plots.py

import streamlit as st
import matplotlib.pyplot as plt
import numpy as np
from mpl_toolkits.mplot3d import Axes3D
import umap
import pandas as pd
from word2vec import *
from sklearn.preprocessing import StandardScaler


def make_3d_plot(new_3d_vectors):
    """
    Turn DataFrame of 3D vectors into a 3D plot
    DataFrame structure: ['word', 'cosine_sim', '3d_vector']
    """
    fig = plt.figure()
    ax = fig.add_subplot(projection='3d')
    
    plt.ion()

    # Unpack vectors and labels from DataFrame
    labels = new_3d_vectors['word']
    x = new_3d_vectors['3d_vector'].apply(lambda v: v[0])
    y = new_3d_vectors['3d_vector'].apply(lambda v: v[1])
    z = new_3d_vectors['3d_vector'].apply(lambda v: v[2])

    # Plot points
    ax.scatter(x, y, z)

    # Add labels
    for i, label in enumerate(labels):
        ax.text(x[i], y[i], z[i], label)

    # Set labels and title
    ax.set_xlabel('X')
    ax.set_ylabel('Y')
    ax.set_zlabel('Z')
    ax.set_title('3D plot of word vectors')

    return fig


import plotly.express as px


def make_3d_plot2(df):
    """
        Turn DataFrame of 3D vectors into a 3D plot using plotly
        DataFrame structure: ['word', 'cosine_sim', '3d_vector']
    """
    vectors = df['3d_vector'].tolist()
    fig = px.scatter_3d(df, x=[v[0] for v in vectors], y=[v[1] for v in vectors], z=[v[2] for v in vectors], text=df['word'])
    return fig


def make_3d_plot3(vectors_list, word, time_slice_model):
    """
    Turn list of 100D vectors into a 3D plot using UMAP and Plotly.
    List structure: [(word, model_name, vector, cosine_sim)]
    """
    # Load model
    model = load_word2vec_model(f'models/{time_slice_model}.model')
    
    # Make UMAP model and fit it to the vectors
    umap_model = umap.UMAP(n_components=3)
    umap_model.fit(model.wv.vectors)
    
    # Transform the vectors to 3D
    transformed_vectors = umap_model.transform(model.wv.vectors)
    
    
    # Create DataFrame from the transformed vectors
    df = pd.DataFrame(transformed_vectors, columns=['x', 'y', 'z'])
    
    # Add word and cosine similarity to DataFrame
    df['word'] = model.wv.index_to_key
    
    # Filter the DataFrame for words in vectors_list and add cosine similarity
    word_list = [v[0] for v in vectors_list]
    cosine_sim_list = [v[3] for v in vectors_list]
    
    # Ensure that the word list and cosine similarity list are aligned properly
    df = df[df['word'].isin(word_list)]
    df['cosine_sim'] = cosine_sim_list
    
    # Create plot
    fig = px.scatter_3d(df, x='x', y='y', z='z', text='word', color='cosine_sim', color_continuous_scale='Reds')
    fig.update_traces(marker=dict(size=5))
    fig.update_layout(title=f'3D plot of nearest neighbours to {word}')
    
    return fig, df



def make_3d_plot4(vectors_list, word, time_slice_model):
    """
    Turn list of 100D vectors into a 3D plot using UMAP and Plotly.
    List structure: [(word, model_name, vector, cosine_sim)]
    """
    # Load model
    model = load_word2vec_model(f'models/{time_slice_model}.model')
    model_dict = model_dictionary(model)
    
    
    # Extract vectors and names from model_dict
    all_vector_names = list(model_dict.keys())
    all_vectors = list(model_dict.values())

    
    # Scale the vectors
    scaler = StandardScaler()
    vectors_scaled = scaler.fit_transform(all_vectors)
    
    # Make UMAP model and fit it to the scaled vectors
    umap_model = umap.UMAP(n_components=3)
    umap_result = umap_model.fit_transform(vectors_scaled)
    
    # Now umap_result contains the 3D representations of the vectors
    # Associate the names with the 3D representations
    result_with_names = [(all_vector_names[i], umap_result[i]) for i in range(len(all_vector_names))]
    
    
    # Only keep the vectors that are in vectors_list and their cosine similarities
    result_with_names = [r for r in result_with_names if r[0] in [v[0] for v in vectors_list]]
    result_with_names = [(r[0], r[1], [v[3] for v in vectors_list if v[0] == r[0]][0]) for r in result_with_names]
    
    
    # Create DataFrame from the transformed vectors
    df = pd.DataFrame(result_with_names, columns=['word', '3d_vector', 'cosine_sim'])
    
    # Sort dataframe by cosine_sim
    df = df.sort_values(by='cosine_sim', ascending=False)
    
    x = df['3d_vector'].apply(lambda v: v[0])
    y = df['3d_vector'].apply(lambda v: v[1])
    z = df['3d_vector'].apply(lambda v: v[2])
    
    
    # Create plot
    fig = px.scatter_3d(df, x=x, y=y, z=z, text='word', color='cosine_sim', color_continuous_scale='Reds')
    fig.update_traces(marker=dict(size=5))
    fig.update_layout(title=f'3D plot of nearest neighbours to {word}')
    
    return fig, df