from __future__ import annotations

import json
import subprocess
import sys
import time
import uuid
import urllib.error
import urllib.request
from pathlib import Path
from typing import Any

from .config import settings


def _run_command(command: list[str], cwd: Path | None = None, timeout: int = 60) -> dict[str, Any]:
    try:
        result = subprocess.run(
            command,
            cwd=str(cwd or settings.project_root),
            capture_output=True,
            text=True,
            timeout=timeout,
        )
        return {
            "passed": result.returncode == 0,
            "returncode": result.returncode,
            "stdout": result.stdout[-4000:],
            "stderr": result.stderr[-4000:],
        }
    except subprocess.TimeoutExpired as exc:
        return {
            "passed": False,
            "returncode": None,
            "stdout": (exc.stdout or "")[-4000:] if isinstance(exc.stdout, str) else "",
            "stderr": (exc.stderr or "")[-4000:] if isinstance(exc.stderr, str) else "",
            "error": f"command timed out after {timeout}s",
        }
    except Exception as exc:
        return {"passed": False, "returncode": None, "stdout": "", "stderr": "", "error": str(exc)}


def _run_snippet(code: str, cwd: Path | None = None, timeout: int = 60) -> dict[str, Any]:
    return _run_command([sys.executable, "-c", code], cwd=cwd, timeout=timeout)


def _run_conda_snippet(env_name: str, code: str, cwd: Path | None = None, timeout: int = 60) -> dict[str, Any]:
    detail = _run_command(["conda", "run", "-n", env_name, "python", "-c", code], cwd=cwd, timeout=timeout)
    detail["env"] = env_name
    return detail


MMSEG_FULL_BUILD_SNIPPET = (
    "from mmseg.utils import register_all_modules; "
    "register_all_modules(init_default_scope=True); "
    "from mmengine.config import Config; "
    "from mmseg.registry import MODELS; "
    "import mmcv._ext; "
    "cfg=Config.fromfile({config_path!r}); "
    "model=MODELS.build(cfg.model); "
    "print(type(model).__name__)"
)


SEGMODEL_TRAIN_STEP_SNIPPET = (
    "import torch, segmentation_models_pytorch as smp; "
    "torch.manual_seed(7); "
    "model=smp.Unet(encoder_name='resnet18', encoder_weights=None, classes=2).train(); "
    "inputs=torch.randn(2,3,64,64); "
    "targets=torch.randint(0,2,(2,64,64)); "
    "optimizer=torch.optim.SGD(model.parameters(), lr=1e-3); "
    "outputs=model(inputs); "
    "loss=torch.nn.functional.cross_entropy(outputs, targets); "
    "loss.backward(); optimizer.step(); "
    "print('loss', round(float(loss.detach()), 6), 'shape', tuple(outputs.shape))"
)


def _yolo_tiny_train_snippet(root: Path, weight: Path) -> str:
    return (
        "import shutil, cv2, numpy as np; "
        "from pathlib import Path; "
        "from ultralytics import YOLO; "
        f"root=Path({str(root)!r}); weight={str(weight)!r}; "
        "shutil.rmtree(root, ignore_errors=True); "
        "[ (root / item).mkdir(parents=True, exist_ok=True) for item in ['images/train','images/val','labels/train','labels/val','runs'] ]; "
        "image=np.zeros((64,64,3), dtype=np.uint8); "
        "cv2.rectangle(image, (16,16), (48,48), (255,255,255), -1); "
        "label='0 0.25 0.25 0.75 0.25 0.75 0.75 0.25 0.75\\n'; "
        "\nfor split in ['train','val']:\n"
        "    cv2.imwrite(str(root / 'images' / split / 'sample.jpg'), image)\n"
        "    (root / 'labels' / split / 'sample.txt').write_text(label, encoding='utf-8')\n"
        "(root / 'data.yaml').write_text('path: '+str(root)+'\\ntrain: images/train\\nval: images/val\\nnc: 1\\nnames:\\n  0: object\\n', encoding='utf-8'); "
        "model=YOLO(weight); "
        "model.train(data=str(root/'data.yaml'), epochs=1, imgsz=64, batch=1, workers=0, device='cpu', project=str(root/'runs'), name='tiny', exist_ok=True, verbose=False, plots=False, val=False); "
        "results=root/'runs'/'tiny'/'results.csv'; best=root/'runs'/'tiny'/'weights'/'best.pt'; "
        "assert results.exists() and results.stat().st_size > 0; "
        "assert best.exists() and best.stat().st_size > 0; "
        "print('results', results, results.stat().st_size, 'best', best.stat().st_size)"
    )


