import http.client as http_client
import json
import logging
import os
import re
import string
import traceback

import gradio as gr
import requests
from huggingface_hub import HfApi

hf_api = HfApi()
#roots_datasets = {dset.id.split("/")[-1]:dset for dset in hf_api.list_datasets(author="bigscience-data", use_auth_token=os.environ.get("bigscience_data_token"))}
roots_datasets = {dset.id.split("/")[-1]:dset for dset in hf_api.list_datasets(author="bigscience-data", use_auth_token=os.environ.get("HF_TOKEN"))}

def get_docid_html(docid):
    data_org, dataset, docid = docid.split("/")
    metadata = roots_datasets[dataset]
    if metadata.private:
        docid_html = (
            f"<a "
            f'class="underline-on-hover"'
            f'title="This dataset is private. See the introductory text for more information"'
            f'style="color:#AA4A44;"'
            f'href="https://huggingface.co/datasets/bigscience-data/{dataset}"'
            f'target="_blank"><b>🔒{dataset}</b></a><span style="color: #7978FF;">/{docid}</span>'
        )
    else:
        docid_html = (
            f"<a "
            f'class="underline-on-hover"'
            f'title="This dataset is licensed {metadata.tags[0].split(":")[-1]}"'
            f'style="color:#2D31FA;"'
            f'href="https://huggingface.co/datasets/bigscience-data/{dataset}"'
            f'target="_blank"><b>{dataset}</b></a><span style="color: #7978FF;">/{docid}</span>'
        )        
    return docid_html


PII_TAGS = {"KEY", "EMAIL", "USER", "IP_ADDRESS", "ID", "IPv4", "IPv6"}
PII_PREFIX = "PI:"


def process_pii(text):
    for tag in PII_TAGS:
        text = text.replace(
            PII_PREFIX + tag,
            """<b><mark style="background: Fuchsia; color: Lime;">REDACTED {}</mark></b>""".format(tag),
        )
    return text


def process_results(results, highlight_terms):
    if len(results) == 0:
        return """<br><p style='font-family: Arial; color:Silver; text-align: center;'>
                No results retrieved.</p><br><hr>"""

    results_html = ""
    for result in results:
        tokens = result["text"].split()
        tokens_html = []
        for token in tokens:
            if token in highlight_terms:
                tokens_html.append("<b>{}</b>".format(token))
            else:
                tokens_html.append(token)
        tokens_html = " ".join(tokens_html)
        tokens_html = process_pii(tokens_html)
        meta_html = (
            """
                <p class='underline-on-hover' style='font-size:12px; font-family: Arial; color:#585858; text-align: left;'>
                <a href='{}' target='_blank'>{}</a></p>""".format(
                result["meta"]["url"], result["meta"]["url"]
            )
            if "meta" in result and result["meta"] is not None and "url" in result["meta"]
            else ""
        )
        docid_html = get_docid_html(result["docid"])
        results_html += """{}
            <p style='font-size:14px; font-family: Arial; color:#7978FF; text-align: left;'>Document ID: {}</p>
            <p style='font-size:12px; font-family: Arial; color:MediumAquaMarine'>Language: {}</p>
            <p style='font-family: Arial;'>{}</p>
            <br>
        """.format(
            meta_html, docid_html, result["lang"], tokens_html
        )
    return results_html + "<hr>"


def scisearch(query, language, num_results=10):
    try:
        query = " ".join(query.split())
        if query == "" or query is None:
            return ""

        post_data = {"query": query, "k": num_results}
        if language != "detect_language":
            post_data["lang"] = language

        output = requests.post(
            os.environ.get("address"),
            headers={"Content-type": "application/json"},
            data=json.dumps(post_data),
            timeout=60,
        )

        payload = json.loads(output.text)

        if "err" in payload:
            if payload["err"]["type"] == "unsupported_lang":
                detected_lang = payload["err"]["meta"]["detected_lang"]
                return f"""
                    <p style='font-size:18px; font-family: Arial; color:MediumVioletRed; text-align: center;'>
                    Detected language <b>{detected_lang}</b> is not supported.<br>
                    Please choose a language from the dropdown or type another query.
                    </p><br><hr><br>"""

        results = payload["results"]
        highlight_terms = payload["highlight_terms"]

        if language == "detect_language":
            results = list(results.values())[0]
            return (
                (
                    f"""<p style='font-family: Arial; color:MediumAquaMarine; text-align: center; line-height: 3em'>
                Detected language: <b>{results[0]["lang"]}</b></p><br><hr><br>"""
                    if len(results) > 0 and language == "detect_language"
                    else ""
                )
                + process_results(results, highlight_terms)
            )

        if language == "all":
            results_html = ""
            for lang, results_for_lang in results.items():
                if len(results_for_lang) == 0:
                    results_html += f"""<p style='font-family: Arial; color:Silver; text-align: left; line-height: 3em'>
                            No results for language: <b>{lang}</b><hr></p>"""
                    continue

                collapsible_results = f"""
                    <details>
                        <summary style='font-family: Arial; color:MediumAquaMarine; text-align: left; line-height: 3em'>
                            Results for language: <b>{lang}</b><hr>
                        </summary>
                        {process_results(results_for_lang, highlight_terms)}
                    </details>"""
                results_html += collapsible_results
            return results_html

