernels]
 # For these ops, we want to manually register to dispatch key Backend and
-# skip codegen-ed registeration to all keys before Backend.
+# skip codegen-ed registration to all keys before Backend.
 # For codegen this means:
 #   - op set below must match ops with manual_kernel_registration=True in native_functions.yaml
 #     where we skip codegen backend kernels

tools/autograd/load_derivatives.py

@@ -336,7 +336,7 @@ def postprocess_forward_derivatives(
             # This transformation is based on the observation that for element-wise functions, the Jacobian
             # matrix is diagonal and thus doing J * v is the same as (v^T J)^T (in practice, we ignore the transpositions)
             # For the complex case, we use hermitian transpose and get (v.conj() J).conj()
-            # So here we are going to re-use the backward formula and replace two things:
+            # So here we are going to reuse the backward formula and replace two things:
             # 1) all occurrences of "grad" with "foo_t.conj()", where foo is the name of the unique differentiable input.
             # 2) all usage of an original input "foo" with its primal value "foo_p".
             # 3) conjugate the final result

tools/bazel_tools/shellwrap.sh

@@ -54,5 +54,5 @@ echo "Entering interactive shell at the execution root:"
 # quote escape all the arguments to use as a single input string
 cmd="'$shell' --noprofile --rcfile '$rcfile'"
 
-# run the command in a script psuedo terminal and dump to null
+# run the command in a script pseudo terminal and dump to null
 /usr/bin/script -c "$cmd" -q /dev/null

tools/build_defs/buck_helpers.bzl

@@ -24,7 +24,7 @@ ONLY_AVAILABLE_IN_BUCK2 = [
 def filter_attributes(kwgs):
     keys = list(kwgs.keys())
 
-    # drop unncessary attributes
+    # drop unnecessary attributes
     for key in keys:
         if key in IGNORED_ATTRIBUTES or key in ONLY_AVAILABLE_IN_BUCK2:
             kwgs.pop(key)

tools/build_defs/type_defs.bzl

@@ -83,7 +83,7 @@ def is_bool(arg):
     """Checks if provided instance is a boolean value.
 
     Args:
-      arg: An instance ot check. type: Any
+      arg: An instance to check. type: Any
 
     Returns:
       True for boolean values, False otherwise. rtype: bool
@@ -96,7 +96,7 @@ def is_number(arg):
     """Checks if provided instance is a number value.
 
     Args:
-      arg: An instance ot check. type: Any
+      arg: An instance to check. type: Any
 
     Returns:
       True for number values, False otherwise. rtype: bool
@@ -109,7 +109,7 @@ def is_struct(arg):
     """Checks if provided instance is a struct value.
 
     Args:
-      arg: An instance ot check. type: Any
+      arg: An instance to check. type: Any
 
     Returns:
       True for struct values, False otherwise. rtype: bool

tools/build_with_debinfo.py

@@ -38,7 +38,7 @@ def get_lib_extension() -> str:
         return "so"
     if sys.platform == "darwin":
         return "dylib"
-    raise RuntimeError(f"Usupported platform {sys.platform}")
+    raise RuntimeError(f"Unsupported platform {sys.platform}")
 
 
 def create_symlinks() -> None:
@@ -78,7 +78,7 @@ def create_build_plan() -> list[tuple[str, str]]:
         if line.startswith(": &&") and line.endswith("&& :"):
             line = line[4:-4]
         line = line.replace("-O2", "-g").replace("-O3", "-g")
-        # Build Metal shaders with debug infomation
+        # Build Metal shaders with debug information
         if "xcrun metal " in line and "-frecord-sources" not in line:
             line += " -frecord-sources -gline-tables-only"
         try:

tools/code_analyzer/gen_operators_yaml.py

@@ -68,13 +68,13 @@ from torchgen.selective_build.selector import merge_kernel_metadata
 #    used by training, and not just the root operators. All Training ops are
 #    also considered for inference, so these are merged into inference ops.
 #
-# 3. Operator Depencency Graph (--dep-graph-yaml-path): A path to the
+# 3. Operator Dependency Graph (--dep-graph-yaml-path): A path to the
 #    operator dependency graph used to determine which operators depend on
 #    which other operators for correct functioning. This is used for
 #    generating the transitive closure of all the operators used by the
 #    model based on the root operators when static selective build is used.
 #    For tracing based selective build, we don't need to perform this
-#    transitive cloure.
+#    transitive closure.
 #
 # 4. Model Metadata (--model-name, --model-versions, --model-assets,
 #    --model-backends): Self-descriptive. These are used to tell this

