experiments/clean_data_scaling_study/dashboard/app.py

"""
Streamlit dashboard for the clean data-scaling study.

Run from the experiments/clean_data_scaling_study/ directory:
    streamlit run dashboard/app.py

Three sections:
  1. Learning curves       — per-epoch metrics for every (model, share) run
  2. Data share vs final   — best/final val mIoU/Dice/IoU/PixelAcc as a function of data share
  3. Inference             — upload an image, see the 4×3 grid of predictions

Reads logs from ../logs and checkpoints from ../checkpoints.
"""
import io
import json
import sys
from pathlib import Path

import numpy as np
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
import streamlit as st
import torch
from PIL import Image
from torchvision import transforms

THIS_DIR = Path(__file__).resolve().parent
EXP_DIR = THIS_DIR.parent
LOGS_DIR = EXP_DIR / "logs"
CKPT_DIR = EXP_DIR / "checkpoints"

sys.path.insert(0, str(EXP_DIR))
from models import MODEL_REGISTRY, PRETTY_NAME  # noqa: E402

MODELS = ["segnet", "unet", "segformer_b0", "segformer_b5"]
SHARES = [25, 50, 100]
METRICS = [
    ("dice", "Dice"),
    ("miou", "mIoU"),
    ("iou", "Foreground IoU"),
    ("pixel_acc", "Pixel Accuracy"),
    ("loss", "Loss"),
]

try:
    st.set_page_config(page_title="Clean Data Scaling Study", layout="wide")
except Exception:
    # already configured by a parent multi-page app
    pass


# ── Loaders ────────────────────────────────────────────────────────────────
@st.cache_data(show_spinner=False)
def load_log(model: str, share: int):
    p = LOGS_DIR / f"{model}_{share}.json"
    if not p.is_file():
        return None
    with open(p) as f:
        return json.load(f)


def log_to_df(log):
    df = pd.DataFrame(log["epochs"])
    df["model"] = log["model"]
    df["share"] = log["share"]
    return df


@st.cache_data(show_spinner=False)
def load_all_logs():
    logs = {}
    for m in MODELS:
        for s in SHARES:
            log = load_log(m, s)
            if log is not None:
                logs[(m, s)] = log
    return logs


def fmt_hms(seconds):
    if seconds is None:
        return "—"
    seconds = int(round(seconds))
    h, rem = divmod(seconds, 3600)
    m, s = divmod(rem, 60)
    return f"{h:d}:{m:02d}:{s:02d}" if h else f"{m:d}:{s:02d}"


def scaling_row(log, kind="best"):
    """One summary row per (model, share). kind ∈ {'best', 'final'}."""
    epochs = log["epochs"]
    row = {"model": PRETTY_NAME[log["model"]], "share": log["share"]}
    if not epochs:
        for k in ("val_dice", "val_miou", "val_iou", "val_pixel_acc"):
            row[k] = None
        return row

    if kind == "best":
        idx = max(range(len(epochs)), key=lambda i: epochs[i].get("val_dice", -1) or -1)
    else:  # final
        idx = len(epochs) - 1
    chosen = epochs[idx]
    row["epoch"] = chosen["epoch"]
    row["val_dice"] = chosen.get("val_dice")
    row["val_miou"] = chosen.get("val_miou")
    row["val_iou"] = chosen.get("val_iou")
    row["val_pixel_acc"] = chosen.get("val_pixel_acc")
    row["wall_clock_seconds"] = log.get("wall_clock_seconds")
    row["wall_clock"] = fmt_hms(log.get("wall_clock_seconds"))
    if epochs:
        per = [e.get("epoch_seconds") for e in epochs if e.get("epoch_seconds") is not None]
        row["sec_per_epoch"] = (sum(per) / len(per)) if per else None
    return row


# ── Inference helpers ──────────────────────────────────────────────────────
@st.cache_resource(show_spinner=False)
def load_ckpt(model_name: str, share: int, kind: str, device: str):
    p = CKPT_DIR / f"{model_name}_{share}_{kind}.pth"
    if not p.is_file():
        return None, False
    builder = MODEL_REGISTRY[model_name]
    model, _, output_is_prob = builder()
    state = torch.load(p, map_location=device, weights_only=False)
    model.load_state_dict(state["model_state_dict"])
    model.to(device).eval()
    output_is_prob = state.get("output_is_prob", output_is_prob)
    return model, output_is_prob


def preprocess(image: Image.Image, image_size: int = 128):
    tf = transforms.Compose([
        transforms.Resize((image_size, image_size)),
        transforms.ToTensor(),
    ])
    return tf(image.convert("RGB")).unsqueeze(0)


def run_inference(model, image_tensor, device, output_is_prob: bool, threshold=0.5):
    with torch.no_grad():
        out = model(image_tensor.to(device))
        probs = out if output_is_prob else torch.sigmoid(out)
        probs = probs.squeeze().cpu().numpy()
    if probs.ndim != 2:
        probs = probs.reshape(probs.shape[-2], probs.shape[-1])
    mask = (probs > threshold).astype(np.float32)
    return probs, mask


