feat: Initial bulk of type checking + setup by Gobot1234 · Pull Request #4761 · ansys/pyfluent

Gobot1234 · 2025-12-18T01:55:54Z

Context

Duplicate of #4500 but on the ansys remote. Looking at this code more, I realise I've ran pyupgrade on this so don't merge this till I get the rest of the ruff stuff set up. I'm also realising there are 3 issues tied up in this, apologies for that.
Fixes #4489 and fixes #4490
Helps with #4543

Change Summary

Added a bunch of initial types for the public facing library part of #4738

Gobot1234 · 2025-12-18T01:57:02Z

-        }
-        self.argvals, self.new_session = _get_argvals_and_session(argvals)
-        if self.argvals["start_timeout"] is None:
+        self.argvals, self.new_session = _get_argvals_and_session(


This was changed as sadly there's no way to type transform a typedict to have optional keys (python isn't typescript, sad)

Gobot1234 · 2025-12-18T02:03:02Z

I'll try and get to the conflicts soonish, probably more useful to get the ruff changes in first

…s/pyfluent into jhilton/typing-improvements

…ements

Copilot

Pull request overview

Copilot reviewed 94 out of 96 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-23T16:08:38Z


-    def __call__(self):
-
+    def __call__(self) -> "Meshing | PureMeshing | Solver | SolverIcing | SolverAero":


DockerLauncher.__call__ is annotated to return only session types, but the dry_run branch returns config_dict (a dict). This makes the public typing incorrect (and inconsistent with launch_fluent() docs). Update the return annotation (and related overloads) to include the dict type, or change dry_run behavior to return a session/tuple consistently.

Suggested change

def __call__(self) -> "Meshing | PureMeshing | Solver | SolverIcing | SolverAero":

def __call__(self) -> "Meshing | PureMeshing | Solver | SolverIcing | SolverAero | dict[str, Any]":

Copilot · 2026-02-23T16:08:39Z

+    *,
+    dry_run: Literal[True],
+    **kwargs: Unpack[LaunchFluentArgs],
+) -> tuple[str, str]: ...


The launch_fluent(..., dry_run=True, ...) overload always claims a tuple[str, str] return, but container dry-runs return a dict (from DockerLauncher.__call__). Add overload(s) that distinguish standalone vs container dry-run returns, and include the container dict return type in the function return union.

Suggested change

) -> tuple[str, str]: ...

) -> tuple[str, str] | dict[str, Any]: ...

Copilot · 2026-02-23T16:08:39Z

+            assert_never(fluent_launch_mode)
+
+    return cast(
+        Meshing | PureMeshing | Solver | SolverIcing | SolverAero | tuple[str, str],


Gobot1234 · 2026-02-27T18:55:46Z

Note to self dimension should be removed from SolverAero, attempting to set anything other than 3 doesn't work

…ements

Copilot

Pull request overview

Copilot reviewed 94 out of 96 changed files in this pull request and generated no new comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 94 out of 96 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (2)

src/ansys/fluent/core/launcher/container_launcher.py:285

DockerLauncher.__call__ is annotated to return only session types, but when dry_run is true it returns config_dict (a dict). This makes the method signature inaccurate and undermines the typing work in this PR. Update the return annotation to include the dry-run return type (and ideally tighten it to dict[str, Any]), or change the dry-run behavior to match the annotated return type.

    def __call__(self) -> "Meshing | PureMeshing | Solver | SolverIcing | SolverAero":
        if self.argvals["dry_run"]:
            config_dict, *_ = configure_container_dict(
                self._args,
                compose_config=self._compose_config,
                **self.argvals["container_dict"],
            )
            dict_str = dict_to_str(config_dict)
            print("\nDocker container run configuration:\n")
            print("config_dict = ")
            print(dict_str)
            return config_dict

src/ansys/fluent/core/launcher/launcher.py:125

The create_launcher() docstring says it raises ValueError for unknown launch modes, but the implementation raises DisallowedValuesError. Either update the docstring to reflect the actual exception type or change the code to raise ValueError as documented.

