import csv
import io
import pandas as pd
from datetime import datetime, timedelta

CANONICAL_COLS = ["idx", "speed", "std_dev", "energy", "power_factor", "time"]
TIME_FORMATS = ["%H:%M:%S.%f", "%H:%M:%S", "%H:%M:%S,%f"]


def parse_csv(stream) -> pd.DataFrame:
    raw = stream.read()
    if isinstance(raw, bytes):
        raw = raw.decode("utf-8-sig")
    # Strip BOM characters that may appear anywhere in the file
    raw = raw.replace("\ufeff", "")

    data_rows = []
    for line in raw.splitlines():
        fields = _split_line(line)
        if len(fields) >= 6 and _is_index(fields[0]) and _is_time(fields[5]):
            data_rows.append(fields[:6])

    if len(data_rows) < 2:
        raise ValueError(
            "Could not find valid data rows in the CSV. "
            "Expected rows with: integer index, 4 numeric values, and a time (HH:MM:SS)."
        )

    df = pd.DataFrame(data_rows, columns=CANONICAL_COLS)

    for col in ("speed", "std_dev", "energy", "power_factor"):
        df[col] = _parse_numeric(df[col])

    df["time"] = _parse_time_column(df["time"])
    df = df.sort_values("time").reset_index(drop=True)
    return df[["speed", "std_dev", "energy", "power_factor", "time"]]


def _split_line(line: str) -> list:
    """Parse one CSV line, respecting quoted fields."""
    for row in csv.reader([line], quotechar='"', doublequote=True, skipinitialspace=True):
        return [f.strip() for f in row]
    return []


def _is_index(val: str) -> bool:
    """True if the value is a non-negative integer (auto-increment row index)."""
    try:
        return int(val.strip()) >= 0
    except (ValueError, AttributeError):
        return False


def _is_time(val: str) -> bool:
    """True if the value parses as HH:MM:SS or HH:MM:SS.fff."""
    cleaned = val.strip()
    for fmt in TIME_FORMATS:
        try:
            datetime.strptime(cleaned, fmt)
            return True
        except ValueError:
            continue
    return False


def _parse_numeric(col: pd.Series) -> pd.Series:
    """Parse a numeric column, accepting both '.' and ',' as decimal separator."""
    result = pd.to_numeric(col, errors="coerce")
    if result.isna().any():
        result = pd.to_numeric(
            col.astype(str).str.replace(",", ".", regex=False),
            errors="coerce",
        )
    if result.isna().any():
        bad = col[result.isna()].tolist()
        raise ValueError(f"Non-numeric values in column: {bad}")
    return result


def _parse_time_column(col: pd.Series) -> pd.Series:
    today = datetime.today().date()
    cleaned = col.astype(str).str.strip()

    parsed = None
    for fmt in TIME_FORMATS:
        candidate = pd.to_datetime(cleaned, format=fmt, errors="coerce")
        if candidate.notna().all():
            parsed = candidate
            break

    if parsed is None:
        candidate = pd.to_datetime(cleaned, errors="coerce")
        if candidate.notna().all():
            parsed = candidate

    if parsed is None:
        raise ValueError(
            "Could not parse time column. Expected format: HH:MM:SS or HH:MM:SS.fff"
        )

    parsed = parsed.apply(lambda t: datetime.combine(today, t.time()))

    times = parsed.tolist()
    for i in range(1, len(times)):
        if times[i] < times[i - 1]:
            times[i] += timedelta(days=1)
    return pd.Series(times, index=col.index)