Medios-Macina/cmdlet/add_file.py

from __future__ import annotations

from typing import Any, Dict, Optional, Sequence, Tuple, List
from pathlib import Path
import sys
import shutil
import tempfile

import models
import pipeline as ctx
from API import HydrusNetwork as hydrus_wrapper
from SYS.logger import log, debug
from SYS.utils_constant import ALL_SUPPORTED_EXTENSIONS
from Store import Store
from ._shared import (
    Cmdlet, CmdletArg, parse_cmdlet_args, SharedArgs,
    extract_tag_from_result, extract_title_from_result, extract_url_from_result,
    merge_sequences, extract_relationships, extract_duration, coerce_to_pipe_object
)
from ._shared import collapse_namespace_tag
from API.folder import read_sidecar, find_sidecar, write_sidecar, API_folder_store
from SYS.utils import sha256_file, unique_path
from metadata import write_metadata

# Canonical supported filetypes for all stores/cmdlets
SUPPORTED_MEDIA_EXTENSIONS = ALL_SUPPORTED_EXTENSIONS

class Add_File(Cmdlet):
    """Add file into the DB"""

    def __init__(self) -> None:
        """Initialize add-file cmdlet."""
        super().__init__(
            name="add-file",
            summary="Upload a media file to specified location (Hydrus, file provider, or local directory).",
            usage="add-file (-path <filepath> | <piped>) (-storage <location> | -provider <fileprovider>) [-delete]",
            arg=[
                SharedArgs.PATH,
                SharedArgs.STORE,
                SharedArgs.HASH,
                CmdletArg(name="provider", type="string", required=False, description="File hosting provider (e.g., 0x0)", alias="prov"),
                CmdletArg(name="delete", type="flag", required=False, description="Delete file after successful upload", alias="del"),
            ],
            detail=[
                "- Storage location options (use -storage):",
                "    hydrus: Upload to Hydrus database with metadata tagging",
                "    local: Copy file to local directory",
                "    <path>: Copy file to specified directory",
                "- File provider options (use -provider):",
                "    0x0: Upload to 0x0.st for temporary hosting",
            ],
            exec=self.run,
        )
        self.register()

    def run(self, result: Any, args: Sequence[str], config: Dict[str, Any]) -> int:
        """Main execution entry point."""
        parsed = parse_cmdlet_args(args, self)

        path_arg = parsed.get("path")
        location = parsed.get("store")
        provider_name = parsed.get("provider")
        delete_after = parsed.get("delete", False)

        stage_ctx = ctx.get_stage_context()
        is_last_stage = (stage_ctx is None) or bool(getattr(stage_ctx, "is_last_stage", False))

        # Decide which items to process.
        # - If user provided -path, treat this invocation as single-item.
        # - Otherwise, if piped input is a list, ingest each item.
        if path_arg:
            items_to_process: List[Any] = [result]
        elif isinstance(result, list) and result:
            items_to_process = list(result)
        else:
            items_to_process = [result]

        debug(f"[add-file] INPUT result type={type(result).__name__}")
        if isinstance(result, list):
            debug(f"[add-file] INPUT result is list with {len(result)} items")
        debug(f"[add-file] PARSED args: location={location}, provider={provider_name}, delete={delete_after}")

        collected_payloads: List[Dict[str, Any]] = []
        successes = 0
        failures = 0

        # Only run the search-store refresh when add-file is the last stage.
        # In the middle of a pipeline, downstream cmdlets should receive the emitted
        # storage payload directly (no need to re-search and risk duplicate emits).
        auto_search_store_after_add = bool(is_last_stage) and len(items_to_process) == 1

        for item in items_to_process:
            pipe_obj = coerce_to_pipe_object(item, path_arg)

            temp_dir_to_cleanup: Optional[Path] = None
            delete_after_item = delete_after
            try:
                media_path_or_url, file_hash = self._resolve_source(item, path_arg, pipe_obj, config)
                debug(f"[add-file] RESOLVED source: path={media_path_or_url}, hash={file_hash[:12] if file_hash else 'N/A'}...")
                if not media_path_or_url:
                    failures += 1
                    continue

                # Update pipe_obj with resolved path
                pipe_obj.path = str(media_path_or_url)

                # URL targets: prefer provider-aware download for OpenLibrary selections.
                if isinstance(media_path_or_url, str) and media_path_or_url.lower().startswith(
                    ("http://", "https://", "magnet:", "torrent:")
                ):
                    table = None
                    full_metadata = None
                    if isinstance(pipe_obj.extra, dict):
                        table = pipe_obj.extra.get("table")
                        full_metadata = pipe_obj.extra.get("full_metadata")

                    is_openlibrary = (str(table or "").lower() == "openlibrary") or ("openlibrary.org/books/" in media_path_or_url.lower())
                    if is_openlibrary:
                        # Enrich tags from OpenLibrary metadata so the stored file has book tags (author/pages/etc).
                        try:
                            from Provider.openlibrary import OpenLibrary as _OpenLibrary

                            olid = None
                            archive_id = None
                            if isinstance(full_metadata, dict):
                                olid = full_metadata.get("openlibrary_id") or full_metadata.get("openlibrary")
                                archive_id = full_metadata.get("archive_id")

                            if not olid:
                                import re
                                m = re.search(r"/books/(OL\d+M)", str(media_path_or_url), flags=re.IGNORECASE)
                                if m:
                                    olid = m.group(1)

                            scraped_tags: List[str] = []
                            if olid:
                                scraped_tags.extend(_OpenLibrary.scrape_openlibrary_metadata(str(olid)) or [])
                            if archive_id:
                                scraped_tags.append(f"internet_archive:{archive_id}")