    Raises
    ------
    ValueError
        If an unknown Fluent launch mode is passed.
    """
    launchers = {
        LaunchMode.STANDALONE: StandaloneLauncher,
        LaunchMode.CONTAINER: DockerLauncher,
        LaunchMode.PIM: PIMLauncher,
        LaunchMode.SLURM: SlurmLauncher,
    }

    if fluent_launch_mode in launchers:
        return launchers[fluent_launch_mode](**kwargs)
    else:
        raise DisallowedValuesError(
            "launch mode",
            fluent_launch_mode,
            [f"LaunchMode.{m.name}" for m in LaunchMode],
        )

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

        ip, port, password = sys.argv[2:5]
-        allow_remote_host, certificates_folder, insecure_mode = sys.argv[6:9]
+        allow_remote_host, certificates_folder, insecure_mode, inside_container = (
+            sys.argv[6:10]
+        )


+      - name: "Make venv"
+        run: uv venv --clear
+
+      - name: "Install extras"
+        run: uv pip install -e ".[ci_types]"
+


 @functools.cache
 def _get_api_tree_data():
    """Get API tree data."""
    api_tree_data_file_path = _get_api_tree_data_file_path()
    if api_tree_data_file_path.exists():
-        json_file = open(api_tree_data_file_path, "r")
+        json_file = open(api_tree_data_file_path)
        api_tree_data = json.load(json_file)
        return api_tree_data


Copilot

Pull request overview

Copilot reviewed 94 out of 96 changed files in this pull request and generated 4 comments.

Comments suppressed due to low confidence (1)

src/ansys/fluent/core/services/_protocols.py:38

ServiceProtocol currently defines an __init__(channel, metadata) signature (under TYPE_CHECKING), but many concrete service classes inheriting it require additional constructor parameters (for example fluent_error_state). Type checkers typically treat protocol inheritance like normal inheritance for signature compatibility, so this is likely to introduce a large number of incompatible-override errors (and StreamingService also doesn’t match channel/metadata). Consider removing __init__ from the protocol entirely, or widening it to accept *args, **kwargs (or adding overloads) so all service implementations can conform.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-23T14:13:03Z

+    "VectorFieldDataRequest",
+)
+
+_to_field_name_str = naming_strategy().to_string


_to_field_name_str is assigned twice in a row; the first assignment is immediately overwritten. This is dead code and can confuse readers about which naming strategy instance is intended. Consider keeping only one assignment (either create a single _naming_strategy_instance and derive _to_field_name_str from it, or inline the call once).

Suggested change

_to_field_name_str = naming_strategy().to_string

Copilot · 2026-03-23T14:13:04Z

    api_tree_data_file_path = _get_api_tree_data_file_path()
    if api_tree_data_file_path.exists():
-        json_file = open(api_tree_data_file_path, "r")
+        json_file = open(api_tree_data_file_path)
        api_tree_data = json.load(json_file)
        return api_tree_data


_get_api_tree_data() opens the JSON file without a context manager, so the file handle may remain open until GC. Use with open(...) as f: (or explicitly close the file) to avoid leaking file descriptors, especially since this function is cached and could be called in long-lived processes.

Copilot · 2026-03-23T14:13:04Z

+from ansys.fluent.core.module_config import *

 # Logging has to be imported before importing other PyFluent modules
-from ansys.fluent.core.logger import set_console_logging_level  # noqa: F401
+from ansys.fluent.core.logger import *



Using from ... import * in the package ansys.fluent.core.__init__ makes the exported namespace depend entirely on each submodule’s __all__ and also stops submodules themselves (e.g., module_config, logger, services, etc.) from being exposed as attributes unless they’re explicitly imported as modules. This conflicts with the public API expectations in tests/test_public_api.py (it expects many submodules to be present in dir(ansys.fluent.core)) and is likely to cause API regressions. Consider switching back to explicit imports/aliases for public symbols and explicitly importing the intended public submodules (or defining a package-level __all__) instead of wildcard imports.