def _yolo_heatmap_snippet(root: Path) -> str:
    script_path = settings.source_root / "Seg_All_In_One_YoloModel" / "yolo_predict_visualize_nn.py"
    return (
        "from pathlib import Path; "
        "import importlib.util, shutil, sys, types; "
        "fake=types.ModuleType('yolo_config'); "
        "fake.MODEL_CONFIGS={'YOLO11n-seg': {}}; "
        "fake.TEST_IMAGE_DIR=''; fake.PREDICT_BEST_MODEL_DIR=Path('.'); fake.show_config_summary=lambda: None; "
        "sys.modules['yolo_config']=fake; "
        f"script=Path({str(script_path)!r}); "
        "spec=importlib.util.spec_from_file_location('yolo_heatmap_mod', script); "
        "mod=importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); "
        f"root=Path({str(root)!r}); "
        "base=root/'runs'/'tiny'; "
        "heatmap_root=base/'HeartMap_Visual'; "
        "shutil.rmtree(heatmap_root, ignore_errors=True); "
        "mod.visualize_nn_comprehensive(str(base/'weights'/'best.pt'), str(root/'images'/'val'/'sample.jpg'), base, 'best.pt', 'GradCAM', 'model.model.model[9]', 'YOLO11n-seg'); "
        "outputs=sorted(heatmap_root.rglob('*.jpg')); "
        "assert len(outputs) >= 2; "
        "print('heatmaps', len(outputs), [str(item.relative_to(base)) for item in outputs[:4]])"
    )


def _mmseg_train_step_snippet(config_path: Path) -> str:
    return (
        "import torch; "
        "from mmengine.config import Config; "
        "from mmengine.structures import PixelData; "
        "from mmseg.registry import MODELS; "
        "from mmseg.structures import SegDataSample; "
        "from mmseg.utils import register_all_modules; "
        "register_all_modules(init_default_scope=True); "
        f"cfg=Config.fromfile({str(config_path)!r}); "
        "cfg.model.backbone.init_cfg=None; cfg.model.pretrained=None; "
        "model=MODELS.build(cfg.model).train(); "
        "sample=SegDataSample(); "
        "sample.gt_sem_seg=PixelData(data=torch.randint(0,19,(1,64,64), dtype=torch.long)); "
        "losses=model(torch.randn(1,3,64,64), [sample], mode='loss'); "
        "loss=sum(value if torch.is_tensor(value) else sum(value) for value in losses.values()); "
        "optimizer=torch.optim.SGD(model.parameters(), lr=1e-4); "
        "loss.backward(); optimizer.step(); "
        "print('loss', round(float(loss.detach()), 6), sorted(losses.keys()))"
    )


def _request_json(method: str, url: str, payload: dict[str, Any] | None = None, timeout: int = 10) -> dict[str, Any]:
    data = None
    headers = {"Accept": "application/json"}
    if payload is not None:
        data = json.dumps(payload, ensure_ascii=False).encode("utf-8")
        headers["Content-Type"] = "application/json"
    request = urllib.request.Request(url, data=data, headers=headers, method=method)
    try:
        with urllib.request.urlopen(request, timeout=timeout) as response:
            body = response.read().decode("utf-8", errors="replace")
            return {"passed": 200 <= response.status < 300, "status": response.status, "body": body, "json": json.loads(body) if body else None}
    except urllib.error.HTTPError as exc:
        body = exc.read().decode("utf-8", errors="replace")
        return {"passed": False, "status": exc.code, "body": body}
    except Exception as exc:
        return {"passed": False, "error": str(exc)}


def _request_text(url: str, timeout: int = 10) -> dict[str, Any]:
    try:
        with urllib.request.urlopen(url, timeout=timeout) as response:
            body = response.read().decode("utf-8", errors="replace")
            return {"passed": 200 <= response.status < 300, "status": response.status, "body": body}
    except Exception as exc:
        return {"passed": False, "error": str(exc)}