                            if scraped_tags:
                                existing = list(pipe_obj.tag or [])
                                pipe_obj.tag = merge_sequences(existing, scraped_tags, case_sensitive=False)
                        except Exception:
                            pass

                        from ProviderCore.registry import get_search_provider
                        from ProviderCore.base import SearchResult

                        provider = get_search_provider("openlibrary", config)
                        if provider is None:
                            log("[add-file] OpenLibrary provider not available", file=sys.stderr)
                            failures += 1
                            continue

                        temp_dir_to_cleanup = Path(tempfile.mkdtemp(prefix="medios_openlibrary_"))
                        sr = SearchResult(
                            table="openlibrary",
                            title=str(getattr(pipe_obj, "title", None) or "Unknown"),
                            path=str(media_path_or_url),
                            full_metadata=full_metadata if isinstance(full_metadata, dict) else {},
                        )
                        downloaded = provider.download(sr, temp_dir_to_cleanup)
                        if downloaded is None:
                            log("[add-file] OpenLibrary download failed", file=sys.stderr)
                            failures += 1
                            continue

                        downloaded_path = Path(downloaded)
                        if downloaded_path.exists() and downloaded_path.is_dir():
                            log(
                                "[add-file] OpenLibrary download produced a directory (missing img2pdf?). Cannot ingest.",
                                file=sys.stderr,
                            )
                            failures += 1
                            continue

                        media_path_or_url = str(downloaded_path)
                        pipe_obj.path = str(downloaded_path)
                        delete_after_item = True

                    # For non-provider URLs, or if still a URL after provider attempt, delegate to download-media.
                    if isinstance(media_path_or_url, str) and media_path_or_url.lower().startswith(
                        ("http://", "https://", "magnet:", "torrent:")
                    ):
                        code = self._delegate_to_download_data(item, media_path_or_url, location, provider_name, args, config)
                        if code == 0:
                            successes += 1
                        else:
                            failures += 1
                        continue

                media_path = Path(media_path_or_url) if isinstance(media_path_or_url, str) else media_path_or_url

                if not self._validate_source(media_path):
                    failures += 1
                    continue

                if provider_name:
                    code = self._handle_provider_upload(media_path, provider_name, pipe_obj, config, delete_after_item)
                    if code == 0:
                        successes += 1
                    else:
                        failures += 1
                    continue

                if location:
                    try:
                        store = Store(config)
                        backends = store.list_backends()
                        if location in backends:
                            code = self._handle_storage_backend(
                                item,
                                media_path,
                                location,
                                pipe_obj,
                                config,
                                delete_after_item,
                                collect_payloads=collected_payloads,
                                suppress_last_stage_overlay=is_last_stage and len(items_to_process) > 1,
                                auto_search_store=auto_search_store_after_add,
                            )
                        else:
                            code = self._handle_local_export(media_path, location, pipe_obj, config, delete_after_item)
                    except Exception as exc:
                        debug(f"[add-file] ERROR: Failed to resolve location: {exc}")
                        log(f"Invalid location: {location}", file=sys.stderr)
                        failures += 1
                        continue

                    if code == 0:
                        successes += 1
                    else:
                        failures += 1
                    continue

                log("No destination specified", file=sys.stderr)
                failures += 1
            finally:
                if temp_dir_to_cleanup is not None:
                    try:
                        shutil.rmtree(temp_dir_to_cleanup, ignore_errors=True)
                    except Exception:
                        pass

        # If we processed multiple storage ingests, present a single consolidated overlay table.
        if is_last_stage and len(items_to_process) > 1 and collected_payloads:
            try:
                from result_table import ResultTable

                table = ResultTable("Result")
                for payload in collected_payloads:
                    table.add_result(payload)
                # Make this the active selectable table so @.. returns here (and playlist table is kept in history).
                ctx.set_last_result_table(table, collected_payloads, subject=collected_payloads)
            except Exception:
                pass

        if successes > 0:
            return 0
        return 1

    @staticmethod
    def _resolve_source(
        result: Any,
        path_arg: Optional[str],
        pipe_obj: models.PipeObject,
        config: Dict[str, Any],
    ) -> Tuple[Optional[Path | str], Optional[str]]:
        """Resolve the source file path from args or pipeline result.

        Returns (media_path_or_url, file_hash)
        where media_path_or_url can be a Path object or a URL string.
        """
        # PRIORITY 1: Try hash+store from result dict (most reliable for @N selections)
        if isinstance(result, dict):
            result_hash = result.get("hash")
            result_store = result.get("store")
            if result_hash and result_store:
                debug(
                    f"[add-file] Using hash+store from result: hash={str(result_hash)[:12]}..., store={result_store}"
                )
                try:
                    store = Store(config)
                    if result_store in store.list_backends():
                        backend = store[result_store]
                        media_path = backend.get_file(result_hash)
                        if isinstance(media_path, Path) and media_path.exists():
                            pipe_obj.path = str(media_path)
                            return media_path, str(result_hash)
                        if isinstance(media_path, str) and media_path.lower().startswith(
                            ("http://", "https://", "magnet:", "torrent:")
                        ):
                            pipe_obj.path = media_path
                            return media_path, str(result_hash)
                except Exception as exc:
                    debug(f"[add-file] Failed to retrieve via hash+store: {exc}")

        # PRIORITY 2: Try explicit path argument
        if path_arg:
            media_path = Path(path_arg)
            pipe_obj.path = str(media_path)
            debug(f"[add-file] Using explicit path argument: {media_path}")
            return media_path, None

