misc: General reductions in lines of code by EdCaunt · Pull Request #3031 · devitocodes/devito

EdCaunt · 2026-09-22T19:58:34Z

No description provided.

EdCaunt · 2026-09-22T20:09:18Z

-            if isinstance(p, type) \
-                    and issubclass(p, sympy.Symbol) \
-                    and any(isinstance(i, p) for i in self.free_symbols):
+            if (isinstance(p, type) and issubclass(p, sympy.Symbol) and


Todo: revert

EdCaunt · 2026-09-22T20:11:08Z

-        last = [ecls(lhs, rhs, implicit_dims=implicit_dims)]

-        return temps + last
+        return temps + [ecls(lhs, rhs, implicit_dims=implicit_dims)]


Todo: revert

EdCaunt · 2026-09-22T20:11:47Z


-class LinearInterpolator(WeightedInterpolator):
+class TabulatedInterpolator(WeightedInterpolator):
+    """Shared plumbing for schemes whose weights are tabulated on the host."""


Todo: adjust docstring

codecov · 2026-09-22T20:15:04Z

Codecov Report

❌ Patch coverage is 78.18930% with 53 lines in your changes missing coverage. Please review.
✅ Project coverage is 83.93%. Comparing base (760a476) to head (e2105ac).

Files with missing lines	Patch %	Lines
devito/types/sparse.py	53.84%	18 Missing ⚠️
devito/finite_differences/differentiable.py	75.86%	6 Missing and 1 partial ⚠️
devito/operator/operator.py	76.66%	6 Missing and 1 partial ⚠️
devito/types/grid.py	69.56%	5 Missing and 2 partials ⚠️
devito/types/dense.py	75.00%	4 Missing and 2 partials ⚠️
devito/tools/data_structures.py	71.42%	4 Missing ⚠️
devito/types/dimension.py	90.90%	3 Missing ⚠️
devito/types/basic.py	95.83%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3031      +/-   ##
==========================================
+ Coverage   83.89%   83.93%   +0.03%     
==========================================
  Files         258      258              
  Lines       55605    55450     -155     
  Branches     4765     4744      -21     
==========================================
- Hits        46652    46544     -108     
+ Misses       8135     8093      -42     
+ Partials      818      813       -5

Flag	Coverage Δ
pytest-gpu-aomp-amdgpuX	`68.72% <60.90%> (+0.07%)`	⬆️
pytest-gpu-gcc-	`78.59% <77.77%> (+0.03%)`	⬆️
pytest-gpu-icx-	`78.53% <77.77%> (+0.04%)`	⬆️
pytest-gpu-nvc-nvidiaX	`69.24% <62.13%> (+0.07%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

EdCaunt · 2026-09-22T20:16:52Z

                "due to excessive register pressure in one of the Operator "
-                "kernels. Try supplying a smaller `par-tile` value."
-            )
+                "kernels. Try supplying a smaller `par-tile` value.")


Todo: revert

EdCaunt · 2026-09-22T20:16:58Z

                "prevents the use of `tbc-tile` in realistic scenarios, but it "
-                "will be removed in future versions."
-            )
+                "will be removed in future versions.")


Todo: revert

EdCaunt · 2026-09-22T20:17:33Z

-    'linearize': False,
-    'place-transfers': False
-}
+rcompile_registry = {'avoid_denormals': False, 'linearize': False,


Todo: revert

EdCaunt · 2026-09-22T20:17:46Z

-    byproduct = byproduct0.filter(key)
-
-    return irs, byproduct
+    return irs, byproduct0.filter(key)


Todo: revert

EdCaunt · 2026-09-22T20:18:00Z


-        return {'platform': self.platform.name,
-                'compiler': compiler,
+        return {'platform': self.platform.name, 'compiler': compiler,


Todo: revert

EdCaunt · 2026-09-22T20:18:09Z

        """
-        key0 = lambda f: (f.is_TimeFunction and
-                          f.save is not None and
+        key0 = lambda f: (f.is_TimeFunction and f.save is not None and


Todo: revert

EdCaunt · 2026-09-22T20:18:44Z

-            self.nbytes_consumed_arrays,
-            self.nbytes_consumed_memmapped
-        )
+        mem_locations = (self.nbytes_consumed_functions, self.nbytes_consumed_arrays,


Todo: revert

JDBetteridge

Thanks, I hate it.