def _post_multipart(url: str, field: str, filename: str, content: bytes, content_type: str = "text/plain", timeout: int = 10) -> dict[str, Any]:
    boundary = f"----SegAcceptance{uuid.uuid4().hex}"
    body = b"".join(
        [
            f"--{boundary}\r\n".encode(),
            f'Content-Disposition: form-data; name="{field}"; filename="{filename}"\r\n'.encode(),
            f"Content-Type: {content_type}\r\n\r\n".encode(),
            content,
            f"\r\n--{boundary}--\r\n".encode(),
        ]
    )
    request = urllib.request.Request(
        url,
        data=body,
        method="POST",
        headers={"Content-Type": f"multipart/form-data; boundary={boundary}", "Accept": "application/json"},
    )
    try:
        with urllib.request.urlopen(request, timeout=timeout) as response:
            text = response.read().decode("utf-8", errors="replace")
            return {"passed": 200 <= response.status < 300, "status": response.status, "body": text, "json": json.loads(text) if text else None}
    except urllib.error.HTTPError as exc:
        text = exc.read().decode("utf-8", errors="replace")
        return {"passed": False, "status": exc.code, "body": text}
    except Exception as exc:
        return {"passed": False, "error": str(exc)}


def _poll_job(base_url: str, job_id: str, timeout: int = 90) -> dict[str, Any]:
    deadline = time.time() + timeout
    last = None
    while time.time() < deadline:
        result = _request_json("GET", f"{base_url}/api/jobs/{job_id}", timeout=10)
        last = result
        job = result.get("json") if result.get("passed") else None
        if job and job.get("status") in {"success", "failed", "cancelled"}:
            return {"passed": job.get("status") == "success", "job": job}
        time.sleep(1)
    return {"passed": False, "error": "job timed out", "last": last}


def _create_job_and_wait(base_url: str, task_type: str, params: dict[str, Any], timeout: int = 90) -> dict[str, Any]:
    created = _request_json("POST", f"{base_url}/api/jobs", {"type": task_type, "params": params}, timeout=10)
    if not created.get("passed") or not created.get("json"):
        return {"passed": False, "created": created}
    job_id = created["json"]["id"]
    polled = _poll_job(base_url, job_id, timeout=timeout)
    events = _request_text(f"{base_url}/api/jobs/{job_id}/events", timeout=10)
    return {"passed": polled["passed"], "created": created["json"], "polled": polled, "events": events}


def _create_job_with_retry(base_url: str, task_type: str, params: dict[str, Any], attempts: int = 2, timeout: int = 90) -> dict[str, Any]:
    results = []
    for _ in range(attempts):
        result = _create_job_and_wait(base_url, task_type, params, timeout=timeout)
        results.append(result)
        if result.get("passed"):
            return {"passed": True, "attempts": results}
        time.sleep(1)
    return {"passed": False, "attempts": results}


def _write_acceptance_images(root: Path) -> tuple[Path, Path, Path]:
    import cv2
    import numpy as np

    image_dir = root / "images"
    label_dir = root / "labels"
    result_dir = root / "stacked"
    image_dir.mkdir(parents=True, exist_ok=True)
    label_dir.mkdir(parents=True, exist_ok=True)
    result_dir.mkdir(parents=True, exist_ok=True)

    image = np.zeros((16, 16, 3), dtype=np.uint8)
    image[:, :, 1] = 180
    label = np.zeros((16, 16, 3), dtype=np.uint8)
    label[4:12, 4:12, 2] = 255
    image_path = image_dir / "sample.png"
    label_path = label_dir / "sample.png"
    cv2.imwrite(str(image_path), image)
    cv2.imwrite(str(label_path), label)
    return image_path, label_path, result_dir


def run_model_family_readiness() -> dict[str, Any]:
    """Exercise the model-family runtime stack without launching full training."""
    source = settings.source_root
    yolo_weight = source / "Seg_All_In_One_YoloModel" / "yolo11n-seg.pt"
    mmseg_config = source / "Seg_All_In_One_MMSeg" / "configs" / "fcn" / "fcn_r18-d8_4xb2-80k_cityscapes-512x1024.py"
    mmseg_pretrained = source / "Seg_All_In_One_MMSeg" / "My_Local_Model" / "mmcls" / "resnet18.pth"