        # PRIORITY 3: Try from pipe_obj.path (check file first before URL)
        pipe_path = getattr(pipe_obj, "path", None)
        if pipe_path:
            pipe_path_str = str(pipe_path)
            debug(f"Resolved pipe_path: {pipe_path_str}")
            if pipe_path_str.startswith("hydrus:"):
                file_hash = pipe_path_str.split(":", 1)[1]
                media_path, success = Add_File._fetch_hydrus_path(file_hash, config)
                return media_path, file_hash if success else None
            if pipe_path_str.lower().startswith(("http://", "https://", "magnet:", "torrent:")):
                return pipe_path_str, None
            return Path(pipe_path_str), None
        
        # PRIORITY 4: Try from pipe_obj.url (for streaming url without downloaded file)
        pipe_url = getattr(pipe_obj, "url", None)
        if pipe_url and isinstance(pipe_url, str):
            # Check if it's a URL
            if pipe_url.lower().startswith(("http://", "https://", "magnet:", "torrent:")):
                debug(f"Detected URL in pipe_obj.url: {pipe_url}")
                return pipe_url, None

        # Try from hydrus hash in pipe_obj.extra or hash
        hydrus_hash = None
        if isinstance(pipe_obj.extra, dict):
            hydrus_hash = pipe_obj.extra.get("hydrus_hash") or pipe_obj.extra.get("hash")
        hydrus_hash = hydrus_hash or pipe_obj.hash

        if hydrus_hash and hydrus_hash != "unknown":
            media_path, success = Add_File._fetch_hydrus_path(str(hydrus_hash), config)
            return media_path, str(hydrus_hash) if success else None

        # Try from result (if it's a string path or URL)
        if isinstance(result, str):
            debug(f"Checking result string: {result}")
            # Check if result is a URL before treating as file path
            if result.lower().startswith(("http://", "https://", "magnet:", "torrent:")):
                debug(f"Detected URL in result string: {result}")
                return result, None  # Return URL string directly
            media_path = Path(result)
            pipe_obj.path = str(media_path)
            return media_path, None

        # Try from result if it's a list (pipeline emits multiple results)
        if isinstance(result, list) and result:
            first_item = result[0]
            # If the first item is a string, it's either a URL or a file path
            if isinstance(first_item, str):
                debug(f"Checking result list[0]: {first_item}")
                if first_item.lower().startswith(("http://", "https://", "magnet:", "torrent:")):
                    debug(f"Detected URL in result list: {first_item}")
                    return first_item, None  # Return URL string directly
                media_path = Path(first_item)
                pipe_obj.path = str(media_path)
                return media_path, None

            # If the first item is a dict, interpret it as a PipeObject-style result
            if isinstance(first_item, dict):
                # Look for path or path-like keys
                path_candidate = first_item.get("path") or first_item.get("filepath") or first_item.get("file")
                # If the dict includes a 'paths' list (multi-part/section download), prefer the first file
                paths_val = first_item.get("paths")
                if not path_candidate and isinstance(paths_val, (list, tuple)) and paths_val:
                    path_candidate = paths_val[0]
                if path_candidate:
                    debug(f"Resolved path from result dict: {path_candidate}")
                    try:
                        media_path = Path(path_candidate)
                        pipe_obj.path = str(media_path)
                        return media_path, first_item.get("hash")
                    except Exception:
                        # Fallback to returning string if not a path
                        return str(path_candidate), first_item.get("hash")

            # If first item is a PipeObject object
            try:
                # models.PipeObject is an actual class; check attribute presence
                import models as _models
                if isinstance(first_item, _models.PipeObject):
                    path_candidate = getattr(first_item, "path", None)
                    if path_candidate:
                        debug(f"Resolved path from PipeObject: {path_candidate}")
                        media_path = Path(path_candidate)
                        pipe_obj.path = str(media_path)
                        return media_path, getattr(first_item, "hash", None)
            except Exception:
                pass

        debug(f"No resolution path matched. pipe_obj.path={pipe_path}, result type={type(result).__name__}")
        log("File path could not be resolved")
        return None, None

    @staticmethod
    def _fetch_hydrus_path(file_hash: str, config: Dict[str, Any]) -> Tuple[Optional[Path], bool]:
        """Fetch the physical path of a file from Hydrus using its hash."""
        if not file_hash:
            return None, False

        try:
            client = hydrus_wrapper.get_client(config)
            if not client:
                log("❌ Hydrus client not available", file=sys.stderr)
                return None, False

            response = client.get_file_path(file_hash)
            file_path_str = response.get("path")
            if not file_path_str:
                log(f"❌ Hydrus file_path endpoint did not return a path", file=sys.stderr)
                return None, False

            media_path = Path(file_path_str)
            if not media_path.exists():
                log(f"❌ Hydrus file path does not exist: {media_path}", file=sys.stderr)
                return None, False

            log(f"✓ Retrieved Hydrus file path: {media_path}", file=sys.stderr)
            return media_path, True
        except Exception as exc:
            log(f"❌ Failed to get Hydrus file path: {exc}", file=sys.stderr)
            return None, False

    @staticmethod
    def _validate_source(media_path: Optional[Path]) -> bool:
        """Validate that the source file exists and is supported."""
        if media_path is None:
            return False

        target_str = str(media_path)

        # If it's a URL target, we skip file existence checks
        if target_str.lower().startswith(("http://", "https://", "magnet:", "torrent:")):
            return True

        if not media_path.exists() or not media_path.is_file():
            log(f"File not found: {media_path}")
            return False

