class Thumbnail(object):
def __init__(self):
self.ee = EarthEngine(settings.EE_TOKEN)
def thumbid(self, id):
MAP_IMAGE = {'id':id};
MAP_IMAGE_BANDS = ['30','20','10'];
cmd = {
'getid':1,
'image':urllib.quote_plus(json.dumps(MAP_IMAGE)),
'bands':','.join(MAP_IMAGE_BANDS), #'30,20,10'
}
return self._execute_cmd("/thumb", cmd)
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.get(url, params)
class Thumbnail(object):
def __init__(self):
self.ee = EarthEngine(settings.EE_TOKEN)
def thumbid(self, id):
MAP_IMAGE = {'id':id};
MAP_IMAGE_BANDS = ['30','20','10'];
cmd = {
'getid':1,
'image':urllib.quote_plus(json.dumps(MAP_IMAGE)),
'bands':','.join(MAP_IMAGE_BANDS), #'30,20,10'
}
return self._execute_cmd("/thumb", cmd)
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.get(url, params)
class EELandsat(object):
def __init__(self, resource):
self.resource = resource
self.ee = EarthEngine(settings.EE_TOKEN)
def list(self, bounds, params={}):
images = self.ee.get("/list?id=%s&bbox=%s&fields=ACQUISITION_DATE" % (self.resource, bounds))
logging.info(images)
if 'data' in images:
return [x['id'] for x in images['data']]
return []
def mapid(self, start, end):
landsat_bands = ['10','20','30','40','50','70','80','61','62']
creator_bands =[{'id':id, 'data_type':'float'} for id in landsat_bands]
MAP_IMAGE1 = {
'creator':'LANDSAT/LandsatTOA',
'input':'LANDSAT/L7_L1T',
'bands':creator_bands,
'start_time': start,
'end_time': end
};
MAP_IMAGE = {
"creator": "SimpleMosaic",
"args": [MAP_IMAGE1]
}
PREVIEW_GAIN = 500;
MAP_IMAGE_BANDS = ['30','20','10'];
cmd = {
'image': json.dumps(MAP_IMAGE), #json.dumps(lanstat),
'bands': ','.join(MAP_IMAGE_BANDS), #'30,20,10',
'gain': PREVIEW_GAIN
}
return self._execute_cmd("/mapid", cmd)
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.post(url, params)
class EELandsat(object):
def __init__(self, resource):
self.resource = resource
self.ee = EarthEngine(settings.EE_TOKEN)
def list(self, bounds, params={}):
images = self.ee.get("/list?id=%s&bbox=%s&fields=ACQUISITION_DATE" % (self.resource, bounds))
logging.info(images)
if 'data' in images:
return [x['id'] for x in images['data']]
return []
def mapid(self, start, end):
landsat_bands = ['10','20','30','40','50','70','80','61','62']
creator_bands =[{'id':id, 'data_type':'float'} for id in landsat_bands]
MAP_IMAGE1 = {
'creator':'LANDSAT/LandsatTOA',
'input':'LANDSAT/L7_L1T',
'bands':creator_bands,
'start_time': start,
'end_time': end
};
MAP_IMAGE = {
"creator": "SimpleMosaic",
"args": [MAP_IMAGE1]
}
PREVIEW_GAIN = 500;
MAP_IMAGE_BANDS = ['30','20','10'];
cmd = {
'image': json.dumps(MAP_IMAGE), #json.dumps(lanstat),
'bands': ','.join(MAP_IMAGE_BANDS), #'30,20,10',
'gain': PREVIEW_GAIN
}
return self._execute_cmd("/mapid", cmd)
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.post(url, params)
class NDFI(object):
""" ndfi info for a period of time
"""
# hardcoded data for request
PRODES_IMAGE = {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ProdesImage',
"args": ["PRODES_2009"]
};
MODIS_BANDS = [
'sur_refl_b01_250m', 'sur_refl_b02_250m', 'sur_refl_b03_500m',
'sur_refl_b04_500m', 'sur_refl_b06_500m', 'sur_refl_b07_500m'];
THIS_POLY = None
def __init__(self, ee_res, last_perdiod, work_period):