def overlay(rgb: np.ndarray, mask: np.ndarray, color=(0, 255, 0), alpha=0.45):
    out = rgb.copy()
    m = mask.astype(bool)
    out[m] = (alpha * np.array(color) + (1 - alpha) * out[m]).astype(np.uint8)
    return out


def heatmap(probs: np.ndarray) -> np.ndarray:
    p = np.clip(probs, 0.0, 1.0)
    rgb = np.zeros((p.shape[0], p.shape[1], 3), dtype=np.uint8)
    rgb[..., 0] = (p * 255).astype(np.uint8)
    rgb[..., 1] = (np.maximum(0, 1 - 2 * np.abs(p - 0.5)) * 255).astype(np.uint8)
    rgb[..., 2] = ((1 - p) * 255).astype(np.uint8)
    return rgb


# ── UI ─────────────────────────────────────────────────────────────────────
st.title("📊 Clean Data-Scaling Study — 4 architectures × 3 data shares")
st.caption(
    "Trained from scratch on a deduplicated dataset (final_data_clean/). "
    "All numbers use the same global confusion-matrix metric code."
)

logs = load_all_logs()

if not logs:
    st.warning("No logs found in `../logs/`. Run training first (`./run_all.sh`).")

tab_curves, tab_scaling, tab_infer = st.tabs(
    ["1 · Learning curves", "2 · Data share vs final", "3 · Inference"]
)


# ── Tab 1: Learning curves ─────────────────────────────────────────────────
with tab_curves:
    st.subheader("Per-epoch metrics")
    if not logs:
        st.info("Waiting for training logs.")
    else:
        col_m, col_split = st.columns([2, 2])
        with col_m:
            metric_key, metric_label = st.selectbox(
                "Metric", METRICS, format_func=lambda x: x[1],
            )
        with col_split:
            split = st.radio("Split", ["val", "train", "both"], horizontal=True, index=0)

        for model in MODELS:
            available = [s for s in SHARES if (model, s) in logs]
            if not available:
                continue
            st.markdown(f"#### {PRETTY_NAME[model]}")
            fig = go.Figure()
            for share in available:
                df = log_to_df(logs[(model, share)])
                if split in ("val", "both"):
                    col = f"val_{metric_key}"
                    if col in df.columns and df[col].notna().any():
                        sub = df.dropna(subset=[col])
                        fig.add_trace(go.Scatter(
                            x=sub["epoch"], y=sub[col], mode="lines",
                            name=f"{share}% val",
                        ))
                if split in ("train", "both"):
                    col = f"train_{metric_key}"
                    if col in df.columns and df[col].notna().any():
                        sub = df.dropna(subset=[col])
                        fig.add_trace(go.Scatter(
                            x=sub["epoch"], y=sub[col], mode="lines", line=dict(dash="dot"),
                            name=f"{share}% train",
                        ))
            fig.update_layout(
                xaxis_title="Epoch", yaxis_title=metric_label, height=360,
                margin=dict(l=10, r=10, t=10, b=10),
                legend=dict(orientation="h", y=-0.2),
            )
            st.plotly_chart(fig, use_container_width=True)


# ── Tab 2: Data share vs final ─────────────────────────────────────────────
with tab_scaling:
    st.subheader("Val metrics vs data share")
    if not logs:
        st.info("Waiting for training logs.")
    else:
        kind = st.radio("Checkpoint", ["best", "final"], horizontal=True, index=0)

        rows = [scaling_row(log, kind=kind) for log in logs.values()]
        df = pd.DataFrame(rows).sort_values(["model", "share"]).reset_index(drop=True)

        display_df = df.drop(columns=["wall_clock_seconds", "sec_per_epoch"], errors="ignore")
        st.dataframe(display_df, use_container_width=True, hide_index=True)

        trained_seconds = df["wall_clock_seconds"].dropna().sum()
        if trained_seconds:
            st.caption(
                f"⏱  Total training wall-clock across all runs: "
                f"**{fmt_hms(trained_seconds)}**  ({trained_seconds:,.0f} s)"
            )

        col1, col2 = st.columns(2)
        with col1:
            fig = px.line(
                df.dropna(subset=["val_miou"]),
                x="share", y="val_miou", color="model",
                markers=True, title=f"Val mIoU ({kind})",
                labels={"share": "Training data (%)", "val_miou": "Val mIoU"},
            )
            fig.update_xaxes(tickvals=SHARES)
            st.plotly_chart(fig, use_container_width=True)
        with col2:
            fig = px.line(
                df.dropna(subset=["val_dice"]),
                x="share", y="val_dice", color="model",
                markers=True, title=f"Val Dice ({kind})",
                labels={"share": "Training data (%)", "val_dice": "Val Dice"},
            )
            fig.update_xaxes(tickvals=SHARES)
            st.plotly_chart(fig, use_container_width=True)