    checks = [
        {
            "name": "segmodel_smp_forward",
            "required": True,
            "detail": _run_snippet(
                "import torch, segmentation_models_pytorch as smp; "
                "m=smp.Unet(encoder_name='resnet18', encoder_weights=None, classes=2).eval(); "
                "torch.set_grad_enabled(False); y=m(torch.randn(1,3,64,64)); "
                "print(tuple(y.shape))"
            ),
        },
        {
            "name": "yolo_seg_predict_cpu",
            "required": True,
            "detail": _run_snippet(
                "from ultralytics import YOLO; import numpy as np; "
                f"model=YOLO({str(yolo_weight)!r}); "
                "r=model.predict(np.zeros((64,64,3), dtype=np.uint8), imgsz=64, verbose=False, save=False, device='cpu'); "
                "print(len(r), r[0].orig_shape)"
            ),
        },
        {
            "name": "mmseg_config_parse",
            "required": True,
            "detail": _run_snippet(
                "from mmengine.config import Config; "
                f"cfg=Config.fromfile({str(mmseg_config)!r}); "
                "print(cfg.model.type, cfg.train_dataloader.batch_size)"
            ),
        },
        {
            "name": "mmseg_local_pretrained_weight",
            "required": True,
            "detail": {"passed": mmseg_pretrained.exists(), "path": str(mmseg_pretrained), "size": mmseg_pretrained.stat().st_size if mmseg_pretrained.exists() else 0},
        },
        {
            "name": "mmseg_full_env_imports",
            "required": True,
            "detail": _run_conda_snippet(
                settings.mmseg_conda_env,
                "import torch, cv2, mmcv, mmengine, mmseg; "
                "import mmcv._ext; "
                "print(torch.__version__, torch.version.cuda, cv2.__version__, mmcv.__version__, mmseg.__version__)",
                timeout=90,
            ),
        },
        {
            "name": "mmseg_full_model_build",
            "required": True,
            "detail": _run_conda_snippet(
                settings.mmseg_conda_env,
                MMSEG_FULL_BUILD_SNIPPET.format(config_path=str(mmseg_config)),
                timeout=90,
            ),
        },
    ]
    for check in checks:
        check["passed"] = bool(check["detail"].get("passed"))
    return {
        "passed": all(item["passed"] for item in checks if item["required"]),
        "warnings": [item for item in checks if not item["required"] and not item["passed"]],
        "checks": checks,
    }


def latest_acceptance_report() -> dict[str, Any]:
    path = settings.project_root / "var" / "acceptance" / "latest.json"
    if not path.exists():
        return {"available": False, "path": str(path)}
    return json.loads(path.read_text(encoding="utf-8"))


def latest_deep_acceptance_report() -> dict[str, Any]:
    path = settings.project_root / "var" / "acceptance" / "deep_latest.json"
    if not path.exists():
        return {"available": False, "path": str(path)}
    return json.loads(path.read_text(encoding="utf-8"))


def run_deep_acceptance() -> dict[str, Any]:
    """Run minimal training loops for each model family without full datasets."""
    acceptance_root = settings.project_root / "var" / "acceptance"
    run_id = uuid.uuid4().hex[:8]
    fixture_root = acceptance_root / f"deep_{run_id}"
    fixture_root.mkdir(parents=True, exist_ok=True)

    yolo_weight = settings.source_root / "Seg_All_In_One_YoloModel" / "yolo11n-seg.pt"
    mmseg_config = settings.source_root / "Seg_All_In_One_MMSeg" / "configs" / "fcn" / "fcn_r18-d8_4xb2-80k_cityscapes-512x1024.py"

