Truthound · All articles · 14/24
Evidence-led Truthound Core guidance on data contracts, check inputs and operational limits.
480 articles
Do not admit a large Float64 remainder from engine agreement alone
Build an independent reference from the integer identity. Numerical-path differences can arise even without NaN inputs.
Check observation presence when a singleton GPU datetime mean is NULL
Use the identity that one valid instant has that same mean. Distinguish operation support from CPU-fallback requirements.
Retain intermediate meaning when a Boolean cast chain produces -24
Check the nonzero-is-true policy beside the final 1/0 codes. An Int8 dtype alone cannot admit the status code.
Do not read truncated Delta timestamp statistics as exact upper bounds
Preserve the precision difference between millisecond statistics and microsecond values. Read duplicate reporting and the follow-up fix together.
Distinguish whole-column format inference failure under strict=False
One invalid row differs from a column with no inferable format. Keep an explicit format beside actual completion status.
Polars decimal-comma CSV: reconcile each scientific token from the actual file
Reconcile the complete numeric token for every output row so previous-row buffer contents are not approved as valid formatting.
Inspect execution when aggregation CSE changes a scalar into a list
Expression reuse can change an aggregate’s role and dtype. Admit the plan and the executed result separately.
Reconcile the Arrow boundary before attributing identifier corruption to Delta
An error discovered after storage need not originate there. Check the original long identifier after a short inline string.
NumPy Unicode strip: character order must not change the stored code
Define chars as a set of codepoints rather than an ordered sequence, then compare with a source-based cleaning reference.
A NaT position from NumPy nanargmin: inspect the original mask before handing off the minimum
Check the selected position together with its original validity so a NaT index is not handed off as the minimum observation.
If the first NaN slice determines NumPy nanquantile dtype: test a reordered handoff
Check final dtype and per-slice quantile references so ordering and missingness do not change approved precision.
N/A after NumPy masked find: hand off result dtype with valid fill metadata
After a dtype change, check whether fill_value is representable under the result contract.
NumPy masked comparison at 0.7: agree on dtype and a strict boundary first
Approve comparison precision for observations and thresholds before checking masked and ordinary paths.
NumPy gradient on integer nanoseconds: preserve the origin and intervals before float conversion
Preserve integer differences before converting large absolute times to float, then approve gradients against the original intervals.
NumPy fixed-width replace: input width must not redefine the search and replacement
Preserve the original search and replacement arguments so input storage width cannot redefine cleaning policy.
An empty weighted bincount: zero sums and a float64 handoff are separate conditions
Even a no-data batch must retain the approved weighted-aggregate dtype and input-length contract.
NumPy einsum optimize: was casting=safe enforced on the final out write?
Verify rejection of a forbidden final cast so optimization cannot change permission for data loss.
Large offsets in NumPy cov: separate positive covariance from an approved error budget
Even positive covariance can lose required accuracy at a large offset.
NumPy corrcoef on constant channels: do not rank a finite 1 as defined correlation
Pearson correlation is undefined for a constant channel.
Rescaling complex sqrt on Windows: a correct real part does not validate the imaginary part
Check both components when restoring scale for large complex inputs.