self.last_perdiod = dict(start=last_perdiod[0],
end=last_perdiod[1])
self.work_period = dict(start=work_period[0],
end=work_period[1])
self.earth_engine_resource = ee_res
self.ee = EarthEngine(settings.EE_TOKEN)
self._image_cache = {}
def mapid2_cmd(self, asset_id, polygon=None, rows=10, cols=10):
return {"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.GetNDFIDelta","args": [
self.last_perdiod['start'],
self.last_perdiod['end'],
self.work_period['start'],
self.work_period['end'],
"MODIS/MOD09GA",
"MODIS/MOD09GQ",
{'type':'FeatureCollection','table_id': 1868251, 'mark': str(timestamp())},
{"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.ProdesImage","args":[asset_id]},
polygon,
rows,
cols]
}
def mapid2(self, asset_id):
cmd = {
"image": json.dumps(self.mapid2_cmd(asset_id)),
"format": 'png'
}
return self._execute_cmd('/mapid', cmd)
def freeze_map(self, asset_id, table, report_id):
"""
"""
cmd ={
"value": json.dumps({
"creator": CALL_SCOPE + "/com.google.earthengine.examples.sad.FreezeMap",
"args":[{
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.ProdesImage",
"args":[asset_id]
},
{'type':'FeatureCollection','table_id':1228540,'mark': str(timestamp())},
#{
#"type":"FeatureCollection",
#"table_id": table, "filters":[
#{"metadata":"report_id", "equals": report_id}
#]
#},
"type"
],
"type":"Image"
}),
}
return self._execute_cmd('/create', cmd)
def rgbid(self):
""" return params to access NDFI rgb image """
# get map id from EE
params = self._RGB_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def smaid(self):
""" return params to access NDFI rgb image """
# get map id from EE
params = self._SMA_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def ndfi0id(self):
# get map id from EE
params = self._NDFI_period_image_command(self.last_perdiod)
return self._execute_cmd('/mapid', params)
def ndfi1id(self):
# get map id from EE
params = self._NDFI_period_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def rgb_strech(self, polygon, bands):
# this is an special call, the application needs to call /value
# before call /mapid in order to google earthn engine makes his work
cmd = self._RGB_streched_command(self.work_period, polygon, bands)
del cmd['bands']
cmd['fields'] = 'stats_sur_refl_b01,stats_sur_refl_b02,stats_sur_refl_b03,stats_sur_refl_b04,stats_sur_refl_b05'
self._execute_cmd('/value', cmd)
cmd = self._RGB_streched_command(self.work_period, polygon, bands)
return self._execute_cmd('/mapid', cmd)
def _get_polygon_bbox(self, polygon):
lats = [x[0] for x in polygon]
lngs = [x[1] for x in polygon]
max_lat = max(lats)
min_lat = min(lats)
max_lng = max(lngs)
min_lng = min(lngs)
return ((min_lat, max_lat), (min_lng, max_lng))
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.post(url, params)
def ndfi_change_value(self, asset_id, polygon, rows=10, cols=10):
img = self.mapid2_cmd(asset_id, polygon, rows, cols)
cmd = {
"image": json.dumps(img),
"fields": 'ndfiSum'#','.join(fields)
}
return self._execute_cmd('/value', cmd)
def ndfi_change_value_old(self, polygons, rows=10, cols=10):
""" return how much NDFI has changed in the time period
``polygons`` are a list of closed polygons defined by lat, lon::
[
[ [lat, lon], [lat, lon]...],
[ [lat, lon], [lat, lon]...]
