from bokeh import mpl
from math import pi
import seaborn as sns
import pandas as pd
import numpy as np
np.random.seed(0)

from bokeh.io import curdoc
from bokeh.charts import Bar
from bokeh.models.widgets import Panel, Tabs
from bokeh.layouts import layout, widgetbox, row, column
from bokeh.models import ColumnDataSource, Select, Slider
from bokeh.charts.attributes import ColorAttr, CatAttr
from bokeh.plotting import figure
from bokeh.palettes import Spectral6

from sklearn import cluster, datasets
from sklearn.neighbors import kneighbors_graph
from sklearn.preprocessing import StandardScaler

data_ngram1 = pd.read_csv('/Users/degravek/Downloads/sorted_ngram1.csv')
data_ngram2 = pd.read_csv('/Users/degravek/Downloads/sorted_ngram2.csv')
data_ngram3 = pd.read_csv('/Users/degravek/Downloads/sorted_ngram3.csv')
data_chunk = pd.read_csv('/Users/degravek/Downloads/sorted_chunk.csv')
data_rake = pd.read_csv('/Users/degravek/Downloads/sorted_rake.csv')

# define some helper functions
def get_dataset(dataset, n_samples):
    if dataset == 'n-gram 1':
        asdata = data_ngram1.copy()
        asdata = asdata.head(n_samples).append(asdata.tail(n_samples))
    elif dataset == 'n-gram 2':
        asdata = data_ngram2.copy()
        asdata = asdata.head(n_samples).append(asdata.tail(n_samples))
    elif dataset == 'n-gram 3':
        asdata = data_ngram3.copy()
        asdata = asdata.head(n_samples).append(asdata.tail(n_samples))
    elif dataset == 'chunk':
        asdata = data_chunk.copy()
        asdata = asdata.head(n_samples).append(asdata.tail(n_samples))
    elif dataset == 'rake':
        asdata = data_rake.copy()
        asdata = asdata.head(n_samples).append(asdata.tail(n_samples))
    return asdata

# set up initial data
n_samples = 15
dataset = 'n-gram 1'

asdata = get_dataset(dataset, n_samples)

# set up plot (styling in theme.yaml)
source = ColumnDataSource(data=dict(x=asdata['aspects'].tolist(), y=asdata['importance'].values))

def mplot(source):
    p = pd.DataFrame()
    p['aspects'] = source.data['x']
    p['importance'] = source.data['y']
    plot = Bar(p, values='importance', label='aspects', legend=False)
    return plot

plot = mplot(source)

datasets_names = [
    'n-gram 1',
    'n-gram 2',
    'n-gram 3',
    'chunk',
    'rake'
]

dataset_select = Select(value='n-gram 1',
                        title='Select dataset:',
                        width=200,
                        options=datasets_names)

samples_slider = Slider(title="Number of aspects",
                        value=15,
                        start=1,
                        end=30,
                        step=1,
                        width=200)

def update_samples_or_dataset(attrname, old, new):
    dataset = dataset_select.value
    n_samples = int(samples_slider.value)

    asdata = get_dataset(dataset, n_samples)
    plot.x_range.factors = asdata['aspects'].tolist() # this was missing
    source.data = dict(x=asdata['aspects'].tolist(), y=asdata['importance'].values)

dataset_select.on_change('value', update_samples_or_dataset)
samples_slider.on_change('value', update_samples_or_dataset)

# set up layout
selects = row(dataset_select, width=420)
inputs = column(selects, widgetbox(samples_slider))

# add to document
curdoc().add_root(row(inputs, plot))
curdoc().title = "Clustering"