        # Validate file type
        file_extension = media_path.suffix.lower()
        if file_extension not in SUPPORTED_MEDIA_EXTENSIONS:
            log(f"❌ Unsupported file type: {file_extension}", file=sys.stderr)
            return False

        return True

    @staticmethod
    def _is_url_target(media_path: Optional[Path]) -> bool:
        """Check if the target is a URL that needs downloading."""
        if media_path and str(media_path).lower().startswith(("http://", "https://")):
            return True
        return False

    def _delegate_to_download_data(
        self,
        result: Any,
        url_str: str,
        location: Optional[str],
        provider_name: Optional[str],
        args: Sequence[str],
        config: Dict[str, Any],
    ) -> int:
        """Delegate URL handling to download-media cmdlet."""
        log(f"Target is a URL, delegating to download-media: {url_str}", file=sys.stderr)
        # Reuse the globally-registered cmdlet instance to avoid duplicative registration
        from cmdlet.download_media import CMDLET as dl_cmdlet
        dl_args = list(args) if args else []

        # Add the URL to the argument list for download-media
        dl_args.insert(0, url_str)
        
        # If result has selection_args (like -item from @N selection), include them
        if isinstance(result, dict) and "_selection_args" in result:
            selection_args = result["_selection_args"]
            if selection_args:
                dl_args.extend(selection_args)
        else:
            extra_val = getattr(result, "extra", None)
            if isinstance(extra_val, dict) and "_selection_args" in extra_val:
                selection_args = extra_val["_selection_args"]
                if selection_args:
                    dl_args.extend(selection_args)

        # download-media doesn't support -storage flag
        # It downloads to the configured directory, then add-file will handle storage
        # Note: Provider uploads (0x0) are not supported via this path

        # Call download-media with the URL in args
        return dl_cmdlet.run(None, dl_args, config)

    @staticmethod
    def _get_url(result: Any, pipe_obj: models.PipeObject) -> List[str]:
        from metadata import normalize_urls

        # Prefer explicit PipeObject.url if present
        urls: List[str] = []
        try:
            urls = normalize_urls(getattr(pipe_obj, "url", None))
        except Exception:
            urls = []

        # Then check extra.url
        if not urls:
            try:
                if isinstance(pipe_obj.extra, dict):
                    urls = normalize_urls(pipe_obj.extra.get("url"))
            except Exception:
                pass

        # Then check result dict
        if not urls and isinstance(result, dict):
            urls = normalize_urls(result.get("url"))

        # Finally, try extractor helper
        if not urls:
            urls = normalize_urls(extract_url_from_result(result))

        return urls

    @staticmethod
    def _get_relationships(result: Any, pipe_obj: models.PipeObject) -> Optional[Dict[str, Any]]:
        try:
            rels = pipe_obj.get_relationships()
            if rels:
                return rels
        except Exception:
            pass
        if isinstance(result, dict) and result.get("relationships"):
            return result.get("relationships")
        try:
            return extract_relationships(result)
        except Exception:
            return None

    @staticmethod
    def _get_duration(result: Any, pipe_obj: models.PipeObject) -> Optional[float]:
        def _parse_duration(value: Any) -> Optional[float]:
            if value is None:
                return None
            if isinstance(value, (int, float)):
                return float(value) if value > 0 else None
            if isinstance(value, str):
                s = value.strip()
                if not s:
                    return None
                try:
                    candidate = float(s)
                    return candidate if candidate > 0 else None
                except ValueError:
                    pass
                if ":" in s:
                    parts = [p.strip() for p in s.split(":") if p.strip()]
                    if len(parts) in {2, 3} and all(p.isdigit() for p in parts):
                        nums = [int(p) for p in parts]
                        if len(nums) == 2:
                            minutes, seconds = nums
                            return float(minutes * 60 + seconds)
                        hours, minutes, seconds = nums
                        return float(hours * 3600 + minutes * 60 + seconds)
            return None

        parsed = _parse_duration(getattr(pipe_obj, "duration", None))
        if parsed is not None:
            return parsed
        try:
            return _parse_duration(extract_duration(result))
        except Exception:
            return None

    @staticmethod
    def _update_pipe_object_destination(
        pipe_obj: models.PipeObject,
        *,
        hash_value: str,
        store: str,
        path: Optional[str],
        tag: List[str],
        title: Optional[str],
        extra_updates: Optional[Dict[str, Any]] = None,
    ) -> None:
        pipe_obj.hash = hash_value
        pipe_obj.store = store
        pipe_obj.path = path
        pipe_obj.tag = tag
        if title:
            pipe_obj.title = title
        if isinstance(pipe_obj.extra, dict):
            pipe_obj.extra.update(extra_updates or {})
        else:
            pipe_obj.extra = dict(extra_updates or {})

    @staticmethod
    def _emit_pipe_object(pipe_obj: models.PipeObject) -> None:
        from result_table import format_result
        log(format_result(pipe_obj, title="Result"), file=sys.stderr)
        ctx.emit(pipe_obj.to_dict())
        ctx.set_current_stage_table(None)

    @staticmethod
    def _emit_storage_result(payload: Dict[str, Any], *, overlay: bool = True, emit: bool = True) -> None:
        """Emit a storage-style result payload.

        - Always emits the dict downstream (when in a pipeline).
        - If this is the last stage (or not in a pipeline), prints a search-store-like table
          and sets an overlay table/items for @N selection.
        """
        # Emit for downstream commands (no-op if not in a pipeline)
        if emit:
            ctx.emit(payload)

        stage_ctx = ctx.get_stage_context()
        is_last = (stage_ctx is None) or bool(getattr(stage_ctx, "is_last_stage", False))
        if not is_last or not overlay:
            return

        try:
            from result_table import ResultTable
            table = ResultTable("Result")
            table.add_result(payload)
            # Overlay so @1 refers to this add-file result without overwriting search history
            ctx.set_last_result_table_overlay(table, [payload], subject=payload)
        except Exception:
            # If table rendering fails, still keep @ selection items
            try:
                ctx.set_last_result_items_only([payload])
            except Exception:
                pass

    @staticmethod
    def _try_emit_search_store_by_hash(*, store: str, hash_value: str, config: Dict[str, Any]) -> bool:
        """Run search-store for a single hash so the final table/payload is consistent.

