"""
Shared extraction logic for VictoriaMetrics pipelines (main + backfill).
"""

import hashlib
import json
import logging
import time
from datetime import datetime, timezone
from typing import Iterator

import requests

logger = logging.getLogger(__name__)

RETRY_ATTEMPTS = 3
RETRY_BACKOFF = 2

DATASET_NAME = "vm_metrics"


def labels_hash(metric: dict) -> str:
    """Stable hash of metric labels (excluding __name__) for deduplication."""
    labels = {k: v for k, v in sorted(metric.items()) if k != "__name__"}
    return hashlib.md5(json.dumps(labels, sort_keys=True).encode()).hexdigest()


def fetch_query_range(
    base_url: str,
    query: str,
    start: int,
    end: int,
    step: int,
) -> list[dict]:
    """POST to /api/v1/query_range with retries. Returns the result array."""
    url = base_url.rstrip("/") + "/query_range"
    params = {"query": query, "start": start, "end": end, "step": step}

    for attempt in range(1, RETRY_ATTEMPTS + 1):
        try:
            resp = requests.post(url, data=params, timeout=300)
            resp.raise_for_status()
            data = resp.json()

            if data.get("status") != "success":
                raise RuntimeError(
                    f"query_range error: {data.get('errorType', '?')}: "
                    f"{data.get('error', 'unknown')}"
                )

            return data["data"]["result"]
        except requests.RequestException as e:
            if attempt == RETRY_ATTEMPTS:
                raise
            wait = RETRY_BACKOFF ** attempt
            logger.warning(
                "Attempt %d/%d failed for %s: %s. Retrying in %ds...",
                attempt, RETRY_ATTEMPTS, url, e, wait,
            )
            time.sleep(wait)


def parse_query_range_result(
    results: list[dict],
    kubernetes: str,
    provider: str,
    cloud_region: str,
    keep_labels: list[str] | None = None,
) -> Iterator[dict]:
    """Flatten query_range result into one dict per (series, timestamp) pair.

    If keep_labels is provided, only those metric labels are retained
    (strips extra labels that VictoriaMetrics may preserve on no-op aggregations).
    """
    for series in results:
        raw_metric = series.get("metric", {})
        if keep_labels is not None:
            metric = {k: v for k, v in raw_metric.items() if k in keep_labels}
        else:
            metric = {k: v for k, v in raw_metric.items() if k != "__name__"}
        lhash = labels_hash(metric)

        for ts, value_str in series.get("values", []):
            ts_int = int(ts)
            row = dict(metric)
            row["metric_timestamp"] = ts_int
            row["metric_value"] = float(value_str)
            row["metric_labels_hash"] = lhash
            row["metric_date"] = datetime.fromtimestamp(ts_int, tz=timezone.utc).date()
            row["kubernetes"] = kubernetes
            row["provider"] = provider
            row["cloud_region"] = cloud_region
            yield row