Medios-Macina/cmdlet/add_tag.py

from __future__ import annotations

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

from SYS.logger import log

import models
import pipeline as ctx
from ._shared import normalize_result_input, filter_results_by_temp
from ._shared import (
    Cmdlet,
    CmdletArg,
    SharedArgs,
    normalize_hash,
    parse_tag_arguments,
    expand_tag_groups,
    parse_cmdlet_args,
    collapse_namespace_tag,
    should_show_help,
    get_field,
)
from Store import Store
from SYS.utils import sha256_file


def _extract_title_tag(tags: List[str]) -> Optional[str]:
    """Return the value of the first title: tag if present."""
    for t in tags:
        if t.lower().startswith("title:"):
            value = t.split(":", 1)[1].strip()
            return value or None
    return None


def _extract_item_tags(res: Any) -> List[str]:
    if isinstance(res, models.PipeObject):
        raw = getattr(res, "tag", None)
    elif isinstance(res, dict):
        raw = res.get("tag")
    else:
        raw = None

    if isinstance(raw, list):
        return [str(t) for t in raw if t is not None]
    if isinstance(raw, str) and raw.strip():
        return [raw]
    return []


def _set_item_tags(res: Any, tags: List[str]) -> None:
    if isinstance(res, models.PipeObject):
        res.tag = tags
    elif isinstance(res, dict):
        res["tag"] = tags


def _apply_title_to_result(res: Any, title_value: Optional[str]) -> None:
    """Update result object/dict title fields and columns in-place."""
    if not title_value:
        return
    if isinstance(res, models.PipeObject):
        res.title = title_value
        # Update columns if present (Title column assumed index 0)
        columns = getattr(res, "columns", None)
        if isinstance(columns, list) and columns:
            label, *_ = columns[0]
            if str(label).lower() == "title":
                columns[0] = (label, title_value)
    elif isinstance(res, dict):
        res["title"] = title_value
        cols = res.get("columns")
        if isinstance(cols, list):
            updated = []
            changed = False
            for col in cols:
                if isinstance(col, tuple) and len(col) == 2:
                    label, _val = col
                    if str(label).lower() == "title":
                        updated.append((label, title_value))
                        changed = True
                    else:
                        updated.append(col)
                else:
                    updated.append(col)
            if changed:
                res["columns"] = updated


def _matches_target(
    item: Any,
    target_hash: Optional[str],
    target_path: Optional[str],
    target_store: Optional[str] = None,
) -> bool:
    """Determine whether a result item refers to the given target.

    Important: hashes can collide across backends in this app's UX (same media in
    multiple stores). When target_store is provided, it must match too.
    """

    def norm(val: Any) -> Optional[str]:
        return str(val).lower() if val is not None else None

    target_hash_l = target_hash.lower() if target_hash else None
    target_path_l = target_path.lower() if target_path else None
    target_store_l = target_store.lower() if target_store else None

    if isinstance(item, dict):
        hashes = [norm(item.get("hash"))]
        paths = [norm(item.get("path"))]
        stores = [norm(item.get("store"))]
    else:
        hashes = [norm(get_field(item, "hash"))]
        paths = [norm(get_field(item, "path"))]
        stores = [norm(get_field(item, "store"))]

    if target_store_l:
        if target_store_l not in stores:
            return False

    if target_hash_l and target_hash_l in hashes:
        return True
    if target_path_l and target_path_l in paths:
        return True
    return False


def _update_item_title_fields(item: Any, new_title: str) -> None:
    """Mutate an item to reflect a new title in plain fields and columns."""
    if isinstance(item, models.PipeObject):
        item.title = new_title
        columns = getattr(item, "columns", None)
        if isinstance(columns, list) and columns:
            label, *_ = columns[0]
            if str(label).lower() == "title":
                columns[0] = (label, new_title)
    elif isinstance(item, dict):
        item["title"] = new_title
        cols = item.get("columns")
        if isinstance(cols, list):
            updated_cols = []
            changed = False
            for col in cols:
                if isinstance(col, tuple) and len(col) == 2:
                    label, _val = col
                    if str(label).lower() == "title":
                        updated_cols.append((label, new_title))
                        changed = True
                    else:
                        updated_cols.append(col)
                else:
                    updated_cols.append(col)
            if changed:
                item["columns"] = updated_cols


