# SPDX-License-Identifier: Apache-2.0
# Copyright 2025-2026 Raykosan (RaykoStudio)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import folder_paths
import json
import os
import struct
import hashlib
import re
from nodes import LoraLoader
from server import PromptServer
import aiohttp

CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
PRESETS_DIR = os.path.join(CURRENT_DIR, "preset_models")
LORA_PRESETS_DIR = os.path.join(PRESETS_DIR, "loras")
RAYKO_LORA_DATA_DIR = os.path.join(CURRENT_DIR, "rayko_lora_data")

os.makedirs(PRESETS_DIR, exist_ok=True)
os.makedirs(LORA_PRESETS_DIR, exist_ok=True)
os.makedirs(RAYKO_LORA_DATA_DIR, exist_ok=True)


def clean_html(text):
    if not text:
        return ""
    clean = re.sub(r'<[^>]+>', ' ', str(text))
    clean = re.sub(r'\s+', ' ', clean).strip()
    return clean


def sanitize_preset_name(name):
    """Sanitize a preset name to a safe filename component.

    Strips all characters except alphanumerics, spaces, underscores and dashes.
    Returns an empty string if the result is empty.
    """
    if not isinstance(name, str):
        return ""
    return "".join(c for c in name.strip() if c.isalnum() or c in " _-").strip()


def _safe_preset_path(name):
    """Return an absolute path inside LORA_PRESETS_DIR for the given preset name.

    Returns None if the name is invalid or if the resulting path escapes
    LORA_PRESETS_DIR (defense in depth against path traversal).
    """
    safe_name = sanitize_preset_name(name)
    if not safe_name:
        return None
    base = os.path.realpath(LORA_PRESETS_DIR)
    candidate = os.path.realpath(os.path.join(base, f"{safe_name}.json"))
    try:
        if os.path.commonpath([base, candidate]) != base:
            return None
    except ValueError:
        # Different drives on Windows / incompatible paths
        return None
    return candidate


def extract_metadata_from_safetensors(lora_full_path):
    try:
        with open(lora_full_path, "rb") as f:
            header_size = struct.unpack("<Q", f.read(8))[0]
            header_bytes = f.read(header_size)
            header = json.loads(header_bytes.decode("utf-8"))
            if "__metadata__" in header:
                meta = header["__metadata__"]
                return {
                    "full_name": meta.get("modelspec.title", ""),
                    "trained_words": [t.strip() for t in meta.get("modelspec.tags", "").split(",") if t.strip()] if "modelspec.tags" in meta else [],
                    "description": clean_html(meta.get("modelspec.description", "") or meta.get("ss_training_comment", ""))
                }
    except Exception:
        pass
    return None


def compute_sha256(file_path):
    sha256_hash = hashlib.sha256()
    with open(file_path, "rb") as f:
        for chunk in iter(lambda: f.read(8192), b""):
            sha256_hash.update(chunk)
    return sha256_hash.hexdigest()


def save_to_rayko_db(file_hash, data, source, overwrite=False):
    db_path = os.path.join(RAYKO_LORA_DATA_DIR, f"{file_hash}.json")
    if not overwrite and os.path.exists(db_path):
        return False
    db_data = {
        "hash": file_hash,
        "name": data.get("full_name", ""),
        "trained_words": data.get("trained_words", []),
        "description": data.get("description", ""),
        "source": source,
        "imported_at": __import__('datetime').datetime.now().isoformat()
    }
    with open(db_path, 'w', encoding='utf-8') as f:
        json.dump(db_data, f, indent=2, ensure_ascii=False)
    return True


class RaykoLoRALoader:
    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "model": ("MODEL", {"tooltip": "Input model to apply LoRA"}),
                "use_clip": ("BOOLEAN", {"default": True, "label": "Apply LoRA to CLIP"}),
                "lora_data": ("STRING", {"default": "[]", "multiline": False, "forceInput": False}),
            },
            "optional": {
                "clip": ("CLIP", {"tooltip": "Input CLIP to apply LoRA (optional)"}),
            }
        }

    RETURN_TYPES = ("MODEL", "CLIP")
    RETURN_NAMES = ("MODEL", "CLIP")
    FUNCTION = "load_lora"
    CATEGORY = "🦊 RaykoStudio"
    DESCRIPTION = "Applies multiple LoRA models to the input model and optionally to CLIP."