Copilot · 2026-03-23T14:13:04Z

+        "data_model_cache",
+        "docker",
+        "examples",
+        "exceptions",
+        "field_data_interfaces",
+        "filereader",
+        "fluent_connection",
+        "generated",
+        "get_build_details",
+        "get_build_version",
+        "get_build_version_string",
+        "journaling",
+        "launch_fluent",
+        "launcher",
+        "logger",
+        "module_config",
+        "parametric",
+        "pyfluent_warnings",


This test assumes many submodules (e.g., docker, exceptions, module_config, logger, etc.) are present in dir(ansys.fluent.core), but the updated ansys.fluent.core.__init__ now mainly does wildcard imports of selected symbols and does not import most submodules as module attributes. As written, this will likely report a large set of “missing symbols” even if the submodules still exist as importable packages. Consider either (1) explicitly importing/exposing those submodules in ansys.fluent.core.__init__, or (2) changing this test to validate importability via importlib.import_module('ansys.fluent.core.<name>') rather than relying on dir().

codacy-production · 2026-04-02T13:52:57Z

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

_{TIP This summary will be updated as you push new changes. Give us feedback}

Copilot

Pull request overview

Copilot reviewed 94 out of 96 changed files in this pull request and generated 6 comments.

Comments suppressed due to low confidence (1)

src/ansys/fluent/core/launcher/launcher.py:125

The create_launcher docstring says it raises ValueError for an unknown launch mode, but the implementation raises DisallowedValuesError. Please align the docstring (or change the exception type) so user-facing docs and behavior match.

    Raises
    ------
    ValueError
        If an unknown Fluent launch mode is passed.
    """
    launchers = {
        LaunchMode.STANDALONE: StandaloneLauncher,
        LaunchMode.CONTAINER: DockerLauncher,
        LaunchMode.PIM: PIMLauncher,
        LaunchMode.SLURM: SlurmLauncher,
    }

    if fluent_launch_mode in launchers:
        return launchers[fluent_launch_mode](**kwargs)
    else:
        raise DisallowedValuesError(
            "launch mode",
            fluent_launch_mode,
            [f"LaunchMode.{m.name}" for m in LaunchMode],
        )

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-02T15:37:16Z

+    zones_svar_data = dict[Any, npt.NDArray[Any] | None]()
    for array in solution_variables_data:
        if array.WhichOneof("array") == "payloadInfo":
            zones_svar_data[array.payloadInfo.zone] = _extract_svar(


dict[Any, ...]() attempts to call a parameterized generic alias and will raise TypeError: 'types.GenericAlias' object is not callable at runtime. Initialize with {} (or dict()) and keep the type via a variable annotation instead.

Copilot · 2026-04-02T15:37:16Z

                        return field_arr

-        fields_data = {}
+        fields_data = dict[Any, dict[str, npt.NDArray[Any]]]()
        for chunk in chunk_iterator:


dict[Any, ...]() is a parameterized generic alias and is not callable at runtime (TypeError). Use {}/dict() and annotate the variable type separately.

Copilot · 2026-04-02T15:37:16Z

+def _get_argvals_and_session(
+    argvals: LauncherArgsT,
+) -> tuple[
+    LauncherArgsT, type["Meshing | PureMeshing | Solver | SolverIcing | SolverAero"]
+]:
+    _validate_gpu(argvals.get("gpu"), argvals.get("dimension"))
    argvals["graphics_driver"] = _get_graphics_driver(
-        argvals["graphics_driver"], argvals["ui_mode"]
+        argvals.get("graphics_driver"), argvals.get("ui_mode")
    )
-    argvals["mode"] = FluentMode(argvals["mode"])
-    del argvals["self"]
+    argvals["mode"] = FluentMode(argvals.get("mode"))
    new_session = argvals["mode"].get_fluent_value()


_get_argvals_and_session calls _validate_gpu(argvals.get('gpu'), argvals.get('dimension')), but when launchers are constructed from **kwargs (e.g., SessionBase.from_install()), dimension may be absent/None. Dimension(dimension) will raise ValueError in that case. Consider defaulting/normalizing required values (dimension, precision, ui_mode, etc.) inside _get_argvals_and_session (or before calling it) so callers can omit them safely.

Copilot · 2026-04-02T15:37:16Z

    """Get API tree data."""
    api_tree_data_file_path = _get_api_tree_data_file_path()
    if api_tree_data_file_path.exists():
-        json_file = open(api_tree_data_file_path, "r")
+        json_file = open(api_tree_data_file_path)
        api_tree_data = json.load(json_file)
        return api_tree_data


open(api_tree_data_file_path) is not closed, which can leak file descriptors. Use a context manager (with open(...) as f:) when loading the JSON.

