#!/usr/bin/env python3 """Build a deterministic, balanced terminal/SWE SFT mixture.""" from __future__ import annotations import argparse import json from pathlib import Path from datasets import Dataset, concatenate_datasets, load_dataset COMMON_COLUMNS = ["messages", "tools", "source", "num_tool_calls"] def normalized_component(path: Path, label: str) -> Dataset: dataset = load_dataset(str(path), split="train").select_columns(COMMON_COLUMNS) return dataset.add_column("mix_component", [label] * len(dataset)) def write_shards(dataset: Dataset, output_dir: Path, shards: int) -> None: output_dir.mkdir(parents=True, exist_ok=True) for index in range(shards): dataset.shard(shards, index, contiguous=True).to_parquet( output_dir / f"train-{index:05d}-of-{shards:05d}.parquet" ) def main() -> None: parser = argparse.ArgumentParser() parser.add_argument("--terminal", type=Path, required=True) parser.add_argument("--swe", type=Path, required=True) parser.add_argument("--adapters", type=Path) parser.add_argument("--output", type=Path, required=True) parser.add_argument("--swe-repeats", type=int, default=2) parser.add_argument("--adapter-repeats", type=int, default=1) parser.add_argument("--seed", type=int, default=20260814) parser.add_argument("--shards", type=int, default=64) args = parser.parse_args() terminal = normalized_component(args.terminal, "nemotron-terminal-skill") swe = normalized_component(args.swe, "swe-smith") parts = [terminal, *([swe] * args.swe_repeats)] adapters = None if args.adapters is not None: adapters = normalized_component(args.adapters, "nemotron-terminal-adapters") parts.extend([adapters] * args.adapter_repeats) mixed = concatenate_datasets(parts).shuffle(seed=args.seed) write_shards(mixed, args.output, min(args.shards, len(mixed))) report = { "terminal_path": str(args.terminal), "terminal_rows": len(terminal), "swe_path": str(args.swe), "swe_rows": len(swe), "swe_repeats": args.swe_repeats, "adapter_path": str(args.adapters) if args.adapters is not None else None, "adapter_rows": len(adapters) if adapters is not None else 0, "adapter_repeats": args.adapter_repeats if adapters is not None else 0, "output_rows": len(mixed), "shuffle_seed": args.seed, "mix": { "nemotron-terminal-skill": len(terminal), "swe-smith": len(swe) * args.swe_repeats, **( {"nemotron-terminal-adapters": len(adapters) * args.adapter_repeats} if adapters is not None else {} ), }, } (args.output / "provenance.json").write_text(json.dumps(report, indent=2) + "\n") print(json.dumps(report, indent=2)) if __name__ == "__main__": main()