    @classmethod
    def IS_CHANGED(cls, lora_data="[]", **kwargs):
        if lora_data is None:
            lora_data = "[]"
        return hashlib.md5(lora_data.encode()).hexdigest()

    def load_lora(self, model, use_clip, clip=None, lora_data="[]"):
        if not lora_data:
            lora_data = "[]"
        try:
            loras = json.loads(lora_data) if lora_data else []
            if not isinstance(loras, list):
                loras = []
        except Exception:
            loras = []

        lora_loader = LoraLoader()
        for lora in loras:
            if not isinstance(lora, dict):
                continue
            name = lora.get("name", "")
            strength_model = float(lora.get("strength_model", 1.0))
            strength_clip = float(lora.get("strength_clip", strength_model))
            enabled = lora.get("enabled", True)

            if name and name != "None" and enabled:
                lora_relative_path = name.replace("\\", "/")
                lora_full_path = folder_paths.get_full_path("loras", lora_relative_path)
                if lora_full_path:
                    try:
                        if use_clip and clip is not None:
                            model, clip = lora_loader.load_lora(model, clip, lora_relative_path, strength_model, strength_clip)
                        else:
                            model, _ = lora_loader.load_lora(model, None, lora_relative_path, strength_model, strength_clip)
                    except Exception as e:
                        print(f"[Rayko] LoRA Error: {e}")

        if use_clip and clip is not None:
            return (model, clip)
        else:
            return (model, None)


NODE_CLASS_MAPPINGS = {"RaykoLoRALoader": RaykoLoRALoader}
NODE_DISPLAY_NAME_MAPPINGS = {"RaykoLoRALoader": "🦊 RS LoRA Loader"}


@PromptServer.instance.routes.get("/rayko_lora_loader/get_loras")
async def get_loras(request):
    return aiohttp.web.json_response(sorted(folder_paths.get_filename_list("loras"), key=lambda x: x.lower()))


@PromptServer.instance.routes.post("/rayko_lora_loader/save_preset")
async def rayko_lora_loader_save_preset(request):
    try:
        data = await request.json()
        filepath = _safe_preset_path(data.get("name", ""))
        if not filepath:
            return aiohttp.web.Response(status=400, text="Invalid name")
        with open(filepath, 'w', encoding='utf-8') as f:
            json.dump({"lora_rows": data.get("lora_rows", [])}, f, indent=2)
        return aiohttp.web.Response(status=200, text="OK")
    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/list_presets")
async def rayko_lora_loader_list_presets(request):
    try:
        presets = [f[:-5] for f in os.listdir(LORA_PRESETS_DIR) if f.endswith('.json')] if os.path.exists(LORA_PRESETS_DIR) else []
        return aiohttp.web.json_response(sorted(presets, key=lambda x: x.lower()))
    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/load_preset")
async def rayko_lora_loader_load_preset(request):
    try:
        data = await request.json()
        filepath = _safe_preset_path(data.get("name", ""))
        if not filepath:
            return aiohttp.web.Response(status=400, text="Invalid name")
        if os.path.exists(filepath):
            with open(filepath, 'r', encoding='utf-8') as f:
                return aiohttp.web.json_response(json.load(f))
        return aiohttp.web.Response(status=404, text="Preset not found")
    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/delete_preset")
async def rayko_lora_loader_delete_preset(request):
    try:
        data = await request.json()
        filepath = _safe_preset_path(data.get("name", ""))
        if not filepath:
            return aiohttp.web.Response(status=400, text="Invalid name")
        if os.path.exists(filepath):
            os.remove(filepath)
            return aiohttp.web.Response(status=200, text="OK")
        return aiohttp.web.Response(status=404, text="Preset not found")
    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/get_lora_info")
async def rayko_lora_loader_get_lora_info(request):
    try:
        data = await request.json()
        lora_name = data.get("name", "")
        if not lora_name:
            return aiohttp.web.Response(status=400, text="No name provided")

