Benchmark API

Tools for measuring and comparing sorting performance.

Functions

compare()

compare(
    size: int = 10000,
    pattern: str = "random",
    algorithms: list[str] | None = None
) -> dict[str, float]

Compare sorting algorithms on specific data pattern.

Parameters: - size: Array size - pattern: Data pattern ("random", "sorted", "reverse", "nearly_sorted", "duplicates") - algorithms: List of algorithm names (None = all)

Returns: Dictionary mapping algorithm names to median times (ms)

Example:

from ordr.benchmark import compare, display_comparison

results = compare(size=10000, pattern="random")
display_comparison(results)

---

profile()

profile(
    sizes: list[int] = [1000, 10000, 100000],
    pattern: str = "random",
    algorithm: str = "smart"
) -> None

Profile an algorithm across different input sizes.

Example:

from ordr.benchmark import profile

profile(
    sizes=[1000, 10000, 100000],
    pattern="random",
    algorithm="smart"
)

---

display_comparison()

display_comparison(
    results: dict[str, float],
    baseline: str = "builtin"
) -> None

Display benchmark results in a rich table.

---

Data Generators

generate_random()

generate_random(size: int) -> list[int]

Generate random array.

---

generate_sorted()

generate_sorted(size: int) -> list[int]

Generate sorted array.

---

generate_reverse()

generate_reverse(size: int) -> list[int]

Generate reverse sorted array.

---

generate_nearly_sorted()

generate_nearly_sorted(size: int) -> list[int]

Generate nearly sorted array (~90% sorted).

---

generate_duplicates()

generate_duplicates(size: int) -> list[int]

Generate array with many duplicates.

---

Example Usage

from ordr.benchmark import compare, profile, display_comparison

# Compare on random data
results = compare(size=10000, pattern="random")
display_comparison(results)

# Compare specific algorithms
results = compare(
    size=10000,
    pattern="nearly_sorted",
    algorithms=["smart", "tim", "pdq"]
)
display_comparison(results)

# Profile across sizes
profile(
    sizes=[1000, 10000, 100000],
    pattern="random",
    algorithm="smart"
)