def _refresh_result_table_title(
    new_title: str,
    target_hash: Optional[str],
    target_store: Optional[str],
    target_path: Optional[str],
) -> None:
    """Refresh the cached result table with an updated title and redisplay it."""
    try:
        last_table = ctx.get_last_result_table()
        items = ctx.get_last_result_items()
        if not last_table or not items:
            return

        updated_items = []
        match_found = False
        for item in items:
            try:
                if _matches_target(item, target_hash, target_path, target_store):
                    _update_item_title_fields(item, new_title)
                    match_found = True
            except Exception:
                pass
            updated_items.append(item)
        if not match_found:
            return

        new_table = last_table.copy_with_title(getattr(last_table, "title", ""))

        for item in updated_items:
            new_table.add_result(item)

        # Keep the underlying history intact; update only the overlay so @.. can
        # clear the overlay then continue back to prior tables (e.g., the search list).
        ctx.set_last_result_table_overlay(new_table, updated_items)
    except Exception:
        pass


def _refresh_tag_view(res: Any, target_hash: Optional[str], store_name: Optional[str], target_path: Optional[str], config: Dict[str, Any]) -> None:
    """Refresh tag display via get-tag. Prefer current subject; fall back to direct hash refresh."""
    try:
        from cmdlet import get as get_cmdlet  # type: ignore
    except Exception:
        return

    if not target_hash or not store_name:
        return

    refresh_args: List[str] = ["-hash", target_hash, "-store", store_name]

    get_tag = None
    try:
        get_tag = get_cmdlet("get-tag")
    except Exception:
        get_tag = None
    if not callable(get_tag):
        return

    try:
        subject = ctx.get_last_result_subject()
        if subject and _matches_target(subject, target_hash, target_path, store_name):
            get_tag(subject, refresh_args, config)
            return
    except Exception:
        pass

    try:
        get_tag(res, refresh_args, config)
    except Exception:
        pass



class Add_Tag(Cmdlet):
    """Class-based add-tag cmdlet with Cmdlet metadata inheritance."""

    def __init__(self) -> None:
        super().__init__(
            name="add-tag",
            summary="Add tag to a file in a store.",
            usage="add-tag -store <store> [-hash <sha256>] [-duplicate <format>] [-list <list>[,<list>...]] [--all] <tag>[,<tag>...]",
            arg=[
                CmdletArg("tag", type="string", required=False, description="One or more tag to add. Comma- or space-separated. Can also use {list_name} syntax. If omitted, uses tag from pipeline payload.", variadic=True),
                SharedArgs.HASH,
                SharedArgs.STORE,
                CmdletArg("-duplicate", type="string", description="Copy existing tag values to new namespaces. Formats: title:album,artist (explicit) or title,album,artist (inferred)"),
                CmdletArg("-list", type="string", description="Load predefined tag lists from adjective.json. Comma-separated list names (e.g., -list philosophy,occult)."),
                CmdletArg("--all", type="flag", description="Include temporary files in tagging (by default, only tag non-temporary files)."),
            ],
            detail=[
                "- By default, only tag non-temporary files (from pipelines). Use --all to tag everything.",
                "- Requires a store backend: use -store or pipe items that include store.",
                "- If -hash is not provided, uses the piped item's hash (or derives from its path when possible).",
                "- Multiple tag can be comma-separated or space-separated.",
                "- Use -list to include predefined tag lists from adjective.json: -list philosophy,occult",
                "- tag can also reference lists with curly braces: add-tag {philosophy} \"other:tag\"",
                "- Use -duplicate to copy EXISTING tag values to new namespaces:",
                "  Explicit format: -duplicate title:album,artist (copies title: to album: and artist:)",
                "  Inferred format: -duplicate title,album,artist (first is source, rest are targets)",
                "- The source namespace must already exist in the file being tagged.",
                "- Target namespaces that already have a value are skipped (not overwritten).",
                "- You can also pass the target hash as a tag token: hash:<sha256>. This overrides -hash and is removed from the tag list.",
            ],
            exec=self.run,
        )
        self.register()