tools/code_coverage/package/tool/print_report.py

@@ -133,7 +133,7 @@ def print_file_oriented_report(
     coverage_percentage = print_file_summary(
         covered_summary, total_summary, summary_file
     )
-    # print test condition (interested folder / tests that are successsful or failed)
+    # print test condition (interested folder / tests that are successful or failed)
     print_test_condition(
         tests,
         tests_type,
@@ -204,7 +204,7 @@ def html_oriented_report() -> None:
     # use lcov to generate the coverage report
     build_folder = os.path.join(get_pytorch_folder(), "build")
     coverage_info_file = os.path.join(SUMMARY_FOLDER_DIR, "coverage.info")
-    # generage coverage report -- coverage.info in build folder
+    # generate coverage report -- coverage.info in build folder
     subprocess.check_call(
         [
             "lcov",

tools/config/defs.bzl

@@ -27,7 +27,7 @@ def if_rocm(if_true, if_false = []):
 def if_sycl(if_true, if_false = []):
     """Helper for selecting based on the whether SYCL/ComputeCPP is configured."""
 
-    # NOTE: Tensorflow expects some stange behavior (see their if_sycl) if we
+    # NOTE: Tensorflow expects some strange behavior (see their if_sycl) if we
     # actually plan on supporting this at some point.
     return select({
         "//conditions:default": if_false,

tools/dynamo/graph_break_registry.json

@@ -270,7 +270,7 @@
     {
       "Gb_type": "Caught non-Exception value",
       "Context": "str(exc_instance)",
-      "Explanation": "Except expects to recieve an object of Exception type but received {exc_instance}.",
+      "Explanation": "Except expects to receive an object of Exception type but received {exc_instance}.",
       "Hints": [
         "Dynamo has detected that tracing the code will result in an error when running in eager. Please double check that your code doesn't contain a similar error when actually running eager/uncompiled."
       ]
@@ -330,7 +330,7 @@
       "Hints": [
         "Use `torch._assert()` to raise a hard AssertionError when the check fails. ",
         "This error will propagate back the user code ",
-        "that called the compiled function (i.e. Dynamo wil not trace any exception handling).",
+        "that called the compiled function (i.e. Dynamo will not trace any exception handling).",
         "Remove the assert statement.",
         "Move the assert statement outside of any context managers in order to graph break with ",
         "partial graph compilation (if fullgraph=False).",
@@ -2050,7 +2050,7 @@
     {
       "Gb_type": "torch.distributed package is not available!",
       "Context": "",
-      "Explanation": "The PyTorch package doesn't include torch.distributed when builing from source.",
+      "Explanation": "The PyTorch package doesn't include torch.distributed when building from source.",
       "Hints": [
         "Set USE_DISTRIBUTED=1 to enable it when building PyTorch from source."
       ]

tools/flight_recorder/components/builder.py

@@ -300,7 +300,7 @@ def build_collectives(
                     for _ in range(1, num_coalesced_entries):
                         all_entries[i].pop(k)
         else:
-            # Iterate through all the ranks and check if there is a mis-match for the current entry.
+            # Iterate through all the ranks and check if there is a mismatch for the current entry.
             check_current_entry_match(
                 all_entries,
                 _pg_guids,

tools/flight_recorder/components/loader.py

@@ -46,7 +46,7 @@ exp = re.compile(r"([\w\-\_]*?)(\d+)$")
 
 def _determine_prefix(files: list[str]) -> str:
     """If the user doesn't specify a prefix, but does pass a dir full of similarly-prefixed files, we should be able to
-    infer the common prefix most of the time.  But if we can't confidently infer, just fall back to requring the user
+    infer the common prefix most of the time.  But if we can't confidently infer, just fall back to requiring the user
     to specify it
     """
     possible_prefixes: defaultdict[str, set[int]] = defaultdict(set)

tools/flight_recorder/components/utils.py

@@ -312,7 +312,7 @@ def match_coalesced_groups_with_non_p2p(
                 {first_rank},
             )
 
-            # Iterate through all the ranks and check if there is a mis-match for the current entry.
+            # Iterate through all the ranks and check if there is a mismatch for the current entry.
             check_current_entry_match(
                 all_coalesced_entries,
                 _pg_guids,

tools/flight_recorder/fr_trace.py

@@ -14,7 +14,7 @@ Not Yet Implemented
 - TODO- tracebacks aren't implemented
 
 Known Issues
-- Flight Recorder buffer sequence_id information is not sufficient to match collectives and coalseced collectives
+- Flight Recorder buffer sequence_id information is not sufficient to match collectives and coalesced collectives
   unless we have the trace data from the beginning of the program.  To enable confident analysis of trace buffers that
   do not start from zero (and to simplify the script's matching logic) we need to add more information to the recorder.
 - Currently, the script omits checking the 'status' of collectives.  We can look for the first 'non completed'

tools/linter/adapters/_linter.py

@@ -413,7 +413,7 @@ class FileLinter(Generic[PythonFileT], ABC):
         return not results or self.args.fix and all(r.is_edit for r in results)
 
