import streamlit as st
import pandas as pd
import numpy as np
from streamlit_echarts import st_echarts
# from streamlit_echarts import JsCode
from streamlit_javascript import st_javascript
# from PIL import Image 

links_dic = {"random": "https://seaeval.github.io/", 
             "meta_llama_3_8b": "https://huggingface.co/meta-llama/Meta-Llama-3-8B", 
             "mistral_7b_instruct_v0_2": "https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2", 
             "sailor_0_5b": "https://huggingface.co/sail/Sailor-0.5B", 
             "sailor_1_8b": "https://huggingface.co/sail/Sailor-1.8B", 
             "sailor_4b": "https://huggingface.co/sail/Sailor-4B", 
             "sailor_7b": "https://huggingface.co/sail/Sailor-7B", 
             "sailor_0_5b_chat": "https://huggingface.co/sail/Sailor-0.5B-Chat", 
             "sailor_1_8b_chat": "https://huggingface.co/sail/Sailor-1.8B-Chat", 
             "sailor_4b_chat": "https://huggingface.co/sail/Sailor-4B-Chat", 
             "sailor_7b_chat": "https://huggingface.co/sail/Sailor-7B-Chat", 
             "sea_mistral_highest_acc_inst_7b": "https://seaeval.github.io/", 
             "meta_llama_3_8b_instruct": "https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct", 
             "flan_t5_base": "https://huggingface.co/google/flan-t5-base", 
             "flan_t5_large": "https://huggingface.co/google/flan-t5-large", 
             "flan_t5_xl": "https://huggingface.co/google/flan-t5-xl", 
             "flan_t5_xxl": "https://huggingface.co/google/flan-t5-xxl", 
             "flan_ul2": "https://huggingface.co/google/flan-t5-ul2", 
             "flan_t5_small": "https://huggingface.co/google/flan-t5-small", 
             "mt0_xxl": "https://huggingface.co/bigscience/mt0-xxl", 
             "seallm_7b_v2": "https://huggingface.co/SeaLLMs/SeaLLM-7B-v2", 
             "gpt_35_turbo_1106": "https://openai.com/blog/chatgpt", 
             "meta_llama_3_70b": "https://huggingface.co/meta-llama/Meta-Llama-3-70B", 
             "meta_llama_3_70b_instruct": "https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct", 
             "sea_lion_3b": "https://huggingface.co/aisingapore/sea-lion-3b", 
             "sea_lion_7b": "https://huggingface.co/aisingapore/sea-lion-7b", 
             "qwen1_5_110b": "https://huggingface.co/Qwen/Qwen1.5-110B", 
             "qwen1_5_110b_chat": "https://huggingface.co/Qwen/Qwen1.5-110B-Chat", 
             "llama_2_7b_chat": "https://huggingface.co/meta-llama/Llama-2-7b-chat-hf", 
             "gpt4_1106_preview": "https://openai.com/blog/chatgpt", 
             "gemma_2b": "https://huggingface.co/google/gemma-2b", 
             "gemma_7b": "https://huggingface.co/google/gemma-7b", 
             "gemma_2b_it": "https://huggingface.co/google/gemma-2b-it", 
             "gemma_7b_it": "https://huggingface.co/google/gemma-7b-it", 
             "qwen_1_5_7b": "https://huggingface.co/Qwen/Qwen1.5-7B", 
             "qwen_1_5_7b_chat": "https://huggingface.co/Qwen/Qwen1.5-7B-Chat", 
             "sea_lion_7b_instruct": "https://huggingface.co/aisingapore/sea-lion-7b-instruct", 
             "sea_lion_7b_instruct_research": "https://huggingface.co/aisingapore/sea-lion-7b-instruct-research", 
             "LLaMA_3_Merlion_8B": "https://seaeval.github.io/", 
             "LLaMA_3_Merlion_8B_v1_1": "https://seaeval.github.io/"}

links_dic = {k.lower().replace('_', '-') : v for k, v in links_dic.items()}

# huggingface_image = Image.open('style/huggingface.jpg')

def nav_to(value):
    try:
        url = links_dic[str(value).lower()]
        js = f'window.open("{url}", "_blank").then(r => window.parent.location.href);'
        st_javascript(js)
    except:
        pass

def draw(folder_name,category_name, dataset_name, sorted):
    
    folder = f"./results/{folder_name}/"

    display_names = {
        'ASR': 'Automatic Speech Recognition',
        'SQA': 'Speech Question Answering',
        'SI': 'Speech Instruction',
        'AC': 'Audio Captioning',
        'ASQA': 'Audio Scene Question Answering',
        'AR': 'Accent Recognition',
        'GR': 'Gender Recognition',
        'ER': 'Emotion Recognition'
    }
    
    data_path = f'{folder}/{category_name.lower()}.csv'
    chart_data = pd.read_csv(data_path).round(2).dropna(axis=0)

    if len(chart_data) == 0:
        return


    if sorted == 'Ascending':
        ascend = True 
    else:
        ascend = False

    sort_by = dataset_name.replace('-', '_').lower()
    
    chart_data = chart_data.sort_values(by=[sort_by], ascending=ascend)
    
    min_value = round(chart_data.iloc[:, 1::].min().min() - 0.1, 1) 
    max_value = round(chart_data.iloc[:, 1::].max().max() + 0.1, 1) 

    columns = list(chart_data.columns)[1:]
    series = []
    for col in columns:
        series.append(
            {
                "name": f"{col.replace('_', '-')}",
                "type": "line",
                "data": chart_data[f'{col}'].tolist(),
            }
            )
        

    options = {
        "title": {"text": f"{display_names[category_name]}"},
        "tooltip": {
            "trigger": "axis",
            "axisPointer": {"type": "cross", "label": {"backgroundColor": "#6a7985"}},
            "triggerOn": 'mousemove',
        },
        "legend": {"data": ['Overall Accuracy']},
        "toolbox": {"feature": {"saveAsImage": {}}},
        "grid": {"left": "3%", "right": "4%", "bottom": "3%", "containLabel": True},
        "xAxis": [
            {
                "type": "category",
                "boundaryGap": False,
                "triggerEvent": True,
                "data": chart_data['Model'].tolist(),
            }
        ],
        "yAxis": [{"type": "value", 
                    "min": min_value,
                    "max": max_value, 
                    # "splitNumber": 10
                    }],
        "series": series,
    }
    
    events = {
        "click": "function(params) { return params.value }"
    }

    value = st_echarts(options=options, events=events, height="500px")
    
    if value != None:
        # print(value)
        nav_to(value)

    # if value != None:
    #     highlight_table_line(value)

    ### create table
    st.divider()
    # chart_data['Link'] = chart_data['Model'].map(links_dic)
    st.dataframe(chart_data,
                #  column_config = {
                #      "Link": st.column_config.LinkColumn(
                #          display_text= st.image(huggingface_image)
                #      ),
                #  }, 
                    hide_index = True, 
                    use_container_width=True)