from unittest import TestCase, SkipTest
import sys

from parameterized import parameterized
import numpy as np
import pandas as pd

from holoviews.core import GridMatrix, NdOverlay
from holoviews.element import (
    Bivariate,
    Distribution,
    HexTiles,
    Histogram,
    Scatter,
)
from hvplot import scatter_matrix


class TestScatterMatrix(TestCase):
    def setUp(self):
        self.df = pd.DataFrame(np.random.randn(1000, 4), columns=['a', 'b', 'c', 'd'])

    def test_returns_gridmatrix(self):
        sm = scatter_matrix(self.df)
        self.assertIsInstance(sm, GridMatrix)

    def test_wrong_diagonal(self):
        with self.assertRaises(ValueError):
            scatter_matrix(self.df, diagonal='wrong')

    def test_wrong_chart(self):
        with self.assertRaises(ValueError):
            scatter_matrix(self.df, chart='wrong')

    def test_diagonal_default(self):
        sm = scatter_matrix(self.df)
        self.assertIsInstance(sm['a', 'a'], Histogram)

    def test_offdiagonal_default(self):
        sm = scatter_matrix(self.df)
        self.assertIsInstance(sm['a', 'b'], Scatter)

    def test_diagonal_kde(self):
        sm = scatter_matrix(self.df, diagonal='kde')
        self.assertIsInstance(sm['a', 'a'], Distribution)

    def test_offdiagonal_bivariate(self):
        sm = scatter_matrix(self.df, chart='bivariate')
        self.assertIsInstance(sm['a', 'b'], Bivariate)

    def test_offdiagonal_hexbin(self):
        sm = scatter_matrix(self.df, chart='hexbin')
        self.assertIsInstance(sm['a', 'b'], HexTiles)

    def test_diagonal_kwargs_mutually_exclusive(self):
        with self.assertRaises(TypeError):
            scatter_matrix(self.df, diagonal_kwds=dict(a=1), hist_kwds=dict(a=1))
        with self.assertRaises(TypeError):
            scatter_matrix(self.df, diagonal_kwds=dict(a=1), density_kwds=dict(a=1))
        with self.assertRaises(TypeError):
            scatter_matrix(self.df, density_kwds=dict(a=1), hist_kwds=dict(a=1))

    def test_diagonal_kwargs(self):
        sm = scatter_matrix(self.df, diagonal_kwds=dict(line_color='red'))
        self.assertEqual(sm['a', 'a'].opts.get().kwargs['line_color'], 'red')

    def test_c(self):
        df = self.df.copy(deep=True)
        df['e'] = np.random.choice(list('xyz'), size=len(df))
        sm = scatter_matrix(df, c='e')

        self.assertIsInstance(sm['a', 'a'], NdOverlay)
        diag_kdims = sm['a', 'a'].kdims
        self.assertEqual(len(diag_kdims), 1)
        self.assertEqual(diag_kdims[0].name, 'e')

        self.assertIsInstance(sm['a', 'b'], Scatter)
        offdiag_vdims = sm['a', 'b'].vdims
        self.assertTrue('e' in (d.name for d in offdiag_vdims))


class TestDatashader(TestCase):
    def setUp(self):
        try:
            import datashader  # noqa
        except ImportError:
            raise SkipTest('Datashader not available')
        if sys.maxsize < 2**32:
            raise SkipTest('Datashader does not support 32-bit systems')
        self.df = pd.DataFrame(np.random.randn(1000, 3), columns=['a', 'b', 'c'])

    def test_rasterize_datashade_mutually_exclusive(self):
        with self.assertRaises(ValueError):
            scatter_matrix(self.df, rasterize=True, datashade=True)

    def test_spread_but_no_rasterize_or_datashade(self):
        with self.assertRaises(ValueError):
            scatter_matrix(self.df, dynspread=True)
        with self.assertRaises(ValueError):
            scatter_matrix(self.df, spread=True)
        with self.assertRaises(ValueError):
            scatter_matrix(self.df, dynspread=True, spread=True)

    @parameterized.expand([('rasterize',), ('datashade',)])
    def test_rasterization(self, operation):
        sm = scatter_matrix(self.df, **{operation: True})
        dm = sm['a', 'b']
        self.assertEqual(dm.callback.operation.name, operation)
        dm[()]
        self.assertEqual(len(dm.last.pipeline.operations), 3)

    @parameterized.expand([('rasterize',), ('datashade',)])
    def test_datashade_aggregator(self, operation):
        sm = scatter_matrix(self.df, aggregator='mean', **{operation: True})
        dm = sm['a', 'b']
        dm[()]
        self.assertEqual(dm.last.pipeline.operations[-1].aggregator, 'mean')

    @parameterized.expand([('spread',), ('dynspread',)])
    def test_spread_rasterize(self, operation):
        sm = scatter_matrix(self.df, rasterize=True, **{operation: True})
        dm = sm['a', 'b']
        dm[()]
        self.assertEqual(len(dm.last.pipeline.operations), 4)

    @parameterized.expand([('spread',), ('dynspread',)])
    def test_spread_datashade(self, operation):
        sm = scatter_matrix(self.df, datashade=True, **{operation: True})
        dm = sm['a', 'b']
        dm[()]
        self.assertEqual(len(dm.last.pipeline.operations), 4)

    @parameterized.expand([('spread',), ('dynspread',)])
    def test_spread_kwargs(self, operation):
        sm = scatter_matrix(self.df, datashade=True, **{operation: True, 'shape': 'circle'})
        dm = sm['a', 'b']
        dm[()]
        self.assertEqual(dm.last.pipeline.operations[-1].args[0].keywords['shape'], 'circle')