# -*- coding: utf-8 -*-

"""
***************************************************************************
*                                                                         *
*   This program is free software; you can redistribute it and/or modify  *
*   it under the terms of the GNU General Public License as published by  *
*   the Free Software Foundation; either version 2 of the License, or     *
*   (at your option) any later version.                                   *
*                                                                         *
***************************************************************************
"""

from qgis.PyQt.QtCore import QCoreApplication
from qgis.core import (QgsProcessingAlgorithm,
                       QgsProcessingParameterFolderDestination,
                       QgsProcessingParameterFile,
                       QgsProcessingParameterString,
                       QgsProcessingParameterNumber,
                       QgsProcessingParameterVectorLayer,
                       QgsProcessingContext)
from qgis import processing
from qgis.utils import iface
import ogr
import os
import sys
import gdal
import calendar
import platform
import datetime

def find_next_date(datelist, date):
    for i,l in enumerate(datelist):
        if l >= date:
            return i
    return None

def find_first_month_day(dates, month=1, day=1):
    start_year = dates[0].year
    test_date = datetime.datetime(start_year,month,day)
    if dates[0] > test_date:
        start_year += 1
        test_date = datetime.datetime(start_year, month, day)

    return find_next_date(dates, test_date)

def get_period_intervals(date_file, md=[1,1], duration=365):
    dates = []
    #orig_dates = []
    with open(date_file) as f:
        for l in f.read().splitlines():
            try:
                dates.append(datetime.datetime.strptime(l, '%Y%m%d'))
                #orig_dates.append(l)
            except:
                continue


    periods = []
    s,e = 0,0
    S = 0
    while s is not None and e is not None:
        s = find_first_month_day(dates, md[0], md[1])
        if s is not None:
            e = find_next_date(dates, dates[s] + datetime.timedelta(days=duration))
            if e is not None:
                periods.append((S+s,S+e))
                S += s+1
                dates = dates[s+1:]

    #for p in periods:
    #    print(orig_dates[p[0]],orig_dates[p[1]],p[1]-p[0]+1)

    return periods
    
def get_mean_expression(periods):
    l = periods[0][1]-periods[0][0]+1
    expr = []
    for i in range(1,l+1):
        s = [p[0]+i for p in periods]
        expr.append('(' + '+'.join(['im1b%d' % x for x in s]) + ')/' + str(len(s)))

    return '{' + ';'.join(expr) + '}'

def ModisToYYYYMMDD(raster_list, output):
    dates=[]
    for path in raster_list :
        modisDate = path.split('_')[-2][3:]
        year = int(modisDate[:-3])
        day = int(modisDate[-3:])
        month = 1
        while day - calendar.monthrange(year,month)[1] > 0 and month <= 12:
            day = day - calendar.monthrange(year,month)[1]
            month = month + 1
        month= '0'+str(month) if month < 10 else str(month)
        day= '0'+str(day) if day < 10 else str(day)
        dates.append(str(year)+str(month)+str(day))
    dates.sort()
    with open(output,'w') as f:
        for date in dates:
            f.write(date+'\n')

def NDVI_subdates_maker(TS_file,datefile,output_folder,begin_date,end_date,context,feedback):
    raster_band_list = []
    dates_list= []
    with open(datefile,'r') as f:
        for i,date in enumerate(f):
            if int(date)>=begin_date and int(date) <=end_date:
                raster_band_list.append(i+1)
                dates_list.append(int(date))
    exp=''
    for i in raster_band_list :
        exp+='im1b'+str(i)+';'
    exp=exp[:-1]
    out = os.path.join(output_folder,"sub_ndvi_ts.tif")
    BMX_parameters = {'il':[TS_file],'exp':exp, 'out':out, 'outcontext':os.path.join(output_folder,"outcontext.txt")}
    processing.run('otb:BandMathX',BMX_parameters, context=context, feedback=feedback)
    os.remove(os.path.join(output_folder,"outcontext.txt"))
    
    with open(os.path.join(output_folder,os.path.basename(datefile)),'w') as f:
        for date in dates_list:
            f.write(str(date)+'\n')
    return out, os.path.join(output_folder,os.path.basename(datefile))
            
def setNoDataValue(fn, val=None):
    """ setNoDataValue(fn,val = 0)
    Sets a value for nodata pixels in a raster GDAL dataset.

    INPUT   :   fn      -   GDAL dataset file
                val     -   nodata value (def. 0)

    OUTPUT  :   None
    """
    ds = gdal.Open(fn, gdal.GA_Update)
    for i in range(0, ds.RasterCount):
        band = ds.GetRasterBand(i + 1)
        if val is None:
            band.DeleteNoDataValue()
        else:
            band.SetNoDataValue(val)
        band = None
    ds.FlushCache()
    ds = None

def getNoDataValue(fn):
    """ getNoDataValue(fn,val = 0)
    Gets the value for nodata pixels in a raster GDAL dataset.