        Important: `add-file` is treated as an action command by the CLI, so the CLI only
        prints tables for it when a display overlay exists. After running search-store,
        this copies the resulting table into the display overlay (when this is the last
        stage) so the canonical store table is what the user sees and can select from.

        Returns True if search-store ran successfully, else False.
        """
        try:
            from cmdlet.search_store import CMDLET as search_store_cmdlet

            args = ["-store", str(store), f"hash:{str(hash_value)}"]
            log(f"[add-file] Refresh: search-store -store {store} \"hash:{hash_value}\"", file=sys.stderr)

            # Run search-store under a temporary stage context so its ctx.emit() calls
            # don't interfere with the outer add-file pipeline stage.
            prev_ctx = ctx.get_stage_context()
            temp_ctx = ctx.PipelineStageContext(stage_index=0, total_stages=1, worker_id=getattr(prev_ctx, "worker_id", None))
            ctx.set_stage_context(temp_ctx)
            try:
                code = search_store_cmdlet.run(None, args, config)
            finally:
                ctx.set_stage_context(prev_ctx)
            if code != 0:
                return False

            # Promote the search-store result to a display overlay so the CLI prints it
            # for action commands like add-file.
            stage_ctx = ctx.get_stage_context()
            is_last = (stage_ctx is None) or bool(getattr(stage_ctx, "is_last_stage", False))
            if is_last:
                try:
                    table = ctx.get_last_result_table()
                    items = ctx.get_last_result_items()
                    if table is not None and items:
                        ctx.set_last_result_table_overlay(table, items, subject={"store": store, "hash": hash_value})
                except Exception:
                    pass

            return True
        except Exception as exc:
            debug(f"[add-file] Failed to run search-store after add-file: {type(exc).__name__}: {exc}")
            return False

    @staticmethod
    def _prepare_metadata(
        result: Any,
        media_path: Path,
        pipe_obj: models.PipeObject,
        config: Dict[str, Any],
    ) -> Tuple[List[str], List[str], Optional[str], Optional[str]]:
        """
        Prepare tags, url, and title for the file.
        Returns (tags, url, preferred_title, file_hash)
        """
        tags_from_result = list(pipe_obj.tag or [])
        if not tags_from_result:
            try:
                tags_from_result = list(extract_tag_from_result(result) or [])
            except Exception:
                tags_from_result = []

        url_from_result = Add_File._get_url(result, pipe_obj)

        preferred_title = pipe_obj.title
        if not preferred_title:
            for t in tags_from_result:
                if str(t).strip().lower().startswith("title:"):
                    candidate = t.split(":", 1)[1].strip().replace("_", " ").strip()
                    if candidate:
                        preferred_title = candidate
                        break
        if not preferred_title:
            preferred_title = extract_title_from_result(result)
            if preferred_title:
                preferred_title = preferred_title.replace("_", " ").strip()

        store = getattr(pipe_obj, "store", None)
        _, sidecar_hash, sidecar_tags, sidecar_url = Add_File._load_sidecar_bundle(
            media_path, store, config
        )

        def normalize_title_tag(tag: str) -> str:
            if str(tag).strip().lower().startswith("title:"):
                parts = tag.split(":", 1)
                if len(parts) == 2:
                    value = parts[1].replace("_", " ").strip()
                    return f"title:{value}"
            return tag

        tags_from_result_no_title = [t for t in tags_from_result if not str(t).strip().lower().startswith("title:")]
        sidecar_tags = collapse_namespace_tag([normalize_title_tag(t) for t in sidecar_tags], "title", prefer="last")
        sidecar_tags_filtered = [t for t in sidecar_tags if not str(t).strip().lower().startswith("title:")]

        merged_tags = merge_sequences(tags_from_result_no_title, sidecar_tags_filtered, case_sensitive=True)

        if preferred_title:
            merged_tags.append(f"title:{preferred_title}")

        merged_url = merge_sequences(url_from_result, sidecar_url, case_sensitive=False)

        file_hash = Add_File._resolve_file_hash(result, media_path, pipe_obj, sidecar_hash)

        # Persist back to PipeObject
        pipe_obj.tag = merged_tags
        if preferred_title and not pipe_obj.title:
            pipe_obj.title = preferred_title
        if file_hash and not pipe_obj.hash:
            pipe_obj.hash = file_hash
        if isinstance(pipe_obj.extra, dict):
            pipe_obj.extra.setdefault("url", merged_url)
        return merged_tags, merged_url, preferred_title, file_hash

    @staticmethod
    def _handle_local_export(
        media_path: Path,
        location: str,
        pipe_obj: models.PipeObject,
        config: Dict[str, Any],
        delete_after: bool,
    ) -> int:
        """Handle exporting to a specific local path (Copy)."""
        try:
            destination_root = Path(location)
        except Exception as exc:
            log(f"❌ Invalid destination path '{location}': {exc}", file=sys.stderr)
            return 1

        log(f"Exporting to local path: {destination_root}", file=sys.stderr)

        result = None
        tags, url, title, f_hash = Add_File._prepare_metadata(result, media_path, pipe_obj, config)