]
for performance reassons bbox greater than 3 degrees are now allowed
"""
# some assertions
for p in polygons:
bbox = self._get_polygon_bbox(p)
if bbox[0][1] - bbox[0][0] > 3.5 or bbox[1][1] - bbox[1][0] > 3.5:
#raise Exception("polygon bbox size must be less than 3 degrees")
pass
cmd = self._NDFI_change_value(
self.last_perdiod,
self.work_period,
polygons,
rows,
cols
)
return self._execute_cmd('/value', cmd)
def _images_for_period(self, period):
cache_key = "%d-%d" %(period['start'], period['end'])
if cache_key in self._image_cache:
img = self._image_cache[cache_key]
else:
reference_images = self.ee.get("/list?id=%s&starttime=%s&endtime=%s" % (
self.earth_engine_resource,
int(period['start']),
int(period['end'])
))
logging.info(reference_images)
img = [x['id'] for x in reference_images['data']]
self._image_cache[cache_key] = img
return img
def _image_composition(self, image_list):
""" create commands to compose images in google earth engine
ok, i really have NO idea what's going on :)
"""
specs = []
for image in image_list:
name = image.split("_", 2)[-1]
specs.append({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ModisCombiner',
"args": ['MOD09GA_005_' + name, 'MOD09GQ_005_' + name]
});
return specs;
def _NDFI_image(self, period):
""" given image list from EE, returns the operator chain to return NDFI image """
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.NDFIImage',
"args": [{
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.UnmixModis',
"args": [{
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.KrigingStub',
"args": [ self._MakeMosaic(period) ]
}]
}]
}
def _change_detection_data(self, reference_period, work_period, polygons=[], cols=10, rows=10):
ndfi_image_1 = self._NDFI_image(reference_period)
ndfi_image_2 = self._NDFI_image(work_period)
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ChangeDetectionData',
"args": [ndfi_image_1,
ndfi_image_2,
self.PRODES_IMAGE,
polygons,
rows,
cols]
}
def _NDFI_change_value(self, reference_period, work_period, polygons, cols=10, rows=10):
""" calc the ndfi change value between two periods inside specified polys
``polygons`` are a list of closed polygons defined by lat, lon::
[
[ [lat, lon], [lat, lon]...],
[ [lat, lon], [lat, lon]...]
]
"""
POLY = []
fields = []
image = self._change_detection_data(reference_period, work_period, [polygons], cols, rows)
return {
"image": json.dumps(image),
"fields": 'ndfiSum'#','.join(fields)
}
def _NDFI_period_image_command(self, period):
""" get NDFI command to get map of NDFI for a period of time """
ndfi_image = self._NDFI_image(period)
return {
"image": json.dumps(ndfi_image),
"bands": 'vis-red,vis-green,vis-blue',
"gain": 1,
"bias": 0.0,
"gamma": 1.6
}
def _RGB_image_command(self, period):
""" commands for RGB image """
return {
"image": json.dumps({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.KrigingStub',
"args": [ self._MakeMosaic(period) ]
}),
"bands": 'sur_refl_b01,sur_refl_b04,sur_refl_b03',
"gain": 0.1,
"bias": 0.0,
"gamma": 1.6
};
def _MakeMosaic(self, period):
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.MakeMosaic',
"args": ["MODIS/MOD09GA","MODIS/MOD09GQ",
{'type':'FeatureCollection','table_id':1868251, 'mark': str(timestamp())}, #this number is an special fusion tables id but i have no idea what it is supposed to do
period['start'], period['end']]
}
def _SMA_image_command(self, period):
return {
"image": json.dumps({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.UnmixModis',
"args": [{
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.KrigingStub',
"args": [self._MakeMosaic(period)]
}]
}),
"bands": 'gv,soil,npv',
"gain": 256,
"bias": 0.0,
"gamma": 1.6
};
def _RGB_streched_command(self, period, polygon, bands):
""" bands in format (1, 2, 3) """
bands = "sur_refl_b0%d,sur_refl_b0%d,sur_refl_b0%d" % bands
