import gzip
import json
from collections import Counter

import pandas as pd
import numpy as np
import jax.numpy as jnp
import tqdm

from sentence_transformers import util
from typing import List, Union
import torch

from backend.utils import load_model, filter_questions, load_embeddings
from sklearn.manifold import TSNE

def cos_sim(a, b):
    return jnp.matmul(a, jnp.transpose(b)) / (jnp.linalg.norm(a) * jnp.linalg.norm(b))


# We get similarity between embeddings.
def text_similarity(anchor: str, inputs: List[str], model_name: str, model_dict: dict):
    print(model_name)
    model = load_model(model_name, model_dict)

    # Creating embeddings
    if hasattr(model, 'encode'):
        anchor_emb = model.encode(anchor)[None, :]
        inputs_emb = model.encode(inputs)
    else:
        assert len(model) == 2
        anchor_emb = model[0].encode(anchor)[None, :]
        inputs_emb = model[1].encode(inputs)

    # Obtaining similarity
    similarity = list(jnp.squeeze(cos_sim(anchor_emb, inputs_emb)))

    # Returning a Pandas' dataframe
    d = {'inputs': inputs,
         'score': [round(similarity[i], 3) for i in range(len(similarity))]}
    df = pd.DataFrame(d, columns=['inputs', 'score'])

    return df


# Search
def text_search(anchor: str, n_answers: int, model_name: str, model_dict: dict):
    # Proceeding with model
    print(model_name)
    assert model_name == "distilbert_qa"
    model = load_model(model_name, model_dict)

    # Creating embeddings
    query_emb = model.encode(anchor, convert_to_tensor=True)[None, :]

    print("loading embeddings")
    corpus_emb = load_embeddings()

    # Getting hits
    hits = util.semantic_search(query_emb, corpus_emb, score_function=util.dot_score, top_k=n_answers)[0]

    filtered_posts = filter_questions("python")
    print(f"{len(filtered_posts)} posts found with tag: python")

    hits_titles = []
    hits_scores = []
    urls = []
    for hit in hits:
        post = filtered_posts[hit['corpus_id']]
        hits_titles.append(post['title'])
        hits_scores.append("{:.3f}".format(hit['score']))
        urls.append(f"https://stackoverflow.com/q/{post['id']}")

    return hits_titles, hits_scores, urls


def text_cluster(anchor: str, n_answers: int, model_name: str, model_dict: dict):
    # Proceeding with model
    print(model_name)
    assert model_name == "distilbert_qa"
    model = load_model(model_name, model_dict)

    # Creating embeddings
    query_emb = model.encode(anchor, convert_to_tensor=True)[None, :]

    print("loading embeddings")
    corpus_emb = load_embeddings()

    # Getting hits
    hits = util.semantic_search(query_emb, corpus_emb, score_function=util.dot_score, top_k=n_answers)[0]

    filtered_posts = filter_questions("python")

    hits_dict = [filtered_posts[hit['corpus_id']] for hit in hits]
    hits_dict.append(dict(id = '1', title = anchor, tags = ['']))

    hits_emb = torch.stack([corpus_emb[hit['corpus_id']] for hit in hits])
    hits_emb = torch.cat((hits_emb, query_emb))

    # Dimensionality reduction with t-SNE
    tsne = TSNE(n_components=3, verbose=1, perplexity=15, n_iter=1000)
    tsne_results = tsne.fit_transform(hits_emb.cpu())
    df = pd.DataFrame(hits_dict)
    tags = list(df['tags'])

    counter = Counter(tags[0])
    for i in tags[1:]:
        counter.update(i)

    df_tags = pd.DataFrame(counter.most_common(), columns=['Tag', 'Mentions'])
    most_common_tags = list(df_tags['Tag'])[1:5]

    labels = []

    for tags_list in list(df['tags']):
        for common_tag in most_common_tags:
            if common_tag in tags_list:
                labels.append(common_tag)
                break
            elif common_tag != most_common_tags[-1]:
                continue
            else:
                labels.append('others')

    df['title'] = [post['title'] for post in hits_dict]
    df['labels'] = labels
    df['tsne_x'] = tsne_results[:, 0]
    df['tsne_y'] = tsne_results[:, 1]
    df['tsne_z'] = tsne_results[:, 2]

    df['size'] = [2 for i in range(len(df))]

    # Making the query bigger than the rest of the observations
    df['size'][len(df) - 1] = 10
    df['labels'][len(df) - 1] = 'QUERY'
    import plotly.express as px

    fig = px.scatter_3d(df, x='tsne_x', y='tsne_y', z='tsne_z', color='labels', size='size',
                        color_discrete_sequence=px.colors.qualitative.D3, hover_data=[df.title])
    return fig



# We get similarity between embeddings.
def tweets_vaccine(anchor: str, model_name: str, model_dict: dict):
    print(model_name)
    model = load_model(model_name, model_dict)

    # Keywords common in disinformation tweets
    keywords = '''abolish big pharma,
    no forced flu shots,
    antivaccine,
    No Forced Vaccines,
    Arrest Bill Gates,
    not mandatory vaccines,
    No Vaccine,
    big pharma mafia,
    No Vaccine For Me,
    big pharma kills,
    no vaccine mandates,
    parents over pharma,
    say no to vaccines,
    stop mandatory vaccination,
    vaccines are poison,
    learn the risk,
    vaccines cause,
    medical freedom,
    vaccines kill,
    medical freedom of choice,
    vaxxed,
    my body my choice,
    vaccines have very dangerous consequences,
    Vaccines harm your organism'''




    # Creating embeddings
    if hasattr(model, 'encode'):
        anchor_emb = model.encode(anchor)[None, :]
        inputs_emb = model.encode(keywords)
    else:
        assert len(model) == 2
        anchor_emb = model[0].encode(anchor)[None, :]
        inputs_emb = model[1].encode(keywords)


    # Obtaining similarity
    similarity = jnp.squeeze(jnp.matmul(anchor_emb, jnp.transpose(inputs_emb)) / (jnp.linalg.norm(anchor_emb) * jnp.linalg.norm(inputs_emb))).tolist()

    # Returning a Pandas' dataframe
    d = dict(tweet = anchor,
             score = [round(similarity, 3)])
    df = pd.DataFrame(d, columns=['tweet', 'score'])


    return df