"""Self-contained, print-friendly HTML report written into every result bundle.""" from __future__ import annotations import html from pathlib import Path STATUS_COLOR = {"pass": "#1baf7a", "warn": "#eda100", "fail": "#e34948", "info": "#2a78d6", "skipped": "#8a8a8a"} def _fmt(v): if isinstance(v, float): return f"{v:.3f}" return html.escape("" if v is None else str(v)) def _table(rows, cols): if not rows: return "

None.

" head = "".join(f"{html.escape(c)}" for c in cols) body = "".join("" + "".join(f"{_fmt(r.get(c))}" for c in cols) + "" for r in rows) return f"{head}{body}
" def write_report_html(res: dict, path: Path) -> None: s = res["summary"] diags = "".join( f"
{d['status']}" f"

{html.escape(d['title'])}

{html.escape(d['headline'])}

" f"

{html.escape(d['detail'])}

For model building: {html.escape(d['ml_implication'])}

" + (f"

{html.escape(d['caveat'])}

" if d["caveat"] else "") + "
" for d in res["diagnostics"]) recs = "".join(f"
  • {html.escape(r['title'])} ({html.escape(r['file'])}): " f"{html.escape(r['text'])}
    • " for r in res["recommendations"]) sections = [ ("Negative-set balance", _table(res["balance"], ["negative_set", "n_negatives", "positive_mean", "negative_mean", "smd", "ks_statistic", "length_smd"])), ("Splits and leakage gates", _table(res["splits"], ["regime", "train_pairs", "test_pairs", "train_proteins", "test_proteins", "protein_overlap", "test_components", "passed"])), ("Protein-novelty curve", _table(res["novelty"], ["representation", "regime", "model", "test_pairs", "roc_auc", "pr_auc", "mcc"])), ("Topology-only and length-only controls", _table([r for r in res["topology"] if r["model"] == "lr"], ["feature", "evaluated_on", "roc_auc", "pr_auc", "mcc"])), ("Consistent-negative revalidation", _table(res["revalidation"], ["representation", "regime", "negative_set", "test_pairs", "roc_auc", "pr_auc", "mcc"])), ("Component-aware uncertainty (R3)", _table(res["uncertainty"], ["representation", "negative_set", "n_test", "n_components", "roc_auc", "boot_low", "boot_high", "jack_low", "jack_high", "boot_p_bh", "jack_p_bh"])), ] if res.get("species"): sp = res["species"] sections += [("Species composition", _table(sp["composition"], ["species", "proteins", "positive_within", "negative_within", "positive_cross", "negative_cross"])), ("Cross-species fraction", _table(sp["cross_species"], ["set", "n_pairs", "cross_species", "cross_species_fraction"])), ("Same-species shortcut", _table(sp["same_species_shortcut"], ["evaluated_on", "n_evaluated", "roc_auc"])), ("Within-species controls", _table(sp["within_metrics"], ["species", "negative_set", "representation", "model", "test_pairs", "roc_auc", "pr_auc"]))] body = "".join(f"

    {html.escape(t)}

    {c}
    " for t, c in sections) doc = f""" PPI Benchmark Report

    PPI benchmark evaluation

    PPI Benchmark Studio · engine {res['engine_version']} (framework {res['framework_version']}) · {res['generated_at']} · seed {res['config']['seed']} · preset {res['config']['preset']}

    {html.escape(s['verdict'])} — {html.escape(s['headline'])}

    Readiness score {s['score']}/100. {html.escape(s['score_note'])}

    Diagnostics

    {diags}

    Recommendations

    {body}

    Methods

    {html.escape(res['methods_text'])}

    """ Path(path).write_text(doc)