return {
"image": json.dumps({
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.StretchImage",
"args":[{
"creator":"ClipToMultiPolygon",
"args":[{
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.KrigingStub",
"args":[ self._MakeMosaic(period)]
},
polygon]
},
["sur_refl_b01","sur_refl_b02","sur_refl_b03","sur_refl_b04","sur_refl_b05"],
2 #EPIC
]
}),
"bands": bands
}
class NDFI(object):
""" ndfi info for a period of time
"""
# hardcoded data for request
PRODES_IMAGE = {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ProdesImage',
"args": ["PRODES_2009"]
};
MODIS_BANDS = [
'sur_refl_b01_250m', 'sur_refl_b02_250m', 'sur_refl_b03_500m',
'sur_refl_b04_500m', 'sur_refl_b06_500m', 'sur_refl_b07_500m'];
THIS_POLY = None
def __init__(self, ee_res, last_period, work_period):
self.last_period = dict(start=last_period[0],
end=last_period[1])
self.work_period = dict(start=work_period[0],
end=work_period[1])
self.earth_engine_resource = ee_res
self.ee = EarthEngine(settings.EE_TOKEN)
self._image_cache = {}
def paint_deforestation(self, asset_id, month, year):
year_str = "%04d" % (year)
#end = "%04d%02d" % (year, month)
return {
"type": "Image", "creator": "Paint", "args": [asset_id,
{
"table_id": int(settings.FT_TABLE_ID), "type": "FeatureCollection",
"filter":[{"property":"type","equals":7}, {"property":"asset_id","contains":year_str}]},
4]
}
def mapid2_cmd(self, asset_id, polygon=None, rows=5, cols=5):
year_msec = 1000 * 60 * 60 * 24 * 365
month_msec = 1000 * 60 * 60 * 24 * 30
six_months_ago = self.work_period['end'] - month_msec * 6
one_month_ago = self.work_period['end'] - month_msec
last_month = time.gmtime(int(six_months_ago / 1000))[1]
last_year = time.gmtime(int(six_months_ago / 1000))[0]
previous_month = time.gmtime(int(one_month_ago / 1000))[1]
previous_year = time.gmtime(int(one_month_ago / 1000))[0]
work_month = self.getMidMonth(self.work_period['start'], self.work_period['end'])
work_year = self.getMidYear(self.work_period['start'], self.work_period['end'])
end = "%04d%02d" % (work_year, work_month)
start = "%04d%02d" % (last_year, last_month)
previous = "%04d%02d" % (previous_year, previous_month)
start_filter = [{'property':'compounddate','greater_than':start},{'property':'compounddate','less_than':end}]
deforested_asset = self.paint_deforestation(asset_id, work_month, work_year)
json_cmd = {"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.GetNDFIDelta","args": [
self.last_period['start'] - year_msec,
self.last_period['end'],
self.work_period['start'],
self.work_period['end'],
"MODIS/MOD09GA",
"MODIS/MOD09GQ",
{'type':'FeatureCollection','table_id': 1868251, 'mark': str(timestamp()), 'filter':start_filter},
{'type':'FeatureCollection','table_id': 1868251, 'mark': str(timestamp()),
'filter':[{"property":"month","equals":work_month},{"property":"year","equals":work_year}]},
{'type':'FeatureCollection','table_id': 4468280, 'mark': str(timestamp()),
'filter':[{"property":"Compounddate","equals":int(previous)}]},
{'type':'FeatureCollection','table_id': 4468280, 'mark': str(timestamp()),
'filter':[{"property":"Compounddate","equals":int(end)}]},
deforested_asset,
polygon,
rows,
cols]
}
logging.info("GetNDFIDelta")
logging.info(json_cmd)
return json_cmd
def getMidMonth(self, start, end):
middle_seconds = int((end + start) / 2000)
this_time = time.gmtime(middle_seconds)
return this_time[1]
def getMidYear(self, start, end):
middle_seconds = int((end + start) / 2000)
this_time = time.gmtime(middle_seconds)
return this_time[0]
def mapid2(self, asset_id):
cmd = {
"image": json.dumps(self.mapid2_cmd(asset_id)),
"format": 'png'
}
return self._execute_cmd('/mapid', cmd)
def freeze_map(self, asset_id, table, report_id):
"""
"""
base_image = {"creator": CALL_SCOPE + "/com.google.earthengine.examples.sad.ProdesImage", "args":[asset_id]};