But in all seriousness, what's the point? It looks like we reduce the number of lines of code in the whole codebase by ~600 at the expense of:

Readability “Programs must be written for people to read, and only incidentally for machines to execute.” Harold Abelson, Structure and Interpretation of Computer Programs
Automatic merge and rebase functionality
Spaghetti code, I'm thinking specifically of early/multiple returns "A single return statement at the end of a function creates a single, known point which is passed through at the termination of function execution.
The single-return structure is easier to change. If there is more to do after a
search, just add the statement(s) between the for loop and the return." C STYLE GUIDE, 1994, National Aeronautics and Space Administration (NASA)

The metric should not be the number of lines we remove. The goal of reducing the number of lines of code in a codebase should be eliminating redundant classes and functions by refining the abstractions and generally making the structure more elegant.

JDBetteridge · 2026-09-23T01:13:50Z

-    'Real',
-    'Weights',
+    'Conj', 'DiffDerivative', 'Differentiable', 'EvalDerivative', 'Imag',
+    'IndexDerivative', 'IndexDerivativeProperty', 'LocalSum', 'Real', 'Weights',


JDBetteridge · 2026-09-23T01:15:16Z

        fd_args = []
        for f in self._args_diff:
-            try:
+            with suppress(AttributeError):


JDBetteridge · 2026-09-23T01:17:10Z

-                    and issubclass(p, sympy.Symbol) \
-                    and any(isinstance(i, p) for i in self.free_symbols):
+            if (isinstance(p, type) and issubclass(p, sympy.Symbol) and
+                    any(isinstance(i, p) for i in self.free_symbols)):


This is less readable

JDBetteridge · 2026-09-23T01:18:09Z

    prio = getattr(expr, '_fd_priority', 0)
-    return max([prio] + [deep_priority(i)
-                         for i in getattr(expr, '_args_diff', ())])
+    return max([prio] + [deep_priority(i) for i in getattr(expr, '_args_diff', ())])


@FabioLuporini doesn't like getattr, if you're going to touch lines of code you should probably fix them too

JDBetteridge · 2026-09-23T01:19:27Z

-            other = self.func(*other)._eval_at(highest_priority(self))
-            return self.func(other, *derivs)
+        other = self.func(*other)._eval_at(highest_priority(self))
+        return self.func(other, *derivs)


I despise multiple returns and consider them the definition of spaghetti code. I'd much rather you consolidate and return once

JDBetteridge · 2026-09-23T01:45:47Z

        """Number of points in the domain region."""
        return DimensionTuple(*self.shape, getters=self.dimensions)

+    def _make_sizes(self, pairs):


This should be a free function and used in at least one other place to justify its existance

JDBetteridge · 2026-09-23T01:45:59Z

-                    for a, b in zip(row_indices, col_indices, strict=True)
-                ]
+                vec = [mat[a]*other_mat[b]
+                       for a, b in zip(row_indices, col_indices, strict=True)]


JDBetteridge · 2026-09-23T01:46:18Z

-            cls = self.function._indexed_cls
-        except AttributeError:
-            cls = Indexed
+        cls = getattr(self.function, '_indexed_cls', Indexed)


@FabioLuporini he's doing it again...

JDBetteridge · 2026-09-23T01:46:50Z

-                "all dimensions are covered"
-            )
+            raise ValueError(f"Cannot translate {self} with mapper {mapper} since not "
+                             "all dimensions are covered")


No description provided.

JDBetteridge · 2026-09-23T01:47:31Z

+                for i, j in zip(self._decomposition, self._size_inhalo.left, strict=True)]
+        right = [max(i.loc_abs_max+j-i.glb_max, 0) if i and not i.loc_empty else 0
+                 for i, j in zip(self._decomposition, self._size_inhalo.right,
+                                 strict=True)]


Let's make this bit less readable!

I'm stopping here, I can't read anymore

JDBetteridge · 2026-09-23T02:00:34Z

Also, if you have time to feed this into a prompt, you probably have time to help me a bit more with my other PRs

EdCaunt added 9 commits September 22, 2026 15:39

misc: Reduction in size of types/dimension

76c4bd6

Further reduction of types

e03a32f

misc: Reduce types/dense

305c6b5

misc: Reduce operator

194002b

misc: Grid and differentiable reductions

06b49de

misc: Reduce data structures

f75d3d6

misc: Reduce interpolators

422277d

misc: Improve readability

4d844ce

misc: Remove underscores

e2105ac

EdCaunt commented Sep 22, 2026

View reviewed changes

JDBetteridge requested changes Sep 23, 2026

View reviewed changes

Conversation

EdCaunt commented Sep 22, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codecov Bot commented Sep 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JDBetteridge left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JDBetteridge commented Sep 23, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

codecov Bot commented Sep 22, 2026 •

edited

Loading