        # Determine Filename (Title-based)
        title_value = title
        if not title_value:
            # Try to find title in tags
            title_tag = next((t for t in tags if str(t).strip().lower().startswith("title:")), None)
            if title_tag:
                title_value = title_tag.split(":", 1)[1].strip()

        if not title_value:
            title_value = media_path.stem.replace("_", " ").strip()

        safe_title = "".join(c for c in title_value if c.isalnum() or c in " ._-()[]{}'`").strip()
        base_name = safe_title or media_path.stem
        new_name = base_name + media_path.suffix

        destination_root.mkdir(parents=True, exist_ok=True)
        target_path = destination_root / new_name

        if target_path.exists():
            target_path = unique_path(target_path)

        # COPY Operation (Safe Export)
        try:
            shutil.copy2(str(media_path), target_path)
        except Exception as exc:
            log(f"❌ Failed to export file: {exc}", file=sys.stderr)
            return 1

        # Copy Sidecars
        Add_File._copy_sidecars(media_path, target_path)

        # Ensure hash for exported copy
        if not f_hash:
            try:
                f_hash = sha256_file(target_path)
            except Exception:
                f_hash = None

        # Write Metadata Sidecars (since it's an export)
        relationships = Add_File._get_relationships(result, pipe_obj)
        try:
            write_sidecar(target_path, tags, url, f_hash)
            write_metadata(target_path, hash_value=f_hash, url=url, relationships=relationships or [])
        except Exception:
            pass

        # Update PipeObject and emit
        extra_updates = {
            "url": url,
            "export_path": str(destination_root),
        }
        if relationships:
            extra_updates["relationships"] = relationships

        chosen_title = title or title_value or pipe_obj.title or target_path.name

        Add_File._update_pipe_object_destination(
            pipe_obj,
            hash_value=f_hash or "unknown",
            store="local",
            path=str(target_path),
            tag=tags,
            title=chosen_title,
            extra_updates=extra_updates,
        )
        Add_File._emit_pipe_object(pipe_obj)

        # Cleanup
        # Only delete if explicitly requested!
        Add_File._cleanup_after_success(media_path, delete_source=delete_after)

        return 0

    @staticmethod
    def _download_soulseek_file(
        result: Any,
        config: Dict[str, Any]
    ) -> Optional[Path]:
        """
        Download a file from Soulseek peer.
        
        Extracts username and filename from soulseek result metadata and initiates download.
        """
        try:
            import asyncio
            from ProviderCore.registry import download_soulseek_file
            from pathlib import Path
            
            # Extract metadata from result
            full_metadata = {}
            if isinstance(result, dict):
                full_metadata = result.get("full_metadata", {})
            elif hasattr(result, "extra") and isinstance(result.extra, dict) and "full_metadata" in result.extra:
                full_metadata = result.extra.get("full_metadata", {})
            elif hasattr(result, "full_metadata"):
                # Direct attribute access (fallback)
                val = getattr(result, "full_metadata", {})
                if isinstance(val, dict):
                    full_metadata = val
            
            username = full_metadata.get("username")
            filename = full_metadata.get("filename")

            if not username or not filename:
                debug(f"[add-file] ERROR: Could not extract soulseek metadata from result (type={type(result).__name__})")
                extra_val = getattr(result, "extra", None)
                if isinstance(extra_val, dict):
                    debug(f"[add-file] Result extra keys: {list(extra_val.keys())}")
                return None
            
            if not username or not filename:
                debug(f"[add-file] ERROR: Missing soulseek metadata (username={username}, filename={filename})")
                return None
            
            debug(f"[add-file] Starting soulseek download: {username} -> {filename}")
            
            # Determine output directory (prefer downloads folder in config)
            output_dir = Path(config.get("output_dir", "./downloads")) if isinstance(config.get("output_dir"), str) else Path("./downloads")
            output_dir.mkdir(parents=True, exist_ok=True)
            
            # Run async download in event loop
            try:
                loop = asyncio.get_event_loop()
                if loop.is_closed():
                    loop = asyncio.new_event_loop()
                    asyncio.set_event_loop(loop)
            except RuntimeError:
                loop = asyncio.new_event_loop()
                asyncio.set_event_loop(loop)
            
            downloaded_path = loop.run_until_complete(
                download_soulseek_file(
                    username=username,
                    filename=filename,
                    output_dir=output_dir,
                    timeout=1200  # 20 minutes
                )
            )
            
            return downloaded_path
            
        except Exception as e:
            log(f"[add-file] Soulseek download error: {type(e).__name__}: {e}", file=sys.stderr)
            debug(f"[add-file] Soulseek download traceback: {e}")
            return None

    @staticmethod
    def _handle_provider_upload(
        media_path: Path,
        provider_name: str,
        pipe_obj: models.PipeObject,
        config: Dict[str, Any],
        delete_after: bool,
    ) -> int:
        """Handle uploading to a file provider (e.g. 0x0)."""
        from ProviderCore.registry import get_file_provider

        log(f"Uploading via {provider_name}: {media_path.name}", file=sys.stderr)

        try:
            file_provider = get_file_provider(provider_name, config)
            if not file_provider:
                log(f"File provider '{provider_name}' not available", file=sys.stderr)
                return 1

            hoster_url = file_provider.upload(str(media_path))
            log(f"File uploaded: {hoster_url}", file=sys.stderr)

            # Associate URL with Hydrus if possible
            f_hash = Add_File._resolve_file_hash(None, media_path, pipe_obj, None)
            if f_hash:
                try:
                    client = hydrus_wrapper.get_client(config)
                    if client:
                        client.associate_url(f_hash, hoster_url)
                except Exception:
                    pass

        except Exception as exc:
            log(f"Upload failed: {exc}", file=sys.stderr)
            return 1