Copilot · 2026-04-02T15:37:17Z

+    "VectorFieldDataRequest",
+)
+
+_to_field_name_str = naming_strategy().to_string


_to_field_name_str is assigned twice in a row; the first assignment is immediately overwritten. Remove the redundant assignment to avoid confusion.

Suggested change

_to_field_name_str = naming_strategy().to_string

Copilot · 2026-04-02T15:37:17Z

+    def _get_nodes(self, surface_id: int) -> tuple[npt.NDArray[np.uint32], np.int16]:
        min_id, max_id = self.get_surface_locs(surface_id)
        nnodes = self._file_handle["meshes"]["1"]["faces"]["nodes"]["1"]["nnodes"]
        nodes = self._file_handle["meshes"]["1"]["faces"]["nodes"]["1"]["nodes"]
        previous = np.sum(nnodes[0:min_id])
        nnodes = nnodes[min_id : max_id + 1]
        nodes = nodes[previous : previous + np.sum(nnodes)]
-        return [nodes, nnodes]
+        return (nodes, nnodes)


_get_nodes returns (nodes, nnodes) where both values are array-like slices, but the return type annotates nnodes as np.int16 (a scalar). This should be an ndarray type (e.g., npt.NDArray[np.integer[Any]] or similar) to match the returned value.

- Add .ci/pylock.toml (PEP 751 format) with ci_types extras pinned - Ensures consistent package versions across local and CI environments - Update basedpyright baseline with locked dependencies - CI workflow now uses 'uv pip sync .ci/pylock.toml' for reproducible type checking

Copilot

Pull request overview

Copilot reviewed 94 out of 97 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (2)

src/ansys/fluent/core/launcher/container_launcher.py:285

DockerLauncher.__call__ is annotated to return only session types, but it returns config_dict (a dict) when dry_run is true. Either include the dict in the return annotation (and propagate that to launch_fluent) or refactor dry-run handling so the annotation matches behavior.

    def __call__(self) -> "Meshing | PureMeshing | Solver | SolverIcing | SolverAero":
        if self.argvals["dry_run"]:
            config_dict, *_ = configure_container_dict(
                self._args,
                compose_config=self._compose_config,
                **self.argvals["container_dict"],
            )
            dict_str = dict_to_str(config_dict)
            print("\nDocker container run configuration:\n")
            print("config_dict = ")
            print(dict_str)
            return config_dict

src/ansys/fluent/core/services/_protocols.py:38

ServiceProtocol's __init__ only allows (channel, metadata), but many service classes in this PR require additional required parameters (for example fluent_error_state). This makes the protocol inaccurate and can trigger incompatible override / structural typing issues. Consider widening __init__ to accept *args, **kwargs (or using a ParamSpec-based protocol / removing __init__ from the protocol) so it matches all service constructors used by FluentConnection.create_grpc_service.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-07T17:28:08Z

    api_tree_data_file_path = _get_api_tree_data_file_path()
    if api_tree_data_file_path.exists():
-        json_file = open(api_tree_data_file_path, "r")
+        json_file = open(api_tree_data_file_path)
        api_tree_data = json.load(json_file)
        return api_tree_data


_get_api_tree_data() opens the JSON file without a context manager, so the file handle is never closed. Use a with open(...) as f: block (or Path.read_text()/json.loads) to avoid leaking descriptors.

Copilot · 2026-04-07T17:28:08Z

+) -> (
+    Meshing
+    | PureMeshing
+    | Solver
+    | SolverIcing


The declared return type for launch_fluent is missing the container dry-run return (dict from DockerLauncher.__call__), so the annotation disagrees with runtime behavior. Update the union return type accordingly (and keep it consistent with the overloads).