        results = list(results.values())[0]
        return process_results(results, highlight_terms)

    except Exception as e:
        results_html = f"""
                <p style='font-size:18px; font-family: Arial; color:MediumVioletRed; text-align: center;'>
                Raised {type(e).__name__}</p>
                <p style='font-size:14px; font-family: Arial; '>
                Check if a relevant discussion already exists in the Community tab. If not, please open a discussion.
                </p>
            """
        print(e)
        print(traceback.format_exc())

    return results_html


def flag(query, language, num_results, issue_description):
    try:
        post_data = {"query": query, "k": num_results, "flag": True, "description": issue_description}
        if language != "detect_language":
            post_data["lang"] = language

        output = requests.post(
            os.environ.get("address"),
            headers={"Content-type": "application/json"},
            data=json.dumps(post_data),
            timeout=120,
        )

        results = json.loads(output.text)
    except:
        print("Error flagging")
    return ""


description = """# <p style="text-align: center;"> 🌸 🔎 Bloom Searcher 🔍 🌸 </p>
Tool design for Roots: [URL](https://huggingface.co/spaces/bigscience-data/scisearch/blob/main/roots_search_tool_specs.pdf).
Bloom on Wikipedia: [URL](https://en.wikipedia.org/wiki/BLOOM_(language_model)).
Bloom Video Playlist: [URL](https://www.youtube.com/playlist?list=PLHgX2IExbFouqnsIqziThlPCX_miiDq14).
Access full corpus check [URL](https://forms.gle/qyYswbEL5kA23Wu99).

Big Science - How to get started
Big Science is a 176B parameter new ML model that was trained on a set of datasets for Natural Language processing, and many other tasks that are not yet explored.. Below is the set of the papers, models, links, and datasets around big science which promises to be the best, most recent large model of its kind benefitting all science pursuits.

Model: https://huggingface.co/bigscience/bloom
Papers:
BLOOM: A 176B-Parameter Open-Access Multilingual Language Model https://arxiv.org/abs/2211.05100
Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism https://arxiv.org/abs/1909.08053
8-bit Optimizers via Block-wise Quantization https://arxiv.org/abs/2110.02861
Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation https://arxiv.org/abs/2108.12409
https://huggingface.co/models?other=doi:10.57967/hf/0003
217 Other Models optimizing use of bloom via specialization: https://huggingface.co/models?other=bloom
Datasets
Universal Dependencies: https://paperswithcode.com/dataset/universal-dependencies
WMT 2014: https://paperswithcode.com/dataset/wmt-2014
The Pile: https://paperswithcode.com/dataset/the-pile
HumanEval: https://paperswithcode.com/dataset/humaneval
FLORES-101: https://paperswithcode.com/dataset/flores-101
CrowS-Pairs: https://paperswithcode.com/dataset/crows-pairs
WikiLingua: https://paperswithcode.com/dataset/wikilingua
MTEB: https://paperswithcode.com/dataset/mteb
xP3: https://paperswithcode.com/dataset/xp3
DiaBLa: https://paperswithcode.com/dataset/diabla

"""


if __name__ == "__main__":
    demo = gr.Blocks(
        css=".underline-on-hover:hover { text-decoration: underline; } .flagging { font-size:12px; color:Silver; }"
    )

    with demo:
        with gr.Row():
            gr.Markdown(value=description)
        with gr.Row():
            query = gr.Textbox(lines=1, max_lines=1, placeholder="Type your query here...", label="Query")
        with gr.Row():
            lang = gr.Dropdown(
                choices=[
                    "ar",
                    "ca",
                    "code",
                    "en",
                    "es",
                    "eu",
                    "fr",
                    "id",
                    "indic",
                    "nigercongo",
                    "pt",
                    "vi",
                    "zh",
                    "detect_language",
                    "all",
                ],
                value="en",
                label="Language",
            )
        with gr.Row():
            k = gr.Slider(1, 100, value=10, step=1, label="Max Results")
        with gr.Row():
            submit_btn = gr.Button("Submit")
        with gr.Row():
            results = gr.HTML(label="Results")
        flag_description = """
            <p class='flagging'>
            If you choose to flag your search, we will save the query, language and the number of results you requested.
            Please consider adding any additional context in the box on the right.</p>"""
        with gr.Column(visible=False) as flagging_form:
            flag_txt = gr.Textbox(
                lines=1,
                placeholder="Type here...",
                label="""If you choose to flag your search, we will save the query, language and the number of results
                    you requested. Please consider adding relevant additional context below:""",
            )
            flag_btn = gr.Button("Flag Results")
            flag_btn.click(flag, inputs=[query, lang, k, flag_txt], outputs=[flag_txt])

        def submit(query, lang, k):
            query = query.strip()
            if query is None or query == "":
                return "", ""
            return {
                results: scisearch(query, lang, k),
                flagging_form: gr.update(visible=True),
            }

        query.submit(fn=submit, inputs=[query, lang, k], outputs=[results, flagging_form])
        submit_btn.click(submit, inputs=[query, lang, k], outputs=[results, flagging_form])

    demo.launch(enable_queue=True, debug=True)