import json
from typing import TYPE_CHECKING, Any, Optional, Tuple, Union

import aiohttp  # Add this import
import httpx  # type: ignore

import litellm
import litellm.litellm_core_utils
import litellm.types
import litellm.types.utils
from litellm.llms.base_llm.chat.transformation import BaseConfig
from litellm.llms.custom_httpx.http_handler import (
    AsyncHTTPHandler,
    HTTPHandler,
    _get_httpx_client,
    get_async_httpx_client,
)
from litellm.utils import CustomStreamWrapper, ModelResponse, ProviderConfigManager

if TYPE_CHECKING:
    from litellm.litellm_core_utils.litellm_logging import Logging as _LiteLLMLoggingObj

    LiteLLMLoggingObj = _LiteLLMLoggingObj
else:
    LiteLLMLoggingObj = Any

DEFAULT_TIMEOUT = 600


class BaseLLMAIOHTTPHandler:

    async def _make_common_async_call(
        self,
        async_httpx_client: AsyncHTTPHandler,
        provider_config: BaseConfig,
        api_base: str,
        headers: dict,
        data: dict,
        timeout: Union[float, httpx.Timeout],
        litellm_params: dict,
        stream: bool = False,
    ) -> aiohttp.ClientResponse:
        """Common implementation across stream + non-stream calls. Meant to ensure consistent error-handling."""
        max_retry_on_unprocessable_entity_error = (
            provider_config.max_retry_on_unprocessable_entity_error
        )

        response: Optional[aiohttp.ClientResponse] = None
        timeout_obj = aiohttp.ClientTimeout(
            total=(
                float(timeout) if isinstance(timeout, (int, float)) else DEFAULT_TIMEOUT
            )
        )

        async with aiohttp.ClientSession(timeout=timeout_obj) as session:
            for i in range(max(max_retry_on_unprocessable_entity_error, 1)):
                try:
                    response = await session.post(
                        url=api_base,
                        headers=headers,
                        json=data,
                    )
                    if not response.ok:
                        response.raise_for_status()
                except aiohttp.ClientResponseError as e:
                    raise self._handle_error(e=e, provider_config=provider_config)
                except Exception as e:
                    raise self._handle_error(e=e, provider_config=provider_config)
                break

        if response is None:
            raise provider_config.get_error_class(
                error_message="No response from the API",
                status_code=422,
                headers={},
            )

        return response

    def _make_common_sync_call(
        self,
        sync_httpx_client: HTTPHandler,
        provider_config: BaseConfig,
        api_base: str,
        headers: dict,
        data: dict,
        timeout: Union[float, httpx.Timeout],
        litellm_params: dict,
        stream: bool = False,
    ) -> httpx.Response:

        max_retry_on_unprocessable_entity_error = (
            provider_config.max_retry_on_unprocessable_entity_error
        )

        response: Optional[httpx.Response] = None

        for i in range(max(max_retry_on_unprocessable_entity_error, 1)):
            try:
                response = sync_httpx_client.post(
                    url=api_base,
                    headers=headers,
                    data=json.dumps(data),
                    timeout=timeout,
                    stream=stream,
                )
            except httpx.HTTPStatusError as e:
                hit_max_retry = i + 1 == max_retry_on_unprocessable_entity_error
                should_retry = provider_config.should_retry_llm_api_inside_llm_translation_on_http_error(
                    e=e, litellm_params=litellm_params
                )
                if should_retry and not hit_max_retry:
                    data = (
                        provider_config.transform_request_on_unprocessable_entity_error(
                            e=e, request_data=data
                        )
                    )
                    continue
                else:
                    raise self._handle_error(e=e, provider_config=provider_config)
            except Exception as e:
                raise self._handle_error(e=e, provider_config=provider_config)
            break

        if response is None:
            raise provider_config.get_error_class(
                error_message="No response from the API",
                status_code=422,  # don't retry on this error
                headers={},
            )

        return response

    async def async_completion(
        self,
        custom_llm_provider: str,
        provider_config: BaseConfig,
        api_base: str,
        headers: dict,
        data: dict,
        timeout: Union[float, httpx.Timeout],
        model: str,
        model_response: ModelResponse,
        logging_obj: LiteLLMLoggingObj,
        messages: list,
        optional_params: dict,
        litellm_params: dict,
        encoding: Any,
        api_key: Optional[str] = None,
        client: Optional[AsyncHTTPHandler] = None,
    ):
        if client is None:
            async_httpx_client = get_async_httpx_client(
                llm_provider=litellm.LlmProviders(custom_llm_provider)
            )
        else:
            async_httpx_client = client

        _response = await self._make_common_async_call(
            async_httpx_client=async_httpx_client,
            provider_config=provider_config,
            api_base=api_base,
            headers=headers,
            data=data,
            timeout=timeout,
            litellm_params=litellm_params,
            stream=False,
        )
        _json_response = await _response.json()

