disclude.ioPyPI / polars-runtime-32 / 2.1.0
polars-runtime-32 2.1.0
- Source
- https://github.com/pola-rs/polars
- Artifact scanned
polars_runtime_32-2.1.0.tar.gz- Compared with
polars_runtime_32-2.0.0.tar.gz(12 signals introduced since)- Files
- 2245 scanned, 70 skipped (binary, oversized, or unrecognised)
- Signals
- 0 critical · 0 warn · 21 info
- Scan time
- 3.2 s
- LLM review
- LLM review by claude-haiku-5-5 (Anthropic): nothing at warn or critical to review.
- Disclude version
- 4.3.0
- Scanned (UTC)
- 2026-10-26 15:42
Signals
Ordered by severity, then signals introduced in this release. Hidden characters are shown as ⟨U+…⟩.
-
// Hoeffding's tail `2 exp(-t² / 2Σw²)` (KLL Lemma 1) at `t = εn`, solved for `Σw²`. let z = f64sqrt; -
// Hoeffding's tail `2 exp(-t² / 2Σw²)` (KLL Lemma 1) at `t = εn`, solved for `Σw²`. let z = f64sqrt; -
// Hoeffding's tail `2 exp(-t² / 2Σw²)` (KLL Lemma 1) at `t = εn`, solved for `Σw²`. let z = f64sqrt; -
// Solve `z * sqrt(Σw²) ≤ error * n` for `k`, with `Σw² = var * (n/k)²`. let k_from_total_variance = ; -
// Solve `z * sqrt(Σw²) ≤ error * n` for `k`, with `Σw² = var * (n/k)²`. let k_from_total_variance = ; -
// `Σw²` of the compactors in units of (n/k)²: // * Where c is the CAPACITY_DECAY. // * Where h is the height of a compactor (counting from 0). -
// * A weight shift at height h moves a rank by at most one item's weight, so it adds // (2^h)² = 4^h to `Σw²`. // * So each compactor adds y_(h+1) * 4^(h+1) / (y_h * 4^h) times as many // weight-shifts as the one below it: -
// `Σw²` of the sampler in units of (n/k)²: // * `feed` conserves weight, so at most `n / 2^h` emissions happen at height `h`, // each a step of size `2^h` (KLL Lemma 2). Summed, that is at most `n * 2^L`. -
// * Level L-1 sits at depth D = H - L and was replaced, so `k * c^D ≤ CUTOFF`. // Hence `2^D ≥ (k / CUTOFF)^α` where `α = ln 2 / ln(1/c)` (KLL Theorem. 2). // * Recall `n/k ≥ 2^(H-2)` ⇒ `2^H ≤ 4 * n/k`. // * Compute the full variance: -
// * Level L-1 sits at depth D = H - L and was replaced, so `k * c^D ≤ CUTOFF`. // Hence `2^D ≥ (k / CUTOFF)^α` where `α = ln 2 / ln(1/c)` (KLL Theorem. 2). // * Recall `n/k ≥ 2^(H-2)` ⇒ `2^H ≤ 4 * n/k`. // * Compute the full variance: -
// * Compute the full variance: // n * 2^L = n * 2^H / 2^D ≤ (4 * n²/k) / (k / CUTOFF)^α // = (n/k)² * 4 * k * (CUTOFF / k)^α let alpha = f64ln / f64ln; -
// n * 2^L = n * 2^H / 2^D ≤ (4 * n²/k) / (k / CUTOFF)^α // = (n/k)² * 4 * k * (CUTOFF / k)^α let alpha = f64ln / f64ln; let sampler_var = ; -
writeln!?; writeln!?; } else -
writeln!?; writeln!?; } -
-
/// # use polars_core::prelude::*; /// let df1: DataFrame = df!("Ray type" => ["α", "β", "X", "γ"])?; /// let df2: DataFrame = df1.drop("Ray type")?; /// -
/// The per-column-set predicates implied by each top-level OR conjunct of /// `predicate`: `(A₁∧X₁) ∨ (A₂∧X₂)` implies `A₁∨A₂` and `X₁∨X₂` when every `Aᵢ` /// references the same set of columns, likewise every `Xᵢ`. Each derived -
/// `predicate`: `(A₁∧X₁) ∨ (A₂∧X₂)` implies `A₁∨A₂` and `X₁∨X₂` when every `Aᵢ` /// references the same set of columns, likewise every `Xᵢ`. Each derived /// predicate is redundant next to the original, but references a single -
Reproduce This Scan
Run disclude locally to verify. These commands scan the published file; the "new in this release" marks come from comparing it with the previous release's file, which they don't repeat. Add --format sarif for SARIF output. Add --llm (with an API key in the environment) to include the LLM review.
|
&&