        # Update PipeObject and emit
        extra_updates: Dict[str, Any] = {
            "provider": provider_name,
            "provider_url": hoster_url,
        }
        if isinstance(pipe_obj.extra, dict):
            # Also track hoster URL as a url for downstream steps
            existing_known = list(pipe_obj.extra.get("url") or [])
            if hoster_url and hoster_url not in existing_known:
                existing_known.append(hoster_url)
            extra_updates["url"] = existing_known

        file_path = pipe_obj.path or (str(media_path) if media_path else None) or ""
        Add_File._update_pipe_object_destination(
            pipe_obj,
            hash_value=f_hash or "unknown",
            store=provider_name or "provider",
            path=file_path,
            tag=pipe_obj.tag,
            title=pipe_obj.title or (media_path.name if media_path else None),
            extra_updates=extra_updates,
        )
        Add_File._emit_pipe_object(pipe_obj)

        Add_File._cleanup_after_success(media_path, delete_source=delete_after)
        return 0

    @staticmethod
    def _handle_storage_backend(
        result: Any,
        media_path: Path,
        backend_name: str,
        pipe_obj: models.PipeObject,
        config: Dict[str, Any],
        delete_after: bool,
        *,
        collect_payloads: Optional[List[Dict[str, Any]]] = None,
        suppress_last_stage_overlay: bool = False,
        auto_search_store: bool = True,
    ) -> int:
        """Handle uploading to a registered storage backend (e.g., 'test' folder store, 'hydrus', etc.)."""
        log(f"Adding file to storage backend '{backend_name}': {media_path.name}", file=sys.stderr)

        delete_after_effective = bool(delete_after)
        if not delete_after_effective:
            # When download-media is piped into add-file, the downloaded artifact is a temp file.
            # After it is persisted to a storage backend, delete the temp copy to avoid duplicates.
            try:
                if (
                    str(backend_name or "").strip().lower() != "temp"
                    and getattr(pipe_obj, "is_temp", False)
                    and getattr(pipe_obj, "action", None) == "cmdlet:download-media"
                ):
                    from config import resolve_output_dir
                    temp_dir = resolve_output_dir(config)
                    try:
                        if media_path.resolve().is_relative_to(temp_dir.expanduser().resolve()):
                            delete_after_effective = True
                            debug(f"[add-file] Auto-delete temp source after ingest: {media_path}")
                    except Exception:
                        # If path resolution fails, fall back to non-destructive behavior
                        pass
            except Exception:
                pass
        
        try:
            store = Store(config)
            backend = store[backend_name]
            
            # Prepare metadata from pipe_obj and sidecars
            tags, url, title, f_hash = Add_File._prepare_metadata(result, media_path, pipe_obj, config)
            
            # Call backend's add_file with full metadata
            # Backend returns hash as identifier
            file_identifier = backend.add_file(
                media_path, 
                title=title,
                tag=tags,
                url=url
            )
            log(f"✓ File added to '{backend_name}': {file_identifier}", file=sys.stderr)

            stored_path: Optional[str] = None
            try:
                maybe_path = backend.get_file(file_identifier)
                if isinstance(maybe_path, Path):
                    stored_path = str(maybe_path)
                elif isinstance(maybe_path, str) and maybe_path:
                    # Some backends may return a browser URL
                    stored_path = maybe_path
            except Exception:
                stored_path = None
            
            Add_File._update_pipe_object_destination(
                pipe_obj,
                hash_value=file_identifier if len(file_identifier) == 64 else f_hash or "unknown",
                store=backend_name,
                path=stored_path,
                tag=tags,
                title=title or pipe_obj.title or media_path.name,
                extra_updates={
                    "url": url,
                },
            )

            # Emit a search-store-like payload for consistent tables and natural piping.
            # Keep hash/store for downstream commands (get-tag, get-file, etc.).
            resolved_hash = file_identifier if len(file_identifier) == 64 else (f_hash or file_identifier or "unknown")

            # If we have url(s), ensure they get associated with the destination file.
            # This mirrors `add-url` behavior but avoids emitting extra pipeline noise.
            if url:
                try:
                    backend.add_url(resolved_hash, list(url))
                except Exception:
                    pass

            meta: Dict[str, Any] = {}
            try:
                meta = backend.get_metadata(resolved_hash) or {}
            except Exception:
                meta = {}

            # Determine size bytes
            size_bytes: Optional[int] = None
            for key in ("size_bytes", "size", "filesize", "file_size"):
                try:
                    raw_size = meta.get(key)
                    if raw_size is not None:
                        size_bytes = int(raw_size)
                        break
                except Exception:
                    pass
            if size_bytes is None:
                try:
                    size_bytes = int(media_path.stat().st_size)
                except Exception:
                    size_bytes = None

            # Determine title/ext
            title_out = (
                meta.get("title")
                or title
                or pipe_obj.title
                or media_path.stem
                or media_path.name
            )
            ext_out = (meta.get("ext") or media_path.suffix.lstrip("."))

            payload: Dict[str, Any] = {
                "title": title_out,
                "ext": str(ext_out or ""),
                "size_bytes": size_bytes,
                "store": backend_name,
                "hash": resolved_hash,
                # Preserve extra fields for downstream commands (kept hidden by default table rules)
                "path": stored_path,
                "tag": list(tags or []),
                "url": list(url or []),
            }
            if collect_payloads is not None:
                try:
                    collect_payloads.append(payload)
                except Exception:
                    pass

            # Keep the add-file 1-row summary overlay (when last stage), then emit the
            # canonical search-store payload/table for piping/selection consistency.
            if auto_search_store and resolved_hash and resolved_hash != "unknown":
                # Show the add-file summary (overlay only) but let search-store provide the downstream payload.
                Add_File._emit_storage_result(payload, overlay=not suppress_last_stage_overlay, emit=False)

                ok = Add_File._try_emit_search_store_by_hash(
                    store=backend_name,
                    hash_value=resolved_hash,
                    config=config,
                )
                if not ok:
                    # Fall back to emitting the add-file payload so downstream stages still receive an item.
                    ctx.emit(payload)
            else:
                Add_File._emit_storage_result(payload, overlay=not suppress_last_stage_overlay, emit=True)
            
