Documentation

cli

Command-line interface for cfabric benchmarks.

Functions

function
cli() → None

Context-Fabric benchmarking tool.

Context-Fabric benchmarking tool. Run performance benchmarks comparing Text-Fabric and Context-Fabric.
function
create_output_dir(base_dir: Path) → Path

Create timestamped output directory.

Parameters
  • base_dir: Path
function
environment() → None

Print test environment information.

function
full(memory_runs: int, latency_runs: int, progressive_runs: int, warmup: int, workers: int, corpora_dir: Path | None, output_dir: Path, queries: int, iterations: int, max_corpora: int, no_pdf: bool) → None

Run full benchmark suite (memory, latency, progressive).

Parameters
  • memory_runs: int
  • latency_runs: int
  • progressive_runs: int
  • warmup: int
  • workers: int
  • corpora_dir: Path | None
  • output_dir: Path
  • queries: int
  • iterations: int
  • max_corpora: int
  • no_pdf: bool
function
get_default_corpora_dir() → Path

Get the default corpora directory.

function
latency(corpus: str, queries: int, iterations: int, runs: int, validation_corpus: str | None, corpora_dir: Path | None, output_dir: Path) → None

Run latency benchmarks only.

Parameters
  • corpus: str
  • queries: int
  • iterations: int
  • runs: int
  • validation_corpus: str | None
  • corpora_dir: Path | None
  • output_dir: Path
function
memory(corpus: tuple[(str, ...)], runs: int, warmup: int, workers: int, corpora_dir: Path | None, output_dir: Path) → None

Run memory benchmarks only.

Parameters
  • corpus: tuple[(str, ...)]
  • runs: int
  • warmup: int
  • workers: int
  • corpora_dir: Path | None
  • output_dir: Path
function
progressive(max_corpora: int, runs: int, corpora_dir: Path | None, output_dir: Path) → None

Run progressive loading benchmark.

Parameters
  • max_corpora: int
  • runs: int
  • corpora_dir: Path | None
  • output_dir: Path
function
validate_patterns(corpora_dir: Path | None, output_dir: Path) → None

Validate curated BHSA patterns on the BHSA corpus.

Parameters
  • corpora_dir: Path | None
  • output_dir: Path
function
visualize(results_dir: Path, format: Literal[('pdf', 'png', 'both')], charts: str) → None

Generate visualizations from existing results.

Generate visualizations from existing results. By default generates: progressive, latency, multicorpus charts. Per-corpus memory charts are not supported (require corpus stats not stored in CSV).
Parameters
  • results_dir: Path
  • format: Literal[('pdf', 'png', 'both')]
  • charts: str