import pandas as pd
import json

custom_css = """
@import url('https://fonts.googleapis.com/css2?family=Vazirmatn&display=swap');
body, .gradio-container, .gr-button, .gr-input, .gr-slider, .gr-dropdown, .gr-markdown {
    font-family: 'Vazirmatn', sans-serif !important;
}

.markdown-text {
    font-size: 16px !important;
}

#models-to-add-text {
    font-size: 18px !important;
}

#citation-button span {
    font-size: 16px !important;
}

#citation-button textarea {
    font-size: 16px !important;
}

#citation-button > label > button {
    margin: 6px;
    transform: scale(1.3);
}

#leaderboard-table {
    margin-top: 15px
}

#leaderboard-table-lite {
    margin-top: 15px
}

#search-bar-table-box > div:first-child {
    background: none;
    border: none;
}

#search-bar {
    padding: 0px;
}

/* Limit the width of the first AutoEvalColumn so that names don't expand too much */
#leaderboard-table td:nth-child(2),
#leaderboard-table th:nth-child(2) {
    max-width: 400px;
    overflow: auto;
    white-space: nowrap;
}

.tab-buttons button {
    font-size: 20px;
}

#scale-logo {
    border-style: none !important;
    box-shadow: none;
    display: block;
    margin-left: auto;
    margin-right: auto;
    max-width: 600px;
}

#scale-logo .download {
    display: none;
}
#filter_type{
    border: 0;
    padding-left: 0;
    padding-top: 0;
}
#filter_type label {
    display: flex;
}
#filter_type label > span{
    margin-top: var(--spacing-lg);
    margin-right: 0.5em;
}
#filter_type label > .wrap{
    width: 103px;
}
#filter_type label > .wrap .wrap-inner{  
    padding: 2px;
}
#filter_type label > .wrap .wrap-inner input{
    width: 1px
}
#filter-columns-type{
    border:0;
    padding:0.5;
}
#filter-columns-size{
    border:0;
    padding:0.5;
}
#box-filter > .form{
    border: 0
}
"""

LLM_BENCHMARKS_ABOUT_TEXT = f"""
## How it works

## Reproducibility
To reproduce our results, here is the commands you can run:

"""


LLM_BENCHMARKS_SUBMIT_TEXT = """## Submit your model
                Submit your model to the leaderboard using the below form AFTER following the following steps: 
                - Create a HuggingFace account and request to join the  [MLSB organization](https://huggingface.co/MLSB)
                - Create a new space in the MLSB organization and add your model using the inference templates: https://huggingface.co/new-space?owner=MLSB
                - Fill the submission form. 
    
                ## Prerequisites:
                To qualify for submission, each team must:
                - Provide an MLSB submission ID (find it on CMT) or a link to a preprint/paper describing their methodology. This publication does not have to specifically report training or evaluation on the P(L)INDER dataset. Previously published methods, such as DiffDock, only need to link their existing paper. Note that entry into this competition does not equate to an MLSB workshop paper submission.
                - Create a copy of the provided [inference templates](https://huggingface.co/MLSB/).
                    - Go to the top right corner of the page of the respective inference template and click on the drop-down menu (vertical ellipsis) right next to the “Community”, then select “Duplicate this space”.
                - Change files in the newly create space to reflect the peculiarities of your model
                    - Edit `requirements.txt` to capture all python dependencies.
                    - Modify the Dockerfile as appropriate (including selecting the right base image)
                    - Include a `inference_app.py` file. This contains a `predict` function that should be modified to reflect the specifics of inference using their model.
                    - Include a `train.py` file to ensure that training and model selection use only the Persian/Base datasets and to clearly show any additional hyperparameters used.
                    - Provide a LICENSE file that allows for reuse, derivative works, and distribution of the provided software and weights (e.g., MIT or Apache2 license).
                - Submit to the leaderboard via the [form below](https://huggingface.co/spaces/MLSB/leaderboard2024).
                    - On submission page, add reference to the newly created space in the format username/space (e.g mlsb/alphafold3). You can create the space on your personal Huggingface account and transfer it to MLSB for the submission to get a GPU assigned. 

                After a brief technical review by our organizers we will grant you a free GPU until MLSB so that anyone can play with the model and we will run the evaluation. 

                If you have a questions please email: workshopmlsb@gmail.com
                """


def load_jsonl(input_file):
    data = []
    with open(input_file, 'r') as f:
        for line in f:
            data.append(json.loads(line))
    return data


def jsonl_to_dataframe(input_file):
    data = load_jsonl(input_file)
    return pd.DataFrame(data)


def add_average_column_to_df(df,columns_to_average, index=3):
    average_column = df[columns_to_average].mean(axis=1)
    df.insert(index, "Average ⬆️", average_column)
    return df


def model_hyperlink(link, model_name):
    return f'<a target="_blank" href="{link}" style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">{model_name}</a>'


def make_clickable_model(model_name):
    link = f"https://huggingface.co/{model_name}"
    return model_hyperlink(link, model_name)


def apply_clickable_model(df, column_name):
    df[column_name] = df[column_name].apply(make_clickable_model)
    return df