def calcFilters(self, pixels, mode):
filteredPixels = []
for filter in loader.getFilters(self.layer_id):
if filter.id == 'Bfast' and mode == 'series':
continue;
elif filter.id != 'Bfast' and mode == 'trend':
continue;
filteredPixels.append({"filterProperties": {"id": filter.id, "label": filter.label}, "pixels": []});
for pixel in pixels:
values = pixel["values"];
longitude = pixel["lon"];
latitude = pixel["lat"];
filteredValues = filter.run(values, longitude, latitude);
filteredValues = filteredValues if type(filteredValues) == list else filteredValues.tolist();
if len(filteredValues) == len(values):
filteredPixels[-1]["pixels"].append({"values": filteredValues, "lon": longitude, "lat": latitude});
else:
filteredPixels[-1]["pixels"].append({"values": [0 for _ in xrange(len(values))], "lon": longitude, "lat": latitude});
return filteredPixels;
def lookup(self, geoJsonGeometry, mode=None):
result = []
filepath = os.path.join(self.imageDbPath, self.file)
mean = self.zonalStatistics(geoJsonGeometry, filepath)
if "month" == self.temporal_resolution_type:
date = self.localDateByMonth()
else:
date = self.LocalDateByDay()
geoJsonGeometry = ast.literal_eval(geoJsonGeometry)
for i, j in zip(mean, date):
count = []
count.append(j)
count.append(float(i))
result.append(count)
series = [{"position": 1, "type": "original", "id": "original", "label": "Valores originais"}]
for filter in loader.getFilters(self.layer_id):
if filter.id == "Bfast" and mode == "series":
continue
elif filter.id != "Bfast" and mode == "trend":
continue
filters = loader.getFilters(self.layer_id)
bfastIndex = utils.findIndexByAttribute(filters, "id", filter.id)
filteredValues = filters[bfastIndex].run(mean, None, None)
filteredValues = filteredValues if type(filteredValues) == list else filteredValues.tolist()
if len(filteredValues) == len(mean):
series.append({"id": filter.id, "label": filter.label, "position": len(result[0]), "type": "filter"})
utils.joinArray(result, filteredValues)
if mode == "trend":
filters = loader.getFilters(self.layer_id)
bfastIndex = utils.findIndexByAttribute(filters, "id", "Bfast")
filteredValues = filters[bfastIndex].run(mean, None, None)
filteredValues = filteredValues if type(filteredValues) == list else filteredValues.tolist()
if len(filteredValues) == len(mean):
series.append({"id": "Bfast", "label": "BFAST", "position": len(result[0]), "type": "trend"}) # 2
utils.joinArray(result, filteredValues)
return {"series": series, "values": result}
def calcTrend(self, resultMean):
filters = loader.getFilters(self.layer_id);
bfastIndex = utils.findIndexByAttribute(filters, 'id', 'Bfast')
origSeriesProp = utils.findIndexByAttribute(resultMean['series'], 'type', 'original');
originalValues = utils.oneArray(resultMean['values'], resultMean['series'][origSeriesProp]['position']-1)
filteredValues = filters[bfastIndex].run(originalValues, None, None);
filteredValues = filteredValues if type(filteredValues) == list else filteredValues.tolist();
result = {
'series': {
'id': "Bfast",
'label': 'BFAST',
'position': len(resultMean['values'][0])+1,
'type': 'trend'
},
'values': filteredValues
}
return result
def run(layerId, longitude, latitude):
datasourceInstance = loader.getDatasource(layerId)
timeserieData = datasourceInstance.lockup(longitude, latitude)
values = utils.oneArray(timeserieData)
position = 1
series = [{'id': 'original', 'label': 'Valores originais', 'position': 1}];
for filter in loader.getFilters(layerId):
filteredValues = filter.run(values, longitude, latitude);
if len(filteredValues) == len(values):
position += 1
series.append({ 'id': filter.id, 'label': filter.label, 'position': position })
utils.joinArray(timeserieData, filteredValues)
return {
'series': series,
'values': timeserieData
}
def trend(layerId, startYear, endYear, interpolation, groupData, timeChange, timeChangeUnits, geoJsonGeometry):
#Obtem a série
datasourceInstance = loader.getDatasource(layerId)
timeserieData = datasourceInstance.lookup(geoJsonGeometry, mode="series")
#Obtem as datas
datesStr = utils.oneArray(timeserieData["values"], 0)
#Obtem os valores entre os anos startYear e endYear
clippedValues = bfast_utils.clipValuesByYear(startYear, endYear, timeserieData["values"], datesStr)
values = clippedValues['values']
dates = clippedValues['dates']
startDate = dates[0]
endDate = dates[-1]
#Encontra o indice do filtro(se houver) e do bfast
filters = loader.getFilters(layerId)
bfastIndex = utils.findIndexByAttribute(filters, 'id', 'Bfast')
# O filtro já foi calculado na datasource
interpolation = interpolation if interpolation != "original" else 'STRING_NOT_FOUND';
interpolationIndex, interpolationPos = next(((i, item['position']) for i, item in enumerate(timeserieData["series"]) if item['id'] == interpolation), (-1, -1))
#Inicializa preparação do objeto de resposta
series = [{'id':'Bfast', 'label':'BFAST', 'position':1, 'type':'trend'}]
#Se há filtro selecionado, executa e adiciona ao objeto de resposta
if interpolationPos != -1:
values = utils.oneArray(timeserieData["values"], interpolationPos)
series.append({'id':timeserieData["series"][interpolationIndex]["id"], 'label': timeserieData["series"][interpolationIndex]["label"], 'position': 2, 'type': 'filter'})
else:
series.append({'id':'original', 'label':'Valores Originais', 'position':2, 'type':'original'})
#Agrupa os valores
groupedData = bfast_utils.groupData(dates, values, groupData)
values = groupedData['values']
frequency = groupedData['frequency']
dates = groupedData['dates']
#Calcula o valor do parâmetro h(minimal segment size) para o bfast
minimalSegmentSize = bfast_utils.calculateMinimalSegmentSize(len(values), timeChange, timeChangeUnits, frequency)
#Executa o BFAST
result = []
if bfastIndex != -1:
result = filters[bfastIndex].run(values, None, None, minimalSegmentSize, frequency, startDate, endDate)
else:
raise IndexError("Bfast filter could not be found.")