    INPUT   :   fn      -   GDAL dataset file

    OUTPUT  :   no-data value
    """
    ds = gdal.Open(fn)
    band = ds.GetRasterBand(1)
    ndv = band.GetNoDataValue()
    ds = None
    return ndv

def moy_NDVI_TS(TS,dates, month, day, duration):
    periods = get_period_intervals(dates, [month,day], duration)
    return get_mean_expression(periods)

class LandscapeStratification(QgsProcessingAlgorithm):
    """
    This is an example algorithm that takes a vector layer and
    creates a new identical one.

    It is meant to be used as an example of how to create your own
    algorithms and explain methods and variables used to do it. An
    algorithm like this will be available in all elements, and there
    is not need for additional work.

    All Processing algorithms should extend the QgsProcessingAlgorithm
    class.
    """

    # Constants used to refer to parameters and outputs. They will be
    # used when calling the algorithm from another algorithm, or when
    # calling from the QGIS console.

    INPUT = 'INPUT'
    INPUTDATE = 'INPUTDATE'
    BEGDATE = 'BEGDATE'
    ENDDATE = 'ENDDATE'
    STARTMONTH = 'STARTMONTH'
    STARTDAY = 'STARTDAY'
    DURATION = 'DURATION'
    CLIP = 'CLIP'
    OUTPUT = 'OUTPUT'
    PREFIX = 'PREFIX'
    CBEGIN = 'CBEGIN'
    CEND = 'CEND'
    SW = 'SW'
    CW = 'CW'
    THRESHOLD = 'THRESHOLD'

    def tr(self, string):
        """
        Returns a translatable string with the self.tr() function.
        """
        return QCoreApplication.translate('Processing', string)

    def createInstance(self):
        return LandscapeStratification()

    def name(self):
        """
        Returns the algorithm name, used for identifying the algorithm. This
        string should be fixed for the algorithm, and must not be localised.
        The name should be unique within each provider. Names should contain
        lowercase alphanumeric characters only and no spaces or other
        formatting characters.
        """
        return 'LandscapeStratification'

    def displayName(self):
        """
        Returns the translated algorithm name, which should be used for any
        user-visible display of the algorithm name.
        """
        return self.tr('LandscapeStratification')

    def group(self):
        """
        Returns the name of the group this algorithm belongs to. This string
        should be localised.
        """
        return self.tr('TSNDVI')

    def groupId(self):
        """
        Returns the unique ID of the group this algorithm belongs to. This
        string should be fixed for the algorithm, and must not be localised.
        The group id should be unique within each provider. Group id should
        contain lowercase alphanumeric characters only and no spaces or other
        formatting characters.
        """
        return 'TSNDVI'

    def shortHelpString(self):
        """
        Returns a localised short helper string for the algorithm. This string
        should provide a basic description about what the algorithm does and the
        parameters and outputs associated with it..
        """
        return self.tr("Compute a landscape stratification metric")

    def initAlgorithm(self, config=None):
        """
        Here we define the inputs and output of the algorithm, along
        with some other properties.
        """

        # We add the input vector features source. It can have any kind of
        # geometry.
        self.addParameter(
            QgsProcessingParameterFile(
                self.INPUT,
                self.tr('Input NDVI stack')
            )
        )

        self.addParameter(
            QgsProcessingParameterFile(
                self.INPUTDATE,
                self.tr('Input NDVI dates file'),
                optional=True
            )
        )
        self.addParameter(
            QgsProcessingParameterString(
                self.BEGDATE,
                self.tr('Beginning date for selection (YYYYMMDD)'),
                defaultValue='20000101'
            )
        )
        self.addParameter(
            QgsProcessingParameterString(
                self.ENDDATE,
                self.tr('End date for selection (YYYYMMDD)'),
                defaultValue=datetime.date.today().strftime("%Y%m%d")
            )
        )
        self.addParameter(
            QgsProcessingParameterNumber(
                self.STARTMONTH,
                self.tr('Starting month for multi-annual mean'),
                type=0,
                defaultValue=1,
                minValue=1,
                maxValue=12
            )
        )
        self.addParameter(
            QgsProcessingParameterNumber(
                self.STARTDAY,
                self.tr('Starting day for multi-annual mean'),
                type=0,
                defaultValue=1,
                minValue=1,
                maxValue=31
            )
        )
        self.addParameter(
            QgsProcessingParameterNumber(
                self.DURATION,
                self.tr('Period length for multi-annual mean'),
                type=0,
                defaultValue=365,
                minValue=1
            )
        )

        self.addParameter(
            QgsProcessingParameterVectorLayer(
                self.CLIP,
                self.tr('Clip vector layer'),
                optional=True,
                defaultValue=None
            )
        )