            Add_File._cleanup_after_success(media_path, delete_source=delete_after_effective)
            return 0
            
        except Exception as exc:
            log(f"❌ Failed to add file to backend '{backend_name}': {exc}", file=sys.stderr)
            import traceback
            traceback.print_exc(file=sys.stderr)
            return 1

    # --- Helpers ---

    @staticmethod
    def _load_sidecar_bundle(
        media_path: Path,
        store: Optional[str],
        config: Dict[str, Any],
    ) -> Tuple[Optional[Path], Optional[str], List[str], List[str]]:
        """Load sidecar metadata."""
        if store and store.lower() == "local":
            try:
                from config import get_local_storage_path
                db_root = get_local_storage_path(config)
                if db_root:
                    with API_folder_store(Path(db_root)) as db:
                        file_hash = db.get_file_hash(media_path)
                        if file_hash:
                            tags = db.get_tags(file_hash) or []
                            metadata = db.get_metadata(file_hash) or {}
                            url = metadata.get("url") or []
                            f_hash = metadata.get("hash") or file_hash
                            if tags or url or f_hash:
                                return None, f_hash, tags, url
            except Exception:
                pass

        try:
            sidecar_path = find_sidecar(media_path)
            if sidecar_path and sidecar_path.exists():
                h, t, u = read_sidecar(sidecar_path)
                return sidecar_path, h, t or [], u or []
        except Exception:
            pass
        return None, None, [], []

    @staticmethod
    def _resolve_file_hash(
        result: Any,
        media_path: Path,
        pipe_obj: models.PipeObject,
        fallback_hash: Optional[str],
    ) -> Optional[str]:
        if pipe_obj.hash and pipe_obj.hash != "unknown":
            return pipe_obj.hash
        if fallback_hash:
            return fallback_hash

        if isinstance(result, dict):
            candidate = result.get('hash')
            if candidate:
                return str(candidate)

        try:
            return sha256_file(media_path)
        except Exception:
            return None

    @staticmethod
    def _resolve_media_kind(path: Path) -> str:
        # Reusing logic
        suffix = path.suffix.lower()
        if suffix in {'.mp3', '.flac', '.wav', '.m4a', '.aac', '.ogg', '.opus', '.wma', '.mka'}:
            return 'audio'
        if suffix in {'.mp4', '.mkv', '.webm', '.mov', '.avi', '.flv', '.mpg', '.mpeg', '.ts', '.m4v', '.wmv'}:
            return 'video'
        if suffix in {'.jpg', '.jpeg', '.png', '.gif', '.webp', '.bmp', '.tiff'}:
            return 'image'
        if suffix in {'.pdf', '.epub', '.txt', '.mobi', '.azw3', '.cbz', '.cbr', '.doc', '.docx'}:
            return 'document'
        return 'other'

    @staticmethod
    def _persist_local_metadata(
        library_root: Path,
        dest_path: Path,
        tags: List[str],
        url: List[str],
        f_hash: Optional[str],
        relationships: Any,
        duration: Any,
        media_kind: str,
    ):
        payload = {
            'hash': f_hash,
            'url': url,
            'relationships': relationships or [],
            'duration': duration,
            'size': None,
            'ext': dest_path.suffix.lower(),
            'media_type': media_kind,
            'media_kind': media_kind,
        }
        try:
            payload['size'] = dest_path.stat().st_size
        except OSError:
            payload['size'] = None

        with API_folder_store(library_root) as db:
            try:
                db.save_file_info(dest_path, payload, tags)
            except Exception as exc:
                log(f"⚠️  Failed to persist metadata: {exc}", file=sys.stderr)

    @staticmethod
    def _copy_sidecars(source_path: Path, target_path: Path):
        possible_sidecars = [
            source_path.with_suffix(source_path.suffix + ".json"),
            source_path.with_name(source_path.name + ".tag"),
            source_path.with_name(source_path.name + ".metadata"),
            source_path.with_name(source_path.name + ".notes"),
        ]
        for sc in possible_sidecars:
            try:
                if sc.exists():
                    suffix_part = sc.name.replace(source_path.name, "", 1)
                    dest_sidecar = target_path.parent / f"{target_path.name}{suffix_part}"
                    dest_sidecar.parent.mkdir(parents=True, exist_ok=True)
                    shutil.copy2(str(sc), dest_sidecar)
            except Exception:
                pass

    @staticmethod
    def _cleanup_after_success(media_path: Path, delete_source: bool):
        if not delete_source:
            return

        # Check if it's a temp file that should always be deleted
        is_temp_merge = "(merged)" in media_path.name or ".dlhx_" in media_path.name

        if delete_source or is_temp_merge:
            log(f"Deleting source file...", file=sys.stderr)
            try:
                media_path.unlink()
                Add_File._cleanup_sidecar_files(media_path)
            except Exception as exc:
                log(f"⚠️  Could not delete file: {exc}", file=sys.stderr)

    @staticmethod
    def _cleanup_sidecar_files(media_path: Path):
        targets = [
            media_path.parent / (media_path.name + '.metadata'),
            media_path.parent / (media_path.name + '.notes'),
            media_path.parent / (media_path.name + '.tag'),
        ]
        for target in targets:
            try:
                if target.exists():
                    target.unlink()
            except Exception:
                pass


# Create and register the cmdlet
CMDLET = Add_File()