"""
<p>
This module implements an algorithm described by Andreas Westfeld in [1,2],
which detects if there was data embedded into an image using JSteg.
It uses the property that JSteg generates pairs of values in the
DCT-coefficients histogram, which can be detected by a \chi^2 test.
</p>

<pre>
[1]: Andreas Westfeld, F5 - A Steganographic Algorithm High Capacity Despite
Better Steganalysis
[2]: Andreas Westfeld, Angriffe auf steganographische Systeme
</pre>
"""

from collections import defaultdict
import os

from PIL import Image
import numpy
from scipy.stats import chisquare
import matplotlib.pyplot as plt
import itertools as it
from msteg.StegBase import StegBase

from msteg.StegBase import *


class ChiSquare(StegBase):
    """
    The module contains only one method, <b>detect</b>.
    """

    def __init__(self, ui, core):
        self.ui = ui
        self.core = core

    @describe_and_annotate((None, None),
                           ("Source image", ImagePath),
                           ("Target image", NewFilePath),
                           ("2nd Target image", NewFilePath))
    def detect(self, src, tgt, tgt2):
        """
        <p>
        Detect if there was data embedded in the <i>source image</i> image with
        JSteg algorithm.
        </p>

        <p>
        Parameters:
        <ol>
        <li><pre>Source image</pre> Image which should be tested</li>
        <li><pre>Target image</pre> Image which displays a graphic with the
        embedding probability</li>
        <li><pre>2nd Target image</pre> Image which displays the embedding
        positions in the image</li>
        </ol>
        </p>
        """
        # --------------------------- Input -----------------------------------
        # If src is from the image pool, test whether the image exists encoded
        # on the file system. Otherwise we can not read DCT-coefficients.
        if self.core.media_manager.is_media_key(src):
            src = self.core.media_manager.get_file(src)
            if hasattr(src, 'tmp_file'):
                src = src.tmp_file
                self.ui.display_error('Trying file: %s' % src)
            else:
                self.ui.display_error('Can not detect anything from \
                        decoded images.')
                return
        # Test whether the file exists.
        if not os.path.isfile(src):
            self.ui.display_error('No such file.')
            return
        # Test if it is a JPEG file.
        if not self._looks_like_jpeg(src):
            self.ui.display_error('Input is probably not a JPEG file.')
            return

        # ---------------------------- Algorithm ------------------------------
        # Build DCT-histogram in steps of \approx 1% of all coefficients and
        # calculate the p-value at each step.
        dct_data = rw_dct.read_dct_coefficients(src)
        hist = defaultdict(int)
        cnt = 0
        l = len(dct_data)
        one_p = l / 100
        result = []
        for block in dct_data:
            # update the histogram with one block of 64 coefficients
            for c in block:
                hist[c] += 1

            cnt += 1
            if not cnt % one_p:
                # calculate p-value
                self.ui.set_progress(cnt * 100 / l)

                # ignore the pair (0, 1), since JSteg does not embed data there
                hl = [hist[i] for i in range(-2048, 2049) if not i in (0, 1)]
                k = len(hl) / 2
                observed = []
                expected = []
                # calculate observed and expected distribution
                for i in range(k):
                    t = hl[2 * i] + hl[2 * i + 1]
                    if t > 3:
                        observed.append(hl[2 * i])
                        expected.append(t / 2)
                # calculate (\chi^2, p)
                p = chisquare(numpy.array(observed), numpy.array(expected))[1]
                result.append(p)

        # ----------------------------- Output --------------------------------
        # Graph displaying the embedding probabilities in relation to the
        # sample size.
        figure = plt.figure()
        plot = figure.add_subplot(111)
        plot.grid(True)
        plot.plot(result, color='r', linewidth=2.0)
        plt.axis([0, 100, 0, 1.1])
        plt.title('Embedding probability for different percentages \
of the file capacity.')
        plt.xlabel('% of file capacity')
        plt.ylabel('Embedding probability')

        if self.core.media_manager.is_media_key(tgt):
            img = figure_to_pil(figure)
            self.core.media_manager.put_media(tgt, img)
        else:
            plt.savefig(tgt)

        # Image displaying the length and position of the embedded data
        # within the image
        img2 = Image.open(src)
        img2.convert("RGB")
        width, height = img2.size

        for i in range(100):
            result[i] = max(result[i:])

        cnt2 = 0
        for (top, left) in it.product(range(0, height, 8), range(0, width, 8)):
            if not cnt2 % one_p:
                r = result[cnt2 / one_p]
                if r >= 0.5:
                    color = (255, int((1 - r) * 2 * 255), 0)
                else:
                    color = (int(r * 2 * 255), 255, 0)
            cnt2 += 1
            img2.paste(color, (left, top, min(left + 8, width),
                                   min(top + 8, height)))
        self.core.media_manager.put_media(tgt2, img2)

    def __str__(self):
        return 'Chi-Square-Test'


def figure_to_pil(figure):
    figure.canvas.draw()
    return Image.fromstring('RGB',
                            figure.canvas.get_width_height(),
                            figure.canvas.tostring_rgb())