#Acrescenta ao resultado as datas, os valores da tendência e os valores originais (agrupados ou não)
datesList = [[i] for i in dates]
utils.joinArray(datesList, result)
result = datesList
utils.joinArray(result, values)
return {
'series': series,
'values': result
}
def lookup(self, geoJsonGeometry, mode=None):
result = []
filepath = os.path.join(self.imageDbPath, self.file)
mean = self.zonalStatistics(geoJsonGeometry, filepath)
if ('month' == self.temporal_resolution_type):
date = self.localDateByMonth()
else:
date = self.LocalDateByDay()
geoJsonGeometry = ast.literal_eval(geoJsonGeometry)
for i, j in zip(mean, date):
count = []
count.append(j)
count.append(float(i))
result.append(count)
series = [{
'position': 1,
'type': 'original',
'id': 'original',
'label': 'Valores originais'
}]
for filter in loader.getFilters(self.layer_id):
if filter.id == 'Bfast' and mode == 'series':
continue
elif filter.id != 'Bfast' and mode == 'trend':
continue
filters = loader.getFilters(self.layer_id)
bfastIndex = utils.findIndexByAttribute(filters, 'id', filter.id)
filteredValues = filters[bfastIndex].run(mean, None, None)
filteredValues = filteredValues if type(
filteredValues) == list else filteredValues.tolist()
if len(filteredValues) == len(mean):
series.append({
'id': filter.id,
'label': filter.label,
'position': len(result[0]),
'type': 'filter'
})
utils.joinArray(result, filteredValues)
if mode == 'trend':
filters = loader.getFilters(self.layer_id)
bfastIndex = utils.findIndexByAttribute(filters, 'id', 'Bfast')
filteredValues = filters[bfastIndex].run(mean, None, None)
filteredValues = filteredValues if type(
filteredValues) == list else filteredValues.tolist()
if len(filteredValues) == len(mean):
series.append({
'id': "Bfast",
'label': 'BFAST',
'position': len(result[0]), #2
'type': 'trend'
})
utils.joinArray(result, filteredValues)
return {
"series": series,
"values": result
}
def trend(layerId, startYear, endYear, interpolation, groupData, timeChange,
timeChangeUnits, geoJsonGeometry):
# Obtem a série
datasourceInstance = loader.getDatasource(layerId)
timeserieData = datasourceInstance.lookup(geoJsonGeometry, mode="series")
# Obtem as datas
datesStr = utils.oneArray(timeserieData["values"], 0)
# Obtem os valores entre os anos startYear e endYear
clippedValues = bfast_utils.clipValuesByYear(startYear, endYear,
timeserieData["values"],
datesStr)
values = clippedValues['values']
dates = clippedValues['dates']
startDate = dates[0]
endDate = dates[-1]
# Encontra o indice do filtro(se houver) e do bfast
filters = loader.getFilters(layerId)
bfastIndex = utils.findIndexByAttribute(filters, 'id', 'Bfast')
# O filtro já foi calculado na datasource
interpolation = interpolation if interpolation != "original" else 'STRING_NOT_FOUND'
interpolationIndex, interpolationPos = next(
((i, item['position'])
for i, item in enumerate(timeserieData["series"])
if item['id'] == interpolation), (-1, -1))
# Inicializa preparação do objeto de resposta
series = [{
'id': 'Bfast',
'label': 'BFAST',
'position': 1,
'type': 'trend'
}]
# Se há filtro selecionado, executa e adiciona ao objeto de resposta
if interpolationPos != -1:
values = utils.oneArray(timeserieData["values"], interpolationPos)
series.append({
'id':
timeserieData["series"][interpolationIndex]["id"],
'label':
timeserieData["series"][interpolationIndex]["label"],
'position':
2,
'type':
'filter'
})
else:
series.append({
'id': 'original',
'label': 'Valores Originais',
'position': 2,
'type': 'original'
})
# Agrupa os valores
groupedData = bfast_utils.groupData(dates, values, groupData)
values = groupedData['values']
frequency = groupedData['frequency']
dates = groupedData['dates']
# Calcula o valor do parâmetro h(minimal segment size) para o bfast
minimalSegmentSize = bfast_utils.calculateMinimalSegmentSize(
len(values), timeChange, timeChangeUnits, frequency)
# Executa o BFAST
result = []
if bfastIndex != -1:
result = filters[bfastIndex].run(values, None, None,
minimalSegmentSize, frequency,
startDate, endDate)
else:
raise IndexError("Bfast filter could not be found.")
# Acrescenta ao resultado as datas, os valores da tendência e os valores originais (agrupados ou não)
datesList = [[i] for i in dates]
utils.joinArray(datesList, result)
result = datesList
utils.joinArray(result, values)
return {'series': series, 'values': result}