     def _error(self, pf: PythonFileT, result: LintResult) -> None:
-        """Called on files that are unparseable"""
+        """Called on files that are unparsable"""
 
     def _replace(self, pf: PythonFileT) -> tuple[str, list[LintResult]]:
         # Because of recursive replacements, we need to repeat replacing and reparsing

tools/linter/dictionary.txt

@@ -1,5 +1,6 @@
 coo
 fro
+froms
 hsa
 nd
 optins

tools/lldb/deploy_debugger.py

@@ -25,7 +25,7 @@ from pathlib import Path
 stem = Path(name).stem
 with NamedTemporaryFile(prefix=stem, suffix='.so', delete=False) as tf:
     tf.write(r)
-    print("torch_deploy registering debug inforation for ", tf.name)
+    print("torch_deploy registering debug information for ", tf.name)
     cmd1 = f"target modules add {tf.name}"
     # print(cmd1)
     lldb.debugger.HandleCommand(cmd1)

tools/nightly.py

@@ -14,7 +14,7 @@ Or if you would like to check out the nightly commit in detached HEAD mode::
     $ ./tools/nightly.py checkout
     $ source venv/bin/activate  # or `& .\venv\Scripts\Activate.ps1` on Windows
 
-Or if you would like to re-use an existing virtual environment, you can pass in
+Or if you would like to reuse an existing virtual environment, you can pass in
 the prefix argument (--prefix)::
 
     $ ./tools/nightly.py checkout -b my-nightly-branch -p my-env
@@ -686,7 +686,7 @@ def _nightly_version(site_dir: Path) -> str:
 
 @timed("Checking out nightly PyTorch")
 def checkout_nightly_version(branch: str | None, site_dir: Path) -> None:
-    """Get's the nightly version and then checks it out."""
+    """Gets the nightly version and then checks it out."""
     nightly_version = _nightly_version(site_dir)
     if branch is None:
         # Detached mode - explicitly use --detach flag

tools/nvcc_fix_deps.py

@@ -1,4 +1,4 @@
-"""Tool to fix the nvcc's dependecy file output
+"""Tool to fix the nvcc's dependency file output
 
 Usage: python nvcc_fix_deps.py nvcc [nvcc args]...
 

tools/packaging/build_wheel.py

@@ -62,7 +62,7 @@ def venv(interpreter: str) -> Iterator[str]:
 
 
 class Builder:
-    # The python interpeter that we should be using
+    # The python interpreter that we should be using
     interpreter: str
 
     def __init__(self, interpreter: str) -> None:
@@ -124,7 +124,7 @@ def main() -> None:
         with venv(interpreter) as venv_interpreter:
             builder = Builder(venv_interpreter)
             # clean actually requires setuptools so we need to ensure we
-            # install requriements before
+            # install requirements before
             builder.install_requirements()
             builder.clean()
 

tools/stats/check_disabled_tests.py

@@ -173,7 +173,7 @@ def save_results(
     all_tests: dict[str, dict[str, int]],
 ) -> None:
     """
-    Save the result to S3, which then gets put into the HUD backened database
+    Save the result to S3, which then gets put into the HUD backend database
     """
     should_be_enabled_tests = {
         name: stats

tools/stats/monitor.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python3
 """
 A Python script that logging the system-level utilization usage in json format.
-Data collected: CPU, memory, GPU memeory utilzation, and GPU utilization if available.
+Data collected: CPU, memory, GPU memory utilization, and GPU utilization if available.
 
 Usage:
 - To run the script with default data collect time setting, use the following command:

tools/test/test_docstring_linter.py

@@ -54,7 +54,7 @@ class TestDocstringLinter(LinterTestCase):
             grandfather_file = f"{td}/grandfather.json"
             grandfather = f"--grandfather={grandfather_file}"
 
-            # Find some faiures
+            # Find some failures
             run("before.txt", grandfather)
 
             # Rewrite grandfather file

tools/testing/target_determination/heuristics/__init__.py

@@ -33,7 +33,7 @@ if TYPE_CHECKING:
 
 
 # All currently running heuristics.
-# To add a heurstic in trial mode, specify the keywork argument `trial_mode=True`.
+# To add a heurstic in trial mode, specify the keyword argument `trial_mode=True`.
 HEURISTICS: list[HeuristicInterface] = [
     PreviouslyFailedInPR(),
     EditedByPR(),

tools/testing/test_run.py

@@ -285,7 +285,7 @@ class ShardedTest:
         if not isinstance(other, ShardedTest):
             raise NotImplementedError
 
-        # This is how the list was implicity sorted when it was a NamedTuple
+        # This is how the list was implicitly sorted when it was a NamedTuple
         if self.name != other.name:
             return self.name < other.name
         if self.shard != other.shard: