API Reference

The public Python API is re-exported from the top-level baleen package and the baleen.eventalign subpackage. Pages here are auto-generated from docstrings.

Top-level convenience imports

The most common entry points are available directly from baleen:

from baleen import (
    run_pipeline_streaming,   # full pipeline (streaming)
    run_pipeline,             # full pipeline (in-memory)
    aggregate_all,            # site aggregation
    write_site_tsv,           # write site_results.tsv
    load_read_results,        # read mod-BAM into a DataFrame
    load_hmm_params,
    save_hmm_params,
)

Reference pages

Page	Covers
Pipeline	run_pipeline_streaming, run_pipeline, results I/O, data classes.
Aggregation	SiteResult, aggregate_all, TSV writers.
Modification Probabilities	The three per-read scoring algorithms.
HMM Training	HMMParams and the three training modes.
Read I/O	mod-BAM writing and loading.
Read-ID Intersection	BAM ∩ FASTQ ∩ BLOW5 enumeration.

Return shape

run_pipeline_streaming returns a 2-tuple (output_paths, metadata). output_paths is a dict with keys site_tsv, read_bam, per_contig_dir, n_total_sites, and n_significant.