Function build_data

codeclash/analysis/viz/line_chart_per_round_changes.py:51–106

Walk logs and build model -> list[list[int]] for rounds 1..ROUNDS. Returns a dict where each model maps to a list of ROUNDS lists, where each inner list contains ints = number of lines changed in that game/round by that model.

(log_dir: Path)

def build_data(log_dir: Path):
    """Walk logs and build model -> list[list[int]] for rounds 1..ROUNDS.

    Returns a dict where each model maps to a list of ROUNDS lists, where each
    inner list contains ints = number of lines changed in that game/round by
    that model.
    """
    model_to_round_lines = {}

    tournaments = [x.parent for x in log_dir.rglob("metadata.json")]
    for game_log_folder in tqdm(tournaments, desc="Scanning tournaments"):
        try:
            with open(game_log_folder / "metadata.json") as f:
                metadata = json.load(f)
        except Exception:
            continue

        try:
            p2m = {
                x["name"]: x["config"]["model"]["model_name"].strip("@").split("/")[-1]
                for x in metadata["config"]["players"]
            }
        except Exception:
            # malformed metadata
            continue

        # ensure models exist in dict
        for model in set(p2m.values()):
            model_to_round_lines.setdefault(model, [[] for _ in range(ROUNDS)])

        # collect changes files per player
        for player_name, model in p2m.items():
            changes_files = (game_log_folder / "players" / player_name).rglob("changes_r*.json")
            for changes_file in changes_files:
                m = re.search(r"changes_r(\d+)\.json", changes_file.name)
                if not m:
                    continue
                round_idx = int(m.group(1))
                if round_idx < 1 or round_idx > ROUNDS:
                    continue

                try:
                    with open(changes_file) as f:
                        changes = json.load(f)
                except Exception:
                    continue

                patch_text = changes.get("incremental_diff")
                if not patch_text:
                    # no diff recorded for this round
                    continue

                num_lines = _lines_changed_from_patch_text(patch_text)
                model_to_round_lines.setdefault(model, [[] for _ in range(ROUNDS)])[round_idx - 1].append(num_lines)

    return model_to_round_lines