Blame for MedEvalKit/visualize_grounding.py

"""
visualize_grounding.py
Visualize layer-wise mean divergence for grounding ablation experiments,
comparing COCO vs VGMED across models and ablation types.

Usage:
    python visualize_grounding.py
    python visualize_grounding.py --ablation mask_region
    python visualize_grounding.py --q_type attribute
    python visualize_grounding.py --output_dir my_plots
"""
import argparse
import os

import matplotlib.pyplot as plt
import numpy as np
import torch

GROUNDING_ROOT = "grounding_results"

MODELS = {
    "HuatuoGPT-V-7B":   "HuatuoGPT-V-7B/HuatuoGPT-Vision-7B",
    "LLaVA-7B":         "LLaVA-7B/llava-1.5-7b-hf",
    "LLaVA-13B":        "LLaVA-13B/llava-1.5-13b-hf",
    "LLaVA-Med-7B":     "LLaVA-Med-7B/llava-med-v1.5-mistral-7b",
    "MedGemma-4B-IT":   "MedGemma-4B-IT/medgemma-4b-it",
    "Qwen2.5VL-3B":     "Qwen2.5VL-3B/Qwen2.5-VL-3B-Instruct",
    "Qwen2.5VL-7B":     "Qwen2.5VL-7B/Qwen2.5-VL-7B-Instruct",
    "Qwen3VL-2B":       "Qwen3VL-2B/Qwen3-VL-2B-Instruct",
    "Qwen3VL-4B":       "Qwen3VL-4B/Qwen3-VL-4B-Instruct",
    "Qwen3VL-8B":       "Qwen3VL-8B/Qwen3-VL-8B-Instruct",
}

ABLATIONS = {
    "mask_region": "Mask Key Regions",
    "mask_text":   "Mask Key Words",
    "diff_region": "Mask Key Regions Compared with Mask Random Regions",
    "diff_text":   "Mask Key Words Compared with Mask Random Words",
}

Q_TYPES = ["localization", "attribute"]
DATASETS = ["COCO", "VGMED"]

# ── shared style constants (kept in sync with visualize_attention.py) ──────────
# Each model is saved as its own standalone PNG, so every figure must be readable
# on its own at a small size: large fonts, thick lines, high DPI.
COLORS       = {"COCO": "#2196F3", "VGMED": "#F44336"}
DIFF_POS     = "#F44336"
DIFF_NEG     = "#2196F3"
FIG_WIDTH    = 8.0
ROW_HEIGHT   = 5.0
TITLE_FS     = 22
AXIS_FS      = 18
TICK_FS      = 16
LEGEND_FS    = 16
MARKER       = "o"
MARKER_SIZE  = 9
LINE_WIDTH   = 3.5
GRID_ALPHA   = 0.3
DPI          = 600
plt.rcParams["font.weight"] = "bold"
plt.rcParams["axes.labelweight"] = "bold"
plt.rcParams["axes.titleweight"] = "bold"
plt.rcParams["axes.linewidth"] = 1.5
# ──────────────────────────────────────────────────────────────────────────────


def load_pt(path):
    data = torch.load(path, map_location="cpu", weights_only=False)
    if isinstance(data, torch.Tensor):
        return data.numpy()
    arr = np.stack(data, axis=0)
    return arr.mean(axis=0)


def plot_ablation(ablation_key, ablation_label, q_type, output_dir, grounding_root):
    """One standalone PNG per model: top = divergence curves, bottom = VGMED-COCO diff bars."""
    for model_name, model_dir in MODELS.items():
        fig, (ax_top, ax_bot) = plt.subplots(
            2, 1, figsize=(FIG_WIDTH, ROW_HEIGHT * 2),
        )
        fig.suptitle(
            f"{model_name} — {ablation_label} ({q_type})",
            fontsize=TITLE_FS, fontweight="bold",
        )

        loaded = {}
        for ds in DATASETS:
            fname = f"{ds}_{q_type}_mean_divergence_per_layer.pt"
            fpath = os.path.join(grounding_root, model_dir, ablation_key, fname)
            if not os.path.exists(fpath):
                print(f"[skip] {fpath}")
                continue
            loaded[ds] = load_pt(fpath)

        # top: curves
        for ds, vals in loaded.items():
            layers = np.arange(len(vals))
            ax_top.plot(layers, vals, label=ds, color=COLORS[ds],
                        linewidth=LINE_WIDTH, marker=MARKER, markersize=MARKER_SIZE)
        ax_top.set_ylabel("Mean Divergence", fontsize=AXIS_FS)
        ax_top.tick_params(labelsize=TICK_FS)
        ax_top.grid(True, alpha=GRID_ALPHA)
        if loaded:
            ax_top.legend(fontsize=LEGEND_FS)

        # bottom: VGMED − COCO diff bars
        if "VGMED" in loaded and "COCO" in loaded:
            vgmed = loaded["VGMED"]
            coco  = loaded["COCO"]
            n = min(len(vgmed), len(coco))
            diff = vgmed[:n] - coco[:n]
            layers = np.arange(n)
            ax_bot.bar(layers, diff,
                       color=[DIFF_POS if d >= 0 else DIFF_NEG for d in diff],
                       alpha=0.7, width=0.8)
            ax_bot.axhline(0, color="black", linewidth=0.8, linestyle="--")
            ax_bot.set_xlabel("Layer", fontsize=AXIS_FS)
            ax_bot.set_ylabel("VGMED − COCO", fontsize=AXIS_FS)
            ax_bot.tick_params(labelsize=TICK_FS)
            ax_bot.grid(True, alpha=GRID_ALPHA, axis="y")
        else:
            ax_bot.set_visible(False)

        plt.tight_layout()
        q_type_dir = os.path.join(output_dir, q_type, ablation_key)
        os.makedirs(q_type_dir, exist_ok=True)
        out_path = os.path.join(q_type_dir, f"{model_name}_{ablation_key}_{q_type}.png")
        plt.savefig(out_path, dpi=DPI, bbox_inches="tight")
        plt.close()
        print(f"Saved: {out_path}")


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("--ablation", choices=list(ABLATIONS.keys()), default=None,
                        help="Plot only one ablation type (default: all)")
    parser.add_argument("--q_type", choices=Q_TYPES, default=None,
                        help="Plot only one q_type (default: all)")
    parser.add_argument("--grounding_root", default=GROUNDING_ROOT)
    parser.add_argument("--output_dir", default="grounding_plots")
    args = parser.parse_args()

    ablations = {args.ablation: ABLATIONS[args.ablation]} if args.ablation else ABLATIONS
    q_types = [args.q_type] if args.q_type else Q_TYPES

    for ablation_key, ablation_label in ablations.items():
        for q_type in q_types:
            plot_ablation(ablation_key, ablation_label, q_type,
                          args.output_dir, args.grounding_root)


if __name__ == "__main__":
    main()