    yolo_root = fixture_root / "yolo_tiny"
    checks = [
        {
            "name": "segmodel_tiny_train_step",
            "passed": False,
            "detail": _run_snippet(SEGMODEL_TRAIN_STEP_SNIPPET, timeout=90),
        },
    ]
    yolo_train = {
        "name": "yolo_tiny_segment_train_epoch",
        "passed": False,
        "detail": _run_snippet(_yolo_tiny_train_snippet(yolo_root, yolo_weight), timeout=180),
    }
    checks.append(yolo_train)
    if yolo_train["detail"].get("passed"):
        checks.append(
            {
                "name": "yolo_tiny_heatmap_generation",
                "passed": False,
                "detail": _run_snippet(_yolo_heatmap_snippet(yolo_root), timeout=90),
            }
        )
    else:
        checks.append(
            {
                "name": "yolo_tiny_heatmap_generation",
                "passed": False,
                "detail": {"passed": False, "error": "skipped because yolo_tiny_segment_train_epoch failed"},
            }
        )
    checks.append(
        {
            "name": "mmseg_tiny_train_step",
            "passed": False,
            "detail": _run_conda_snippet(settings.mmseg_conda_env, _mmseg_train_step_snippet(mmseg_config), timeout=120),
        }
    )
    for check in checks:
        check["passed"] = bool(check["detail"].get("passed"))

    report = {
        "available": True,
        "run_id": run_id,
        "fixture_root": str(fixture_root),
        "passed": all(item["passed"] for item in checks),
        "checks": checks,
        "created_at": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
    }
    latest = acceptance_root / "deep_latest.json"
    latest.write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8")
    return report


def run_live_acceptance(base_url: str = "http://127.0.0.1:8010") -> dict[str, Any]:
    """Run a lightweight end-to-end smoke against the live API and job runner."""
    acceptance_root = settings.project_root / "var" / "acceptance"
    run_id = uuid.uuid4().hex[:8]
    fixture_root = acceptance_root / f"run_{run_id}"
    fixture_root.mkdir(parents=True, exist_ok=True)

    checks: list[dict[str, Any]] = []

    dataset_name = f"acceptance_{run_id}"
    created_dataset = _request_json("POST", f"{base_url}/api/datasets", {"name": dataset_name, "description": "acceptance smoke"}, timeout=10)
    checks.append({"name": "create_dataset_api", "passed": created_dataset.get("passed", False), "detail": created_dataset})

    upload = _post_multipart(
        f"{base_url}/api/datasets/{dataset_name}/upload/labels",
        "files",
        "label 01.txt",
        b"0 0.5 0.5 0.25 0.25\n",
        "text/plain",
        timeout=10,
    )
    checks.append({"name": "upload_label_api", "passed": upload.get("passed", False), "detail": upload})

    artifact_ok = False
    artifact_detail: dict[str, Any] = {"skipped": True}
    try:
        relative_path = upload["json"]["saved"][0]["relative_path"]
        artifact_detail = _request_text(f"{base_url}/api/artifacts/{relative_path}", timeout=10)
        artifact_ok = artifact_detail.get("passed", False) and "0 0.5" in artifact_detail.get("body", "")
    except Exception as exc:
        artifact_detail = {"error": str(exc)}
    checks.append({"name": "artifact_api_for_uploaded_label", "passed": artifact_ok, "detail": artifact_detail})

    mock = _create_job_and_wait(base_url, "mock.echo", {"message": f"acceptance {run_id}"}, timeout=45)
    mock_log = mock.get("polled", {}).get("job", {}).get("log_tail", "")
    checks.append({"name": "mock_job_runner", "passed": mock.get("passed", False) and f"acceptance {run_id}" in mock_log, "detail": mock})

    image_path, label_path, result_dir = _write_acceptance_images(fixture_root)
    stack = _create_job_with_retry(
        base_url,
        "dataset.stack_single",
        {
            "image_path": str(image_path),
            "label_path": str(label_path),
            "result_dir": str(result_dir),
            "alpha": 0.35,
        },
        attempts=2,
        timeout=90,
    )
    output_path = result_dir / "sample.png"
    checks.append(
        {
            "name": "legacy_stack_job_runner",
            "passed": stack.get("passed", False) and output_path.exists() and output_path.stat().st_size > 0,
            "detail": {**stack, "output_path": str(output_path), "output_exists": output_path.exists()},
        }
    )

    readiness = run_model_family_readiness()
    checks.append({"name": "model_family_readiness", "passed": readiness["passed"], "detail": readiness})

    report = {
        "available": True,
        "run_id": run_id,
        "base_url": base_url,
        "passed": all(item["passed"] for item in checks),
        "checks": checks,
        "model_family_readiness": readiness,
        "created_at": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
    }
    latest = acceptance_root / "latest.json"
    latest.write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8")
    return report