import pandas as pd
import numpy as np

def find_similar(p_index, similarity_matrix, projects_df, top_x):
    selected_row = similarity_matrix[p_index]
    top_indexes = np.argsort(selected_row)[-top_x:][::-1]
    top_values = selected_row[top_indexes]

    top_projects_df = projects_df.iloc[top_indexes]

    top_projects_df["similarity"] = top_values

    return top_projects_df