        col3, col4 = st.columns(2)
        with col3:
            fig = px.bar(
                df.dropna(subset=["val_iou"]),
                x="share", y="val_iou", color="model", barmode="group",
                title=f"Val foreground IoU ({kind})",
                labels={"share": "Training data (%)", "val_iou": "Val IoU (foreground)"},
            )
            fig.update_xaxes(tickvals=SHARES)
            st.plotly_chart(fig, use_container_width=True)
        with col4:
            fig = px.bar(
                df.dropna(subset=["val_pixel_acc"]),
                x="share", y="val_pixel_acc", color="model", barmode="group",
                title=f"Val pixel accuracy ({kind})",
                labels={"share": "Training data (%)", "val_pixel_acc": "Val pixel acc"},
            )
            fig.update_xaxes(tickvals=SHARES)
            st.plotly_chart(fig, use_container_width=True)

        st.markdown("##### Training time")
        time_df = df.dropna(subset=["wall_clock_seconds"]).assign(
            wall_minutes=lambda d: d["wall_clock_seconds"] / 60.0
        )
        if not time_df.empty:
            tcol1, tcol2 = st.columns(2)
            with tcol1:
                fig = px.bar(
                    time_df, x="share", y="wall_minutes", color="model", barmode="group",
                    title="Total training time (minutes)",
                    labels={"share": "Training data (%)", "wall_minutes": "Wall clock (min)"},
                )
                fig.update_xaxes(tickvals=SHARES)
                st.plotly_chart(fig, use_container_width=True)
            with tcol2:
                fig = px.bar(
                    time_df.dropna(subset=["sec_per_epoch"]),
                    x="share", y="sec_per_epoch", color="model", barmode="group",
                    title="Average seconds per epoch",
                    labels={"share": "Training data (%)", "sec_per_epoch": "Seconds / epoch"},
                )
                fig.update_xaxes(tickvals=SHARES)
                st.plotly_chart(fig, use_container_width=True)


# ── Tab 3: Inference ───────────────────────────────────────────────────────
with tab_infer:
    st.subheader("Upload an image — 4 models × 3 shares = 12 segmentations")
    st.caption(
        "Each cell uses one (model, data-share) checkpoint. "
        "Toggle best vs final to see end-of-training behavior."
    )

    debug = st.checkbox("debug", value=False, key="infer_debug")
    col_a, col_b, col_c, col_d = st.columns([2, 2, 2, 2])
    with col_a:
        kind = st.radio("Checkpoint", ["best", "final"], horizontal=True, key="infer_kind")
    with col_b:
        threshold = st.slider("Threshold", 0.0, 1.0, 0.5, 0.05, key="infer_thr")
    with col_c:
        view = st.radio("View", ["mask", "overlay", "heatmap"], horizontal=True, key="infer_view")
    with col_d:
        cell_w = st.select_slider(
            "Cell size (px)", options=[140, 180, 220, 260], value=180, key="infer_cell"
        )

    uploaded = st.file_uploader(
        "Drop an image (jpg/png)", type=["jpg", "jpeg", "png"], key="infer_upload"
    )

    if uploaded is not None:
        device = "cuda" if torch.cuda.is_available() else "cpu"
        try:
            raw_bytes = uploaded.getvalue()
            img = Image.open(io.BytesIO(raw_bytes)).convert("RGB")
        except Exception as e:
            st.error(f"Could not decode uploaded image: {e}")
            st.stop()

        st.caption(f"📁 `{uploaded.name}` — {img.size[0]}×{img.size[1]} px, {len(raw_bytes)/1024:.1f} KB")

        x = preprocess(img, image_size=128)
        rgb_small = (x.squeeze().permute(1, 2, 0).numpy() * 255).astype(np.uint8)
        st.image([img, rgb_small], width=cell_w * 2, caption=["original", "128×128 (model input)"])

        if debug:
            st.write(f"tensor={tuple(x.shape)}  device={device}  models={MODELS}  shares={SHARES}")

        st.markdown("##### Predictions  (rows = model, columns = data share)")

        def render_cell(probs, mask, rgb):
            if view == "mask":
                return (mask * 255).astype(np.uint8)
            if view == "overlay":
                return overlay(rgb, mask)
            return heatmap(probs)

        for model_name in MODELS:
            cols = st.columns(len(SHARES))
            for col, share in zip(cols, SHARES):
                with col:
                    st.markdown(f"**{PRETTY_NAME[model_name]} · {share}%**")
                    try:
                        with st.spinner(f"loading {model_name} {share}%…"):
                            m, output_is_prob = load_ckpt(model_name, share, kind, device)
                        if m is None:
                            st.warning(f"missing `{model_name}_{share}_{kind}.pth`")
                            continue
                        probs, mask = run_inference(m, x, device, output_is_prob, threshold)
                        cell_img = render_cell(probs, mask, rgb_small)
                        st.image(cell_img, width=cell_w)
                        st.caption(
                            f"cov={float(mask.mean())*100:.1f}%   "
                            f"p[{probs.min():.2f},{probs.max():.2f}]"
                        )
                    except Exception as e:
                        st.error(f"{model_name} {share}% — {type(e).__name__}")
                        st.exception(e)
    else:
        st.info("Upload an image to run inference across all 12 trained models.")