"""
subtitle_parser.py — Subtitle Parsing & Writing  (Step 3)
Handles .srt / .vtt / .ass via pysubs2.

Public API
----------
parse_subtitle(file_bytes, filename)  -> (SSAFile, fmt_str)
get_plain_lines(ssafile)              -> list[str]
replace_lines(ssafile, new_lines)     -> SSAFile
write_subtitle(ssafile, fmt_str)      -> bytes
clean_workspace()                     -> None
workspace_size_mb()                   -> float
"""

import copy
import logging
import os
import re
import shutil
import tempfile

import pysubs2

logger = logging.getLogger(__name__)

# ── Workspace dir ──────────────────────────────────────────────────────────
WORKSPACE_DIR = os.path.join(tempfile.gettempdir(), "subsync_workspace")
os.makedirs(WORKSPACE_DIR, exist_ok=True)

# ── Format map ─────────────────────────────────────────────────────────────
_EXT_MAP = {".srt": "srt", ".vtt": "vtt", ".ass": "ass", ".ssa": "ass"}


def _detect_format(filename: str) -> str:
    ext = os.path.splitext(filename.lower())[1]
    fmt = _EXT_MAP.get(ext)
    if not fmt:
        raise ValueError(
            f"'{ext}' format ကို support မလုပ်ပါ — .srt / .vtt / .ass သာ လက်ခံသည်"
        )
    return fmt


# ── Parse ──────────────────────────────────────────────────────────────────

def parse_subtitle(file_bytes: bytes, filename: str):
    """
    Parse subtitle bytes to pysubs2.SSAFile.
    Returns (SSAFile, fmt_string).
    Raises ValueError with user-friendly message on failure.
    """
    fmt = _detect_format(filename)

    # Decode with multiple encoding fallbacks; strip BOM
    text = None
    for enc in ("utf-8-sig", "utf-8", "latin-1", "cp1252"):
        try:
            text = file_bytes.decode(enc)
            break
        except (UnicodeDecodeError, LookupError):
            continue
    if text is None:
        raise ValueError("ဖိုင် encoding ဖတ်၍ မရပါ (utf-8 / latin-1 ကြိုးစားပြီး)")

    text = text.lstrip("\ufeff")   # strip BOM if still present

    # --- Strategy 1: from_string (no format_ arg — let pysubs2 auto-detect)
    try:
        subs = pysubs2.SSAFile.from_string(text)
        logger.info("Parsed %d events from '%s' (%s)", len(subs), filename, fmt)
        return subs, fmt
    except Exception as e1:
        logger.warning("from_string failed (%s): %s", filename, e1)

    # --- Strategy 2: write to tempfile and load by path
    try:
        suffix = os.path.splitext(filename)[1] or ("." + fmt)
        tmp_fd, tmp_path = tempfile.mkstemp(suffix=suffix)
        try:
            os.write(tmp_fd, file_bytes)
        finally:
            os.close(tmp_fd)

        subs = pysubs2.SSAFile.load(tmp_path)
        os.unlink(tmp_path)
        logger.info("Parsed via tempfile: %d events", len(subs))
        return subs, fmt
    except Exception as e2:
        logger.error("tempfile parse also failed: %s", e2)
        raise ValueError(
            f"Subtitle ဖိုင် ဖတ်၍ မရပါ။\n"
            f"Attempt 1: {e1}\n"
            f"Attempt 2: {e2}"
        ) from e2


# ── Tag stripping ──────────────────────────────────────────────────────────
_ASS_TAG_RE  = re.compile(r"\{[^}]*\}")
_HTML_TAG_RE = re.compile(r"<[^>]+>")


def _strip_tags(text: str) -> str:
    text = _ASS_TAG_RE.sub("", text)
    text = _HTML_TAG_RE.sub("", text)
    return text.strip()


def get_plain_lines(ssafile) -> list:
    """
    Return list[str] same length as ssafile.events.
    Tags stripped; \\N/\\n normalised to Python newline.
    """
    lines = []
    for event in ssafile.events:
        raw   = event.text.replace("\\N", "\n").replace("\\n", "\n")
        clean = _strip_tags(raw)
        lines.append(clean)
    return lines


# ── Replace lines (timing-safe deep copy) ──────────────────────────────────

def replace_lines(ssafile, new_lines: list):
    """
    Return new SSAFile with translated text injected index-for-index.

    TIMESTAMP SAFETY:
    - event.start / event.end are NEVER touched (deep-copy preserves them)
    - Only event.text is replaced
    - If len(new_lines) != len(events): raise immediately — never partial
    - Empty/whitespace translation → keep original text (safe fallback)
    - No re-ordering, no shifting, index i always maps to same timestamp
    """
    n_events = len(ssafile.events)
    n_lines  = len(new_lines)

    if n_lines != n_events:
        raise ValueError(
            f"TIMESTAMP SAFETY: cannot replace — "
            f"translated lines ({n_lines}) != subtitle events ({n_events}). "
            f"Every line must have a matching translation."
        )

    result = copy.deepcopy(ssafile)   # timestamps copied intact

    fallback_count = 0
    for i, event in enumerate(result.events):
        translated = str(new_lines[i]).strip()
        if not translated:
            # Keep original — do NOT leave a blank event (timing stays correct)
            fallback_count += 1
            continue
        # Restore multi-line joins: Python \n → ASS \N
        event.text = translated.replace("\n", "\\N")
        # event.start and event.end are untouched

    if fallback_count:
        logger.warning(
            "replace_lines: %d/%d events kept original (empty translation)",
            fallback_count, n_events
        )

    return result


# ── Write ──────────────────────────────────────────────────────────────────

def write_subtitle(ssafile, fmt: str) -> bytes:
    """Serialise SSAFile to bytes in original format (UTF-8)."""
    try:
        return ssafile.to_string(fmt).encode("utf-8")
    except Exception as exc:
        raise RuntimeError(f"Subtitle ဖိုင် write မရပါ: {exc}") from exc


def output_filename(original_name: str) -> str:
    """'movie.en.srt' → 'movie.en.my.srt'"""
    base, ext = os.path.splitext(original_name)
    return f"{base}.my{ext}"


# ── Workspace management ───────────────────────────────────────────────────

def clean_workspace() -> None:
    """Delete all workspace files (called by Clean Workspace button)."""
    try:
        if os.path.exists(WORKSPACE_DIR):
            shutil.rmtree(WORKSPACE_DIR)
        os.makedirs(WORKSPACE_DIR, exist_ok=True)
    except Exception as exc:
        raise RuntimeError(f"Workspace ရှင်းမရပါ: {exc}") from exc


def workspace_size_mb() -> float:
    total = 0
    for dirpath, _, filenames in os.walk(WORKSPACE_DIR):
        for f in filenames:
            try:
                total += os.path.getsize(os.path.join(dirpath, f))
            except OSError:
                pass
    return round(total / (1024 * 1024), 2)