alfred/alfred/application/filesystem/manage_subtitles.py

"""ManageSubtitlesUseCase — orchestrates the full subtitle pipeline for a video file."""

import logging
from pathlib import Path

from alfred.domain.shared.value_objects import ImdbId
from alfred.domain.subtitles.entities import SubtitleCandidate
from alfred.domain.subtitles.services.identifier import SubtitleIdentifier
from alfred.domain.subtitles.services.matcher import SubtitleMatcher
from alfred.domain.subtitles.services.pattern_detector import PatternDetector
from alfred.application.subtitles.placer import (
    PlacedTrack,
    SubtitlePlacer,
    _build_dest_name,
)
from alfred.domain.subtitles.services.utils import available_subtitles
from alfred.domain.subtitles.value_objects import ScanStrategy
from alfred.infrastructure.filesystem.scanner import PathlibFilesystemScanner
from alfred.infrastructure.knowledge.subtitles.base import SubtitleKnowledgeBase
from alfred.infrastructure.knowledge.subtitles.loader import KnowledgeLoader
from alfred.infrastructure.persistence.context import get_memory
from alfred.infrastructure.probe.ffprobe_prober import FfprobeMediaProber
from alfred.infrastructure.subtitle.metadata_store import SubtitleMetadataStore
from alfred.infrastructure.subtitle.rule_repository import RuleSetRepository

from .dto import (
    AvailableSubtitle,
    ManageSubtitlesResponse,
    PlacedSubtitle,
    UnresolvedTrack,
)

logger = logging.getLogger(__name__)


def _infer_library_root(dest_video: Path, media_type: str) -> Path:
    """
    Infer the media library root folder from the destination video path.

    TV show:  video → Season 01 → The X-Files   (3 levels up)
    Movie:    video → Inception (2010)           (1 level up)
    """
    if media_type == "tv_show":
        return dest_video.parent.parent
    return dest_video.parent


def _to_imdb_id(raw: str | None) -> ImdbId | None:
    if not raw:
        return None
    try:
        return ImdbId(raw)
    except Exception:
        return None


class ManageSubtitlesUseCase:
    """
    Full subtitle pipeline:

    1. Load knowledge base
    2. Detect (or confirm) the release pattern
    3. Identify all tracks (ffprobe + filesystem scan)
    4. Load + resolve rules for this media
    5. Match tracks against rules
    6. If any tracks are unresolved → return needs_clarification (don't place yet)
    7. Place matched tracks via hard-link
    8. Persist to .alfred/metadata.yaml

    The use case is stateless — all dependencies are instantiated inline.
    """

    def execute(
        self,
        source_video: str,
        destination_video: str,
        imdb_id: str | None = None,
        media_type: str = "tv_show",
        release_group: str | None = None,
        season: int | None = None,
        episode: int | None = None,
        confirmed_pattern_id: str | None = None,
        dry_run: bool = False,
    ) -> ManageSubtitlesResponse:
        source_path = Path(source_video)
        dest_path = Path(destination_video)

        if not source_path.exists() and not source_path.parent.exists():
            return ManageSubtitlesResponse(
                status="error",
                error="source_not_found",
                message=f"Source video not found: {source_video}",
            )

        kb = SubtitleKnowledgeBase(KnowledgeLoader())
        prober = FfprobeMediaProber()
        scanner = PathlibFilesystemScanner()
        library_root = _infer_library_root(dest_path, media_type)
        store = SubtitleMetadataStore(library_root)
        repo = RuleSetRepository(library_root)

        # --- Pattern resolution ---
        pattern = self._resolve_pattern(
            kb,
            prober,
            scanner,
            store,
            source_path,
            confirmed_pattern_id,
            release_group,
        )
        if pattern is None:
            return ManageSubtitlesResponse(
                status="error",
                error="pattern_not_found",
                message="Could not determine subtitle pattern for this release.",
            )

        # --- Identify ---
        media_id = _to_imdb_id(imdb_id)
        identifier = SubtitleIdentifier(kb, prober, scanner)
        metadata = identifier.identify(
            video_path=source_path,
            pattern=pattern,
            media_id=media_id,
            media_type=media_type,
            release_group=release_group,
        )

        if metadata.total_count == 0:
            logger.info(
                f"ManageSubtitles: no subtitle tracks found for {source_path.name}"
            )
            return ManageSubtitlesResponse(
                status="ok",
                video_path=destination_video,
                placed=[],
                skipped_count=0,
            )

        # --- Embedded short-circuit ---
        if pattern.scan_strategy == ScanStrategy.EMBEDDED:
            logger.info("ManageSubtitles: embedded pattern — skipping matcher")
            available = [
                AvailableSubtitle(
                    language=t.language.code if t.language else "?",
                    subtitle_type=t.subtitle_type.value,
                )
                for t in available_subtitles(metadata.embedded_tracks)
            ]
            return ManageSubtitlesResponse(
                status="ok",
                video_path=destination_video,
                placed=[],
                skipped_count=0,
                available=available,
            )