        lora_relative_path = lora_name.replace("\\", "/")
        lora_full_path = folder_paths.get_full_path("loras", lora_relative_path)

        if not lora_full_path or not os.path.exists(lora_full_path):
            return aiohttp.web.json_response({"error": "File not found"})

        file_hash = compute_sha256(lora_full_path)

        db_path = os.path.join(RAYKO_LORA_DATA_DIR, f"{file_hash}.json")
        if os.path.exists(db_path):
            try:
                with open(db_path, 'r', encoding='utf-8') as f:
                    db_data = json.load(f)
                return aiohttp.web.json_response({
                    "full_name": db_data.get("name", lora_name),
                    "trained_words": db_data.get("trained_words", []),
                    "description": clean_html(db_data.get("description", "")),
                    "source": db_data.get("source", "rayko_db")
                })
            except Exception:
                pass

        safetensors_meta = extract_metadata_from_safetensors(lora_full_path)
        if safetensors_meta and (safetensors_meta["trained_words"] or safetensors_meta["description"]):
            save_to_rayko_db(file_hash, safetensors_meta, "safetensors")
            return aiohttp.web.json_response({
                "full_name": safetensors_meta["full_name"] or lora_name,
                "trained_words": safetensors_meta["trained_words"],
                "description": safetensors_meta["description"],
                "source": "safetensors"
            })

        return aiohttp.web.json_response({
            "full_name": lora_name,
            "trained_words": [],
            "description": "",
            "source": "none"
        })

    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/fetch_civitai_info")
async def rayko_lora_loader_fetch_civitai_info(request):
    try:
        data = await request.json()
        lora_name = data.get("name", "")
        if not lora_name:
            return aiohttp.web.Response(status=400, text="No name provided")

        lora_relative_path = lora_name.replace("\\", "/")
        lora_full_path = folder_paths.get_full_path("loras", lora_relative_path)

        if not lora_full_path or not os.path.exists(lora_full_path):
            return aiohttp.web.json_response({"error": "File not found"})

        file_hash = compute_sha256(lora_full_path)
        civitai_url = f"https://civitai.com/api/v1/model-versions/by-hash/{file_hash}"

        async with aiohttp.ClientSession() as session:
            async with session.get(civitai_url, headers={"User-Agent": "ComfyUI-RaykoStudio/1.0"}) as response:
                if response.status == 404:
                    return aiohttp.web.json_response({"error": "not_found", "message": "Not found on Civitai."})
                elif response.status == 429:
                    return aiohttp.web.json_response({"error": "rate_limit", "message": "Rate limit reached."})
                elif response.status != 200:
                    return aiohttp.web.json_response({"error": "api_error", "message": f"API status {response.status}"})

                civitai_data = await response.json()

            model_id = civitai_data.get("modelId")
            model_tags = []

            if model_id:
                try:
                    model_url = f"https://civitai.com/api/v1/models/{model_id}"
                    async with session.get(model_url, headers={"User-Agent": "ComfyUI-RaykoStudio/1.0"}) as model_response:
                        if model_response.status == 200:
                            model_data = await model_response.json()
                            model_tags = model_data.get("tags", [])
                except Exception as e:
                    print(f"[Rayko] Error fetching full model details: {e}")

        raw_trained = civitai_data.get("trainedWords", [])

        trained_words = []
        if isinstance(raw_trained, list):
            trained_words.extend([str(t).strip() for t in raw_trained if str(t).strip()])
        if isinstance(model_tags, list):
            trained_words.extend([str(t).strip() for t in model_tags if str(t).strip()])

        trained_words = list(dict.fromkeys(trained_words))

        raw_description = civitai_data.get("description", "")
        clean_description = clean_html(raw_description)