        self.addParameter(
            QgsProcessingParameterString(
                self.PREFIX,
                self.tr('Output files prefix'),
                ''
            )
        )
 
        self.addParameter(
            QgsProcessingParameterNumber(
                self.CBEGIN,
                self.tr('Starting PCA component'),
                type=0,
                defaultValue=2))
        self.addParameter(
            QgsProcessingParameterNumber(
                self.CEND,
                self.tr('Final PCA component (0 for last)'),
                type=0,
                defaultValue=5))
        self.addParameter(
            QgsProcessingParameterNumber(
                self.SW,
                self.tr('Weight for the spatial homogeneity'),
                type=1,
                defaultValue=0.5,
                minValue=0,
                maxValue=1))
        self.addParameter(
            QgsProcessingParameterNumber(
                self.CW,
                self.tr('Weight for the spectral homogeneity'),
                type=1,
                defaultValue=0.5,
                minValue=0,
                maxValue=1))
        self.addParameter(
            QgsProcessingParameterNumber(
                self.THRESHOLD,
                self.tr('Threshold for Generic region merging operation'),
                defaultValue=850
                ))

        self.addParameter(
            QgsProcessingParameterFolderDestination(
                self.OUTPUT,
                self.tr('Output folder')
            )
        )

    def processAlgorithm(self, parameters, context, feedback):
        """
        Here is where the processing itself takes place.
        """
        if platform.system() == 'Linux':
            sh = False
        elif platform.system() == 'Windows':
            sh = True
        else:
            sys.exit("Platform not supported!")
        # Retrieve the feature source and sink. The 'dest_id' variable is used
        # to uniquely identify the feature sink, and must be included in the
        # dictionary returned by the processAlgorithm function.
        Smooth_TS = self.parameterAsString(
            parameters,
            self.INPUT,
            context
        )
        
        datefile = self.parameterAsString(
            parameters,
            self.INPUTDATE,
            context
        )

        clip = self.parameterAsVectorLayer(
            parameters,
            self.CLIP,
            context
        )

        output_folder = self.parameterAsString(
            parameters,
            self.OUTPUT,
            context
        )

        prefix = parameters['PREFIX'] if parameters['PREFIX'] != None else ''
        if len(prefix)>0 and prefix[-1] != '_':
            prefix+='_'
        begin_date = parameters['BEGDATE'] if parameters['BEGDATE'] != None else 20000101
        end_date = parameters['ENDDATE'] if parameters['ENDDATE'] != None else int(datetime.date.today().strftime("%Y%m%d"))
        #write_temp_files = parameters['KEEPFILES']
        
        threshold=parameters['THRESHOLD']
        cw=parameters['CW']
        sw=parameters['SW']
        cbegin=parameters['CBEGIN']
        cend=parameters['CEND']

        tmp_name = prefix + 'metrics'
        tmp_folder = os.path.join(output_folder,tmp_name)
        if not os.path.exists(tmp_folder):
            os.mkdir(tmp_folder)

        print('Check Datefile')
        print(datefile)
        if datefile == '':
            datefile = os.path.splitext(Smooth_TS)[0] + '_dates.txt'
        print(datefile)

        if int(begin_date)!=20000101 or int(end_date)!=int(datetime.date.today().strftime("%Y%m%d")):
            TS_file, datefile = NDVI_subdates_maker(Smooth_TS,datefile,tmp_folder,int(begin_date),int(end_date),context,feedback)
        else :
            TS_file = Smooth_TS

        ndv = getNoDataValue(TS_file)

        # Compute mean time series over the period
        exp = moy_NDVI_TS(TS_file,datefile,parameters['STARTMONTH'], parameters['STARTDAY'], parameters['DURATION'])
        Moy_TS = os.path.join(tmp_folder, prefix+"LandStrat_mean.tif")
        BMX_parameters = {'il':[TS_file],'exp':exp, 'out':Moy_TS, "outputpixeltype":2, 'outcontext':os.path.join(output_folder,"outcontext.txt")}
        processing.run("otb:BandMathX", BMX_parameters, context=context, feedback=feedback)
        os.remove(os.path.join(output_folder,"outcontext.txt"))
        if ndv is not None:
            setNoDataValue(Moy_TS, ndv)