        # cast to httpx.Response
        # Todo - use this until we migrate fully to aiohttp
        response = httpx.Response(
            status_code=_response.status,
            headers=_response.headers,
            json=_json_response,
        )
        return provider_config.transform_response(
            model=model,
            raw_response=response,
            model_response=model_response,
            logging_obj=logging_obj,
            api_key=api_key,
            request_data=data,
            messages=messages,
            optional_params=optional_params,
            litellm_params=litellm_params,
            encoding=encoding,
        )

    def completion(
        self,
        model: str,
        messages: list,
        api_base: str,
        custom_llm_provider: str,
        model_response: ModelResponse,
        encoding,
        logging_obj: LiteLLMLoggingObj,
        optional_params: dict,
        timeout: Union[float, httpx.Timeout],
        litellm_params: dict,
        acompletion: bool,
        stream: Optional[bool] = False,
        fake_stream: bool = False,
        api_key: Optional[str] = None,
        headers: Optional[dict] = {},
        client: Optional[Union[HTTPHandler, AsyncHTTPHandler]] = None,
    ):
        provider_config = ProviderConfigManager.get_provider_chat_config(
            model=model, provider=litellm.LlmProviders(custom_llm_provider)
        )
        # get config from model, custom llm provider
        headers = provider_config.validate_environment(
            api_key=api_key,
            headers=headers or {},
            model=model,
            messages=messages,
            optional_params=optional_params,
            api_base=api_base,
        )

        api_base = provider_config.get_complete_url(
            api_base=api_base,
            model=model,
            optional_params=optional_params,
            stream=stream,
        )

        data = provider_config.transform_request(
            model=model,
            messages=messages,
            optional_params=optional_params,
            litellm_params=litellm_params,
            headers=headers,
        )

        ## LOGGING
        logging_obj.pre_call(
            input=messages,
            api_key=api_key,
            additional_args={
                "complete_input_dict": data,
                "api_base": api_base,
                "headers": headers,
            },
        )

        if acompletion is True:
            return self.async_completion(
                custom_llm_provider=custom_llm_provider,
                provider_config=provider_config,
                api_base=api_base,
                headers=headers,
                data=data,
                timeout=timeout,
                model=model,
                model_response=model_response,
                logging_obj=logging_obj,
                api_key=api_key,
                messages=messages,
                optional_params=optional_params,
                litellm_params=litellm_params,
                encoding=encoding,
                client=(
                    client
                    if client is not None and isinstance(client, AsyncHTTPHandler)
                    else None
                ),
            )

        if stream is True:
            if fake_stream is not True:
                data["stream"] = stream
            completion_stream, headers = self.make_sync_call(
                provider_config=provider_config,
                api_base=api_base,
                headers=headers,  # type: ignore
                data=data,
                model=model,
                messages=messages,
                logging_obj=logging_obj,
                timeout=timeout,
                fake_stream=fake_stream,
                client=(
                    client
                    if client is not None and isinstance(client, HTTPHandler)
                    else None
                ),
                litellm_params=litellm_params,
            )
            return CustomStreamWrapper(
                completion_stream=completion_stream,
                model=model,
                custom_llm_provider=custom_llm_provider,
                logging_obj=logging_obj,
            )

        if client is None or not isinstance(client, HTTPHandler):
            sync_httpx_client = _get_httpx_client()
        else:
            sync_httpx_client = client

        response = self._make_common_sync_call(
            sync_httpx_client=sync_httpx_client,
            provider_config=provider_config,
            api_base=api_base,
            headers=headers,
            data=data,
            timeout=timeout,
            litellm_params=litellm_params,
        )
        return provider_config.transform_response(
            model=model,
            raw_response=response,
            model_response=model_response,
            logging_obj=logging_obj,
            api_key=api_key,
            request_data=data,
            messages=messages,
            optional_params=optional_params,
            litellm_params=litellm_params,
            encoding=encoding,
        )

    def make_sync_call(
        self,
        provider_config: BaseConfig,
        api_base: str,
        headers: dict,
        data: dict,
        model: str,
        messages: list,
        logging_obj,
        litellm_params: dict,
        timeout: Union[float, httpx.Timeout],
        fake_stream: bool = False,
        client: Optional[HTTPHandler] = None,
    ) -> Tuple[Any, dict]:
        if client is None or not isinstance(client, HTTPHandler):
            sync_httpx_client = _get_httpx_client()
        else:
            sync_httpx_client = client
        stream = True
        if fake_stream is True:
            stream = False

        response = self._make_common_sync_call(
            sync_httpx_client=sync_httpx_client,
            provider_config=provider_config,
            api_base=api_base,
            headers=headers,
            data=data,
            timeout=timeout,
            litellm_params=litellm_params,
            stream=stream,
        )

        if fake_stream is True:
            completion_stream = provider_config.get_model_response_iterator(
                streaming_response=response.json(), sync_stream=True
            )
        else:
            completion_stream = provider_config.get_model_response_iterator(
                streaming_response=response.iter_lines(), sync_stream=True
            )

        # LOGGING
        logging_obj.post_call(
            input=messages,
            api_key="",
            original_response="first stream response received",
            additional_args={"complete_input_dict": data},
        )

        return completion_stream, dict(response.headers)

    def _handle_error(self, e: Exception, provider_config: BaseConfig):
        status_code = getattr(e, "status_code", 500)
        error_headers = getattr(e, "headers", None)
        error_text = getattr(e, "text", str(e))
        error_response = getattr(e, "response", None)
        if error_headers is None and error_response:
            error_headers = getattr(error_response, "headers", None)
        if error_response and hasattr(error_response, "text"):
            error_text = getattr(error_response, "text", error_text)
        if error_headers:
            error_headers = dict(error_headers)
        else:
            error_headers = {}
        raise provider_config.get_error_class(
            error_message=error_text,
            status_code=status_code,
            headers=error_headers,
        )