        model_name = civitai_data.get("model", {}).get("name", "")
        version_name = civitai_data.get("name", "")

        if model_name:
            if version_name and version_name.lower() not in ["v1", "v1.0", "default", "latest", ""]:
                full_name = f"{model_name} ({version_name})"
            else:
                full_name = model_name
        else:
            full_name = lora_name

        cache_data = {
            "full_name": full_name,
            "trained_words": trained_words,
            "description": clean_description
        }

        save_to_rayko_db(file_hash, cache_data, "civitai", overwrite=True)

        return aiohttp.web.json_response({
            "full_name": full_name,
            "trained_words": trained_words,
            "description": clean_description,
            "source": "civitai"
        })

    except aiohttp.ClientError as e:
        return aiohttp.web.json_response({"error": "network", "message": f"Network error: {str(e)}"})
    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/add_tags")
async def rayko_add_tags(request):
    try:
        data = await request.json()
        lora_name = data.get("name", "")
        new_tags = data.get("tags", [])

        if not lora_name:
            return aiohttp.web.Response(status=400, text="No name provided")
        if not isinstance(new_tags, list) or not new_tags:
            return aiohttp.web.json_response({"error": "No tags provided"})

        lora_relative_path = lora_name.replace("\\", "/")
        lora_full_path = folder_paths.get_full_path("loras", lora_relative_path)

        if not lora_full_path or not os.path.exists(lora_full_path):
            return aiohttp.web.json_response({"error": "File not found"})

        file_hash = compute_sha256(lora_full_path)
        db_path = os.path.join(RAYKO_LORA_DATA_DIR, f"{file_hash}.json")

        existing_data = {}
        if os.path.exists(db_path):
            try:
                with open(db_path, 'r', encoding='utf-8') as f:
                    existing_data = json.load(f)
            except Exception:
                pass

        existing_tags = existing_data.get("trained_words", [])
        if not isinstance(existing_tags, list):
            existing_tags = []

        combined_tags = existing_tags.copy()
        for tag in new_tags:
            tag = str(tag).strip()
            if tag and tag not in combined_tags:
                combined_tags.append(tag)

        save_data = {
            "full_name": existing_data.get("name", ""),
            "trained_words": combined_tags,
            "description": existing_data.get("description", "")
        }

        save_to_rayko_db(file_hash, save_data, "manual", overwrite=True)

        return aiohttp.web.json_response({
            "trained_words": combined_tags,
            "added": len(combined_tags) - len(existing_tags)
        })

    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))


@PromptServer.instance.routes.post("/rayko_lora_loader/remove_tag")
async def rayko_remove_tag(request):
    try:
        data = await request.json()
        lora_name = data.get("name", "")
        tag_to_remove = data.get("tag", "")

        if not lora_name or not tag_to_remove:
            return aiohttp.web.Response(status=400, text="Missing parameters")

        lora_relative_path = lora_name.replace("\\", "/")
        lora_full_path = folder_paths.get_full_path("loras", lora_relative_path)

        if not lora_full_path or not os.path.exists(lora_full_path):
            return aiohttp.web.json_response({"error": "File not found"})

        file_hash = compute_sha256(lora_full_path)
        db_path = os.path.join(RAYKO_LORA_DATA_DIR, f"{file_hash}.json")

        if not os.path.exists(db_path):
            return aiohttp.web.json_response({"error": "No data found"})

        with open(db_path, 'r', encoding='utf-8') as f:
            existing_data = json.load(f)

        existing_tags = existing_data.get("trained_words", [])
        if tag_to_remove in existing_tags:
            existing_tags.remove(tag_to_remove)

        save_data = {
            "full_name": existing_data.get("name", ""),
            "trained_words": existing_tags,
            "description": existing_data.get("description", "")
        }

        save_to_rayko_db(file_hash, save_data, existing_data.get("source", "manual"), overwrite=True)

        return aiohttp.web.json_response({
            "trained_words": existing_tags,
            "removed": tag_to_remove
        })

    except Exception as e:
        return aiohttp.web.Response(status=500, text=str(e))