        # --- Match (external only) ---
        subtitle_prefs = None
        try:
            memory = get_memory()
            subtitle_prefs = memory.ltm.subtitle_preferences
        except Exception:
            pass
        rules = repo.load(release_group, subtitle_prefs).resolve(kb.default_rules())
        matcher = SubtitleMatcher()
        matched, unresolved = matcher.match(metadata.external_tracks, rules)

        if unresolved:
            logger.info(
                f"ManageSubtitles: {len(unresolved)} unresolved track(s) — needs clarification"
            )
            return ManageSubtitlesResponse(
                status="needs_clarification",
                video_path=destination_video,
                placed=[],
                unresolved=[_to_unresolved_dto(t) for t in unresolved],
            )

        if not matched:
            return ManageSubtitlesResponse(
                status="ok",
                video_path=destination_video,
                placed=[],
                skipped_count=metadata.total_count,
            )

        # --- Dry run: skip placement ---
        if dry_run:
            placed_dtos = []
            for t in matched:
                if not t.file_path:
                    continue
                try:
                    filename = _build_dest_name(t, dest_path.stem)
                except ValueError:
                    continue
                placed_dtos.append(
                    PlacedSubtitle(
                        source=str(t.file_path),
                        destination=str(dest_path.parent / filename),
                        filename=filename,
                    )
                )
            return ManageSubtitlesResponse(
                status="ok",
                video_path=destination_video,
                placed=placed_dtos,
                skipped_count=0,
            )

        # --- Place ---
        placer = SubtitlePlacer()
        place_result = placer.place(matched, dest_path)

        # --- Persist ---
        if place_result.placed:
            pairs = _pair_placed_with_tracks(place_result.placed, matched)
            store.append_history(pairs, season, episode, release_group)

        placed_dtos = [
            PlacedSubtitle(
                source=str(p.source),
                destination=str(p.destination),
                filename=p.filename,
            )
            for p in place_result.placed
        ]

        return ManageSubtitlesResponse(
            status="ok",
            video_path=destination_video,
            placed=placed_dtos,
            skipped_count=place_result.skipped_count,
        )

    def _resolve_pattern(
        self,
        kb: SubtitleKnowledgeBase,
        prober: FfprobeMediaProber,
        scanner: PathlibFilesystemScanner,
        store: SubtitleMetadataStore,
        source_path: Path,
        confirmed_pattern_id: str | None,
        release_group: str | None,
    ):
        # 1. Explicit override from caller
        if confirmed_pattern_id:
            p = kb.pattern(confirmed_pattern_id)
            if p:
                return p
            logger.warning(f"ManageSubtitles: unknown pattern '{confirmed_pattern_id}'")

        # 2. Previously confirmed in metadata store
        stored_id = store.confirmed_pattern()
        if stored_id:
            p = kb.pattern(stored_id)
            if p:
                logger.debug(f"ManageSubtitles: using confirmed pattern '{stored_id}'")
                return p

        # 3. Auto-detect
        release_root = source_path.parent
        detector = PatternDetector(kb, prober, scanner)
        result = detector.detect(release_root, source_path)

        if result["detected"] and result["confidence"] >= 0.6:
            logger.info(
                f"ManageSubtitles: auto-detected pattern '{result['detected'].id}' "
                f"(confidence={result['confidence']:.2f})"
            )
            return result["detected"]

        # 4. Fallback — adjacent (safest default)
        logger.info("ManageSubtitles: falling back to 'adjacent' pattern")
        return kb.pattern("adjacent")


def _to_unresolved_dto(
    track: SubtitleCandidate, min_confidence: float = 0.7
) -> UnresolvedTrack:
    reason = "unknown_language" if track.language is None else "low_confidence"
    return UnresolvedTrack(
        raw_tokens=track.raw_tokens,
        file_path=str(track.file_path) if track.file_path else None,
        file_size_kb=track.file_size_kb,
        reason=reason,
    )


def _pair_placed_with_tracks(
    placed: list[PlacedTrack],
    tracks: list[SubtitleCandidate],
) -> list[tuple[PlacedTrack, SubtitleCandidate]]:
    """
    Pair each PlacedTrack with its originating SubtitleCandidate by source path.
    Falls back to positional matching if paths don't align.
    """
    track_by_path = {t.file_path: t for t in tracks if t.file_path}
    pairs = []
    for p in placed:
        track = track_by_path.get(p.source)
        if track is None and tracks:
            track = tracks[0]  # positional fallback
        if track:
            pairs.append((p, track))
    return pairs