        # manage mask (nodata from input U clip mask)
        if clip is not None:
            maskc = os.path.join(tmp_folder, prefix+"mask_clip.tif")
            processing.run("otb:Rasterization", {"in":clip,"im":Smooth_TS,"mode.binary.foreground":1,"out":maskc, "outputpixeltype":0}, context=context, feedback=feedback)
            vald = os.path.join(tmp_folder, prefix+"valid.tif")
            ND_app_parameters = {"in": TS_file, "out":vald, "mode":"buildmask", 'mode.apply.mask':maskc, "outputpixeltype":0}
            processing.run('otb:ManageNoData', ND_app_parameters, context=context, feedback=feedback)
            mask = os.path.join(tmp_folder, prefix+"mask.tif")
            BMX_parameters = {'il': [maskc, vald], 'exp':'im1b1*im2b1', 'out':mask, 'outputpixeltype':0, 'outcontext':os.path.join(output_folder,"outcontext.txt")}
            processing.run("otb:BandMathX", BMX_parameters, context=context, feedback=feedback)
            os.remove(os.path.join(output_folder,"outcontext.txt"))
            #os.remove(maskc)
            #os.remove(vald)
        else:
            mask = os.path.join(tmp_folder, prefix + "mask.tif")
            ND_app_parameters = {"in": TS_file, "out": mask, "mode": "buildmask", 'mode.apply.mask': TS_file,
                                 "outputpixeltype": 0}
            processing.run('otb:ManageNoData', ND_app_parameters, context=context, feedback=feedback)

        # Compute metric using no-data value
        LS_Strat_pre = os.path.join(tmp_folder, prefix+"LandStrat_metric_pre.tif")
        LS_app_parameters = {"ndvits": Moy_TS, "out": LS_Strat_pre, "outputpixeltype": 5}
        if ndv is not None:
            LS_app_parameters["bv"] = ndv
        if cbegin != None :
            LS_app_parameters["cbegin"]=cbegin
        if cend != None :
            LS_app_parameters["cend"]=cend
        processing.run("otb:LandscapeStratificationMetric", LS_app_parameters, context=context, feedback=feedback)

        LS_Strat = os.path.join(tmp_folder, prefix + "LandStrat_metric.tif")
        ND_app_parameters = {'in': LS_Strat_pre, 'out': LS_Strat, 'mode': 'apply', 'mode.apply.mask': mask,
                             'outputpixeltype': 5}
        processing.run('otb:ManageNoData', ND_app_parameters, context=context, feedback=feedback)
        setNoDataValue(LS_Strat, ndv)
        
        LS_Strat_norm = os.path.join(tmp_folder, prefix+"LandStrat_metric_norm.tif")
        DC_app_parameters={"in":LS_Strat, "out":LS_Strat_norm, "mask":mask, "outmin":0, "outmax":2048, "outputpixeltype":2}
        processing.run("otb:DynamicConvert", DC_app_parameters, context=context, feedback=feedback)

        LS_Strat = os.path.join(tmp_folder,prefix+"LandStrat_metric_norm_masked.tif")
        ND_app_parameters = {'in': LS_Strat_norm, 'out':LS_Strat, 'mode':'apply', 'mode.apply.mask':mask, 'outputpixeltype':2}
        processing.run('otb:ManageNoData', ND_app_parameters, context=context, feedback=feedback)
        setNoDataValue(LS_Strat, ndv)

        #os.remove(LS_Strat_pre)
        #os.remove(LS_Strat_norm)

        GRM_output_pre = os.path.join(tmp_folder,prefix+"LandStrat_map_pre.tif")
        GRM_app_parameters = {'in':LS_Strat, 'threshold':float(threshold), 'criterion':'bs', 'out':GRM_output_pre}
        if cw != 0.5:
            GRM_app_parameters['criterion.bs.cw']= cw
        if sw != 0.5:
            GRM_app_parameters['criterion.bs.sw']= sw
        processing.run("otb:LSGRM", GRM_app_parameters, context=context, feedback=feedback)

        GRM_output = os.path.join(tmp_folder, prefix + "LandStrat_map.tif")
        BMX_parameters = {'il': [GRM_output_pre, mask], 'exp': 'im1b1*im2b1', 'out': GRM_output, 'outputpixeltype': 3,
                          'outcontext': os.path.join(output_folder, "outcontext.txt")}
        processing.run("otb:BandMathX", BMX_parameters, context=context, feedback=feedback)
        os.remove(os.path.join(output_folder, "outcontext.txt"))
        #os.remove(GRM_output_pre)
        setNoDataValue(GRM_output, 0)

        GRM_vec_output = os.path.join(output_folder,prefix+"LandscapeStratification.shp")
        if os.path.exists(GRM_vec_output):
            drv = ogr.GetDriverByName('ESRI Shapefile')
            drv.DeleteDataSource(GRM_vec_output)
        processing.run("gdal:polygonize", {'INPUT':GRM_output, 'BAND':1, 'OUTPUT':GRM_vec_output}, context=context, feedback=feedback)

        nm = os.path.basename(os.path.splitext(GRM_vec_output)[0])
        context.addLayerToLoadOnCompletion(GRM_vec_output,
                                           QgsProcessingContext.LayerDetails(name=nm, project=context.project()))
        
        return {'OUT':GRM_vec_output}