Copilot · 2026-04-07T17:28:08Z

+    """Working directory for the Fluent client."""
+    fluent_path: str | None
+    """User provided Fluent installation path."""
+    topy: str | list[Any] | None


Gobot1234 and others added 10 commits November 5, 2025 10:21

Add type level support for passing Path instances to download_file

4f2c990

Add typing_extensions

1ddf8e8

Fix inheritance for Meshing

7825ff4

Fix create_launcher allowing None

90d69a0

Update launchers typing to be more specific

654deb4

Fix ci

ec258c3

Fix ci again

00c3a85

remove accidental commits

b6d2e3e

run pyupgrade for moving typing to builtins and a couple of other things

0a48f90

fix some np.arrays that are meant to be ndarrays in types

7b36a1c

Gobot1234 commented Dec 18, 2025

View reviewed changes

mkundu1 mentioned this pull request Dec 30, 2025

100% type hints in PyFluent code #4738

Open

Gobot1234 added 5 commits January 5, 2026 16:50

pre-fix the conflict that will happen

c04b413

add types for services

08e5525

Merge branch 'jhilton/typing-improvements' of https://github.com/ansy…

7c2e9c3

…s/pyfluent into jhilton/typing-improvements

fix some low hanging fruit

0acf4ea

Merge remote-tracking branch 'origin/main' into jhilton/typing-improv…

e6a7bf9

…ements

github-actions Bot added documentation Documentation related (improving, adding, etc) examples Publishing PyFluent examples maintenance General maintenance of the repo (libraries, cicd, etc) dependencies Related to dependencies labels Jan 15, 2026

Gobot1234 commented Jan 15, 2026

View reviewed changes

Comment thread src/ansys/fluent/core/launcher/slurm_launcher.py Outdated

Gobot1234 marked this pull request as ready for review January 15, 2026 03:19

Gobot1234 requested review from hpohekar, mkundu1, prmukherj and seanpearsonuk as code owners January 15, 2026 03:19

Copilot AI review requested due to automatic review settings January 15, 2026 03:19

Gobot1234 requested a review from mayankansys as a code owner January 15, 2026 03:19

Copilot AI reviewed Feb 23, 2026

View reviewed changes

Merge branch 'main' into jhilton/typing-improvements

d464ad0

Merge remote-tracking branch 'origin/main' into jhilton/typing-improv…

74171c9

…ements

Copilot AI review requested due to automatic review settings March 12, 2026 16:56

Copilot started reviewing on behalf of Gobot1234 March 12, 2026 16:56 View session

Copilot AI reviewed Mar 12, 2026

View reviewed changes

Gobot1234 added 2 commits March 13, 2026 10:31

add tomli

4eb9f20

maybe the reordering fixes it

83c5f72

Copilot AI review requested due to automatic review settings March 17, 2026 13:28

Copilot started reviewing on behalf of Gobot1234 March 17, 2026 13:29 View session

Copilot AI reviewed Mar 17, 2026

View reviewed changes

maybe this fixes it?

317d529

seanpearsonuk approved these changes Mar 17, 2026

View reviewed changes

rebasline?

bfb4e3a

Copilot AI review requested due to automatic review settings March 23, 2026 14:05

Copilot started reviewing on behalf of Gobot1234 March 23, 2026 14:05 View session

Copilot AI reviewed Mar 23, 2026

View reviewed changes

Merge branch 'main' into jhilton/typing-improvements

cbb3038

maybe this fixes some of the errors

5afbed7

Copilot AI review requested due to automatic review settings April 2, 2026 15:30

Copilot started reviewing on behalf of Gobot1234 April 2, 2026 15:31 View session

Copilot AI reviewed Apr 2, 2026

View reviewed changes

Gobot1234 added 2 commits April 7, 2026 11:59

maybe this fixes some of the errors

726a227

Copilot AI review requested due to automatic review settings April 7, 2026 17:20

Copilot started reviewing on behalf of Gobot1234 April 7, 2026 17:21 View session

Copilot AI reviewed Apr 7, 2026

View reviewed changes

forgot to use lock file

4b4cd40


		def __call__(self):

		def __call__(self) -> "Meshing \| PureMeshing \| Solver \| SolverIcing \| SolverAero":

	) -> tuple[str, str]: ...
	) -> tuple[str, str] \| dict[str, Any]: ...

Conversation

Gobot1234 commented Dec 18, 2025

Context

Change Summary

Uh oh!

Gobot1234 Dec 18, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Gobot1234 commented Dec 18, 2025

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

Gobot1234 commented Feb 27, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

codacy-production Bot commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Up to standards ✅

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Apr 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

codacy-production Bot commented Apr 2, 2026 •

edited

Loading