    def run(self, result: Any, args: Sequence[str], config: Dict[str, Any]) -> int:
        """Add tag to a file with smart filtering for pipeline results."""
        if should_show_help(args):
            log(f"Cmdlet: {self.name}\nSummary: {self.summary}\nUsage: {self.usage}")
            return 0

        # Parse arguments
        parsed = parse_cmdlet_args(args, self)

        # Check for --all flag
        include_temp = parsed.get("all", False)

        # Normalize input to list
        results = normalize_result_input(result)

        # Filter by temp status (unless --all is set)
        if not include_temp:
            results = filter_results_by_temp(results, include_temp=False)

        if not results:
            log("No valid files to tag (all results were temporary; use --all to include temporary files)", file=sys.stderr)
            return 1

        # Get tag from arguments (or fallback to pipeline payload)
        raw_tag = parsed.get("tag", [])
        if isinstance(raw_tag, str):
            raw_tag = [raw_tag]

        # Fallback: if no tag provided explicitly, try to pull from first result payload
        if not raw_tag and results:
            first = results[0]
            payload_tag = None
            
            # Try multiple tag lookup strategies in order
            tag_lookups = [
                lambda x: getattr(x, "tag", None),
                lambda x: x.get("tag") if isinstance(x, dict) else None,
            ]
            
            for lookup in tag_lookups:
                try:
                    payload_tag = lookup(first)
                    if payload_tag:
                        break
                except (AttributeError, TypeError, KeyError):
                    continue
            
            if payload_tag:
                if isinstance(payload_tag, str):
                    raw_tag = [payload_tag]
                elif isinstance(payload_tag, list):
                    raw_tag = payload_tag

        # Handle -list argument (convert to {list} syntax)
        list_arg = parsed.get("list")
        if list_arg:
            for l in list_arg.split(','):
                l = l.strip()
                if l:
                    raw_tag.append(f"{{{l}}}")

        # Parse and expand tag
        tag_to_add = parse_tag_arguments(raw_tag)
        tag_to_add = expand_tag_groups(tag_to_add)

        # Allow hash override via namespaced token (e.g., "hash:abcdef...")
        extracted_hash = None
        filtered_tag: List[str] = []
        for tag in tag_to_add:
            if isinstance(tag, str) and tag.lower().startswith("hash:"):
                _, _, hash_val = tag.partition(":")
                if hash_val:
                    extracted_hash = normalize_hash(hash_val.strip())
                continue
            filtered_tag.append(tag)
        tag_to_add = filtered_tag

        if not tag_to_add:
            log("No tag provided to add", file=sys.stderr)
            return 1

        # Get other flags (hash override can come from -hash or hash: token)
        hash_override = normalize_hash(parsed.get("hash")) or extracted_hash
        duplicate_arg = parsed.get("duplicate")

        # tag ARE provided - apply them to each store-backed result
        total_added = 0
        total_modified = 0

        store_override = parsed.get("store")

        store_registry = Store(config)

        for res in results:
            store_name: Optional[str]
            raw_hash: Optional[str]
            raw_path: Optional[str]

            if isinstance(res, models.PipeObject):
                store_name = store_override or res.store
                raw_hash = res.hash
                raw_path = res.path
            elif isinstance(res, dict):
                store_name = store_override or res.get("store")
                raw_hash = res.get("hash")
                raw_path = res.get("path")
            else:
                ctx.emit(res)
                continue

            if not store_name:
                store_name = None

            # If the item isn't in a configured store backend yet (e.g., store=PATH) but has a local file,
            # treat add-tag as a pipeline mutation (carry tags forward for add-file) instead of a store write.
            if not store_override:
                store_name_str = str(store_name) if store_name is not None else ""
                local_mode_requested = (not store_name_str) or (store_name_str.upper() == "PATH") or (store_name_str.lower() == "local")
                is_known_backend = bool(store_name_str) and store_registry.is_available(store_name_str)

                if local_mode_requested and raw_path:
                    try:
                        if Path(str(raw_path)).expanduser().exists():
                            existing_tag_list = _extract_item_tags(res)
                            existing_lower = {t.lower() for t in existing_tag_list if isinstance(t, str)}

