ponder-lab / Hybridize-Functions-Refactoring / 26464324828
72%
main: 72%

LAST BUILD BRANCH: feat/run-output-fixtures
DEFAULT BRANCH: main
Ran
Jobs 1
Files 31
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 72.219%. Remained the same
26464324828

push

github

web-flow 
Extend precision audit to fixtures 52-54, 130-133 (#554)

* Extend precision audit to fixtures 52-54 and 130-133

Seven fixtures share the multi-function single-arg shape: two functions
in the test file, `func()` (decorated `@tf.function`) constructs tensors
via `c = tf.matmul(a, b)` and passes the result through a low-level
tensor wrapper into `func2(t)`. The function under test is `func2`; the
parameter under audit is `t`.

Across the seven fixtures, only the import shape and the constructor
name vary—the runtime data flow is identical:

- 52-54: `tf.Tensor(c.op, 0, tf.float32)` (different imports)
- 130-133: `tf.experimental.numpy.ndarray(c.op, 0, tf.float32)`
  (different imports)

All seven previously asserted only the structural shape (one parameter
named `t`, `getHasTensorParameter() == true`). Add a new helper
`testHasLikelyTensorParameterHelperMultiFunction(...)` that handles the
multi-function single-arg case—selects the target function by name,
asserts the structural shape, and runs the precision audit—and migrate
all seven fixtures to use it.

Audit verdict for all seven: IDEAL. Ariadne tracks the
`tf.matmul(a, b)` → `c.op` → tensor-wrapper-constructor → `t` chain
correctly, emitting FLOAT32 with shape (2, 2). The inferred input
signature carries the same `TensorType` at the single parameter
position.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* Tighten `testHasLikelyTensorParameterHelperMultiFunction` per Copilot review

Two findings from Copilot on #554:

(1) `expectingTensorParameter` was a parameter, but the body
unconditionally asserted `t.getTensorTypes()` is a singleton and
`inferInputSignature()` is present, so calling with
`expectingTensorParameter == false` would fail. Drop the parameter and
hardcode `assertTrue(target.getHasTensorParameter())`. If a future
fixture needs the no-tensor or signature-drop case, add a sibling
helper analogous to `testHasLikelyTensorParameterHelperExpectingDrop`.

(2) The target-functio... (continued)

1188 of 1645 relevant lines covered (72.22%)

0.72 hits per line

Jobs
ID Job ID Ran Files Coverage
1 26464324828.1 31
72.22
 GitHub Action Run
Source Files on build 26464324828