use serde::{Deserialize, Serialize};

use crate::dto::api;

#[derive(Debug, Serialize, Deserialize)]
pub struct OllamaModels {
    pub models: Vec<OllamaModel>,
}

#[derive(Debug, Serialize, Deserialize)]
pub struct OllamaModel {
    pub name: String,

    pub details: Option<OllamaModelDetails>,

    pub size: Option<u64>,
    pub digest: Option<String>,
    pub modified_at: Option<String>,
}

#[derive(Debug, Serialize, Deserialize)]
pub struct OllamaModelDetails {
    pub family: Option<String>,
    pub parameter_size: Option<String>,
    pub quantization_level: Option<String>,
}

#[derive(Debug, Serialize, Deserialize)]
pub struct OllamaOptions {
    pub temperature: Option<f32>,
    pub top_p: Option<f32>,
    pub top_k: Option<u32>,
    pub repeat_penalty: Option<f32>,
    pub seed: Option<i64>,

    pub num_ctx: Option<u32>,
    pub num_predict: Option<u32>,
}

#[derive(Debug, Serialize)]
pub struct OllamaGenerateRequest<'a> {
    pub model: &'a str,
    pub prompt: &'a str,
    pub stream: bool,
    pub options: OllamaOptions,
}

#[derive(Debug, Serialize, Deserialize)]
pub struct OllamaGenerateResponse {
    pub model: String,
    pub created_at: Option<String>,
    pub response: String,
    pub done: bool,

    #[serde(default)]
    pub context: Option<Vec<u64>>,

    pub total_duration: Option<u64>,
    pub load_duration: Option<u64>,
    pub prompt_eval_count: Option<u32>,
    pub eval_count: Option<u32>,
}

#[derive(Debug, Serialize)]
pub struct OllamaChatRequest<'a> {
    pub model: &'a str,
    pub messages: &'a [api::Message],
    pub stream: bool,
    pub options: OllamaOptions,
}

#[derive(Debug, Deserialize)]
pub struct OllamaChatResponse {
    pub model: String,
    pub message: api::Message,
    pub done: bool,

    pub prompt_eval_count: Option<u32>,
    pub eval_count: Option<u32>,
}