                            item_tag_to_add = list(tag_to_add)
                            item_tag_to_add = collapse_namespace_tag(item_tag_to_add, "title", prefer="last")

                            if duplicate_arg:
                                parts = str(duplicate_arg).split(':')
                                source_ns = ""
                                targets: list[str] = []

                                if len(parts) > 1:
                                    source_ns = parts[0]
                                    targets = [t.strip() for t in parts[1].split(',') if t.strip()]
                                else:
                                    parts2 = str(duplicate_arg).split(',')
                                    if len(parts2) > 1:
                                        source_ns = parts2[0]
                                        targets = [t.strip() for t in parts2[1:] if t.strip()]

                                if source_ns and targets:
                                    source_prefix = source_ns.lower() + ":"
                                    for t in existing_tag_list:
                                        if not t.lower().startswith(source_prefix):
                                            continue
                                        value = t.split(":", 1)[1]
                                        for target_ns in targets:
                                            new_tag = f"{target_ns}:{value}"
                                            if new_tag.lower() not in existing_lower:
                                                item_tag_to_add.append(new_tag)

                            removed_namespace_tag: list[str] = []
                            for new_tag in item_tag_to_add:
                                if not isinstance(new_tag, str) or ":" not in new_tag:
                                    continue
                                ns = new_tag.split(":", 1)[0].strip()
                                if not ns:
                                    continue
                                ns_prefix = ns.lower() + ":"
                                for t in existing_tag_list:
                                    if t.lower().startswith(ns_prefix) and t.lower() != new_tag.lower():
                                        removed_namespace_tag.append(t)
                            removed_namespace_tag = sorted({t for t in removed_namespace_tag})

                            actual_tag_to_add = [
                                t
                                for t in item_tag_to_add
                                if isinstance(t, str) and t.lower() not in existing_lower
                            ]

                            updated_tag_list = [t for t in existing_tag_list if t not in removed_namespace_tag]
                            updated_tag_list.extend(actual_tag_to_add)

                            _set_item_tags(res, updated_tag_list)
                            final_title = _extract_title_tag(updated_tag_list)
                            _apply_title_to_result(res, final_title)

                            total_added += len(actual_tag_to_add)
                            total_modified += 1 if (removed_namespace_tag or actual_tag_to_add) else 0

                            ctx.emit(res)
                            continue
                    except Exception:
                        pass

                if local_mode_requested:
                    log("[add_tag] Error: Missing usable local path for tagging (or provide -store)", file=sys.stderr)
                    return 1

                if store_name_str and not is_known_backend:
                    log(f"[add_tag] Error: Unknown store '{store_name_str}'. Available: {store_registry.list_backends()}", file=sys.stderr)
                    return 1

            resolved_hash = normalize_hash(hash_override) if hash_override else normalize_hash(raw_hash)
            if not resolved_hash and raw_path:
                try:
                    p = Path(str(raw_path))
                    stem = p.stem
                    if len(stem) == 64 and all(c in "0123456789abcdef" for c in stem.lower()):
                        resolved_hash = stem.lower()
                    elif p.exists() and p.is_file():
                        resolved_hash = sha256_file(p)
                except Exception:
                    resolved_hash = None

            if not resolved_hash:
                log("[add_tag] Warning: Item missing usable hash (and could not derive from path); skipping", file=sys.stderr)
                ctx.emit(res)
                continue

            try:
                backend = store_registry[str(store_name)]
            except Exception as exc:
                log(f"[add_tag] Error: Unknown store '{store_name}': {exc}", file=sys.stderr)
                return 1

            try:
                existing_tag, _src = backend.get_tag(resolved_hash, config=config)
            except Exception:
                existing_tag = []

            existing_tag_list = [t for t in (existing_tag or []) if isinstance(t, str)]
            existing_lower = {t.lower() for t in existing_tag_list}
            original_title = _extract_title_tag(existing_tag_list)

            # Per-item tag list (do not mutate shared list)
            item_tag_to_add = list(tag_to_add)
            item_tag_to_add = collapse_namespace_tag(item_tag_to_add, "title", prefer="last")

            # Handle -duplicate logic (copy existing tag to new namespaces)
            if duplicate_arg:
                parts = str(duplicate_arg).split(':')
                source_ns = ""
                targets: list[str] = []

                if len(parts) > 1:
                    source_ns = parts[0]
                    targets = [t.strip() for t in parts[1].split(',') if t.strip()]
                else:
                    parts2 = str(duplicate_arg).split(',')
                    if len(parts2) > 1:
                        source_ns = parts2[0]
                        targets = [t.strip() for t in parts2[1:] if t.strip()]

                if source_ns and targets:
                    source_prefix = source_ns.lower() + ":"
                    for t in existing_tag_list:
                        if not t.lower().startswith(source_prefix):
                            continue
                        value = t.split(":", 1)[1]
                        for target_ns in targets:
                            new_tag = f"{target_ns}:{value}"
                            if new_tag.lower() not in existing_lower:
                                item_tag_to_add.append(new_tag)

            # Namespace replacement: delete old namespace:* when adding namespace:value
            removed_namespace_tag: list[str] = []
            for new_tag in item_tag_to_add:
                if not isinstance(new_tag, str) or ":" not in new_tag:
                    continue
                ns = new_tag.split(":", 1)[0].strip()
                if not ns:
                    continue
                ns_prefix = ns.lower() + ":"
                for t in existing_tag_list:
                    if t.lower().startswith(ns_prefix) and t.lower() != new_tag.lower():
                        removed_namespace_tag.append(t)

            removed_namespace_tag = sorted({t for t in removed_namespace_tag})

            actual_tag_to_add = [t for t in item_tag_to_add if isinstance(t, str) and t.lower() not in existing_lower]

            changed = False
            if removed_namespace_tag:
                try:
                    ok_del = backend.delete_tag(resolved_hash, removed_namespace_tag, config=config)
                    if ok_del:
                        changed = True
                except Exception as exc:
                    log(f"[add_tag] Warning: Failed deleting namespace tag: {exc}", file=sys.stderr)

            if actual_tag_to_add:
                try:
                    ok_add = backend.add_tag(resolved_hash, actual_tag_to_add, config=config)
                    if ok_add:
                        changed = True
                    else:
                        log("[add_tag] Warning: Store rejected tag update", file=sys.stderr)
                except Exception as exc:
                    log(f"[add_tag] Warning: Failed adding tag: {exc}", file=sys.stderr)

            if changed:
                total_added += len(actual_tag_to_add)
                total_modified += 1

            try:
                refreshed_tag, _src2 = backend.get_tag(resolved_hash, config=config)
                refreshed_list = [t for t in (refreshed_tag or []) if isinstance(t, str)]
            except Exception:
                refreshed_list = existing_tag_list

            # Update the result's tag using canonical field
            if isinstance(res, models.PipeObject):
                res.tag = refreshed_list
            elif isinstance(res, dict):
                res["tag"] = refreshed_list

            final_title = _extract_title_tag(refreshed_list)
            _apply_title_to_result(res, final_title)

            if final_title and (not original_title or final_title.lower() != original_title.lower()):
                _refresh_result_table_title(final_title, resolved_hash, str(store_name), raw_path)

            if changed:
                _refresh_tag_view(res, resolved_hash, str(store_name), raw_path, config)

            ctx.emit(res)

        log(
            f"[add_tag] Added {total_added} new tag(s) across {len(results)} item(s); modified {total_modified} item(s)",
            file=sys.stderr,
        )
        return 0


CMDLET = Add_Tag()