remapped = {"algorithm": "Image.remap", "image":base_image,
"from":[0,1,2,3,4,5,6,7,8,9], "to":[0,1,2,3,4,5,6,2,3,9]}
def_image = self.paint_call(remapped, int(report_id), table, 7)
selected_def = {"algorithm": "Image.select", "input": def_image,
"bandSelectors":["remapped"]}
deg_image = self.paint_call(selected_def, int(report_id), table, 8)
renamed_image = {"algorithm": "Image.select", "input": deg_image,
"bandSelectors":["remapped"], "newNames":["classification"]}
clipped_image = {"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.AddBB",
"args":[renamed_image, asset_id, "classification"]}
map_image = {"algorithm": "Image.addBands", "dstImg": asset_id, "srcImg": clipped_image,
"names": ["classification"], "overwrite": True}
cmd = {"value": json.dumps(map_image)}
return self._execute_cmd('/create', cmd)
def paint_call(self, current_asset, report_id, table, value):
return {
"algorithm": "Image.paint", "image": current_asset,
"featureCollection": {
"table_id": int(table), "type": "FeatureCollection",
"filter":[{"property":"report_id","equals":int(report_id)},
{"property":"type", "equals": value}]},
"color": value
}
def rgbid(self):
""" return params to access NDFI rgb image """
# get map id from EE
params = self._RGB_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def smaid(self):
""" return params to access NDFI rgb image """
# get map id from EE
params = self._SMA_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def ndfi0id(self):
# get map id from EE set long_span=1
params = self._NDFI_period_image_command(self.last_period, 1)
return self._execute_cmd('/mapid', params)
def baseline(self):
params = self._baseline_image_command()
return self._execute_cmd('/mapid', params)
def ndfi1id(self):
# get map id from EE
params = self._NDFI_period_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def rgb_strech(self, polygon, sensor, bands):
# this is an special call, the application needs to call /value
# before call /mapid in order to google earthn engine makes his work
cmd = self._RGB_streched_command(self.work_period, polygon, sensor, bands)
del cmd['bands']
if (sensor=="modis"):
cmd['fields'] = 'stats_sur_refl_b01,stats_sur_refl_b02,stats_sur_refl_b03,stats_sur_refl_b04,stats_sur_refl_b05'
else:
cmd['fields'] = 'stats_30,stats_20,stats_10'
self._execute_cmd('/value', cmd)
cmd = self._RGB_streched_command(self.work_period, polygon, sensor, bands)
return self._execute_cmd('/mapid', cmd)
def _get_polygon_bbox(self, polygon):
lats = [x[0] for x in polygon]
lngs = [x[1] for x in polygon]
max_lat = max(lats)
min_lat = min(lats)
max_lng = max(lngs)
min_lng = min(lngs)
return ((min_lat, max_lat), (min_lng, max_lng))
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.post(url, params)
def ndfi_change_value(self, asset_id, polygon, rows=5, cols=5):
img = self.mapid2_cmd(asset_id, polygon, rows, cols)
cmd = {
"image": json.dumps(img),
"fields": 'ndfiSum'#','.join(fields)
}
return self._execute_cmd('/value', cmd)
def ndfi_change_value_old(self, polygons, rows=5, cols=5):
""" return how much NDFI has changed in the time period
``polygons`` are a list of closed polygons defined by lat, lon::
[
[ [lat, lon], [lat, lon]...],
[ [lat, lon], [lat, lon]...]
]
for performance reassons bbox greater than 3 degrees are now allowed
"""
# some assertions
for p in polygons:
bbox = self._get_polygon_bbox(p)
if bbox[0][1] - bbox[0][0] > 3.5 or bbox[1][1] - bbox[1][0] > 3.5:
#raise Exception("polygon bbox size must be less than 3 degrees")
pass
cmd = self._NDFI_change_value(
self.last_period,
self.work_period,
polygons,
rows,
cols
)
return self._execute_cmd('/value', cmd)
def _images_for_period(self, period):
cache_key = "%d-%d" %(period['start'], period['end'])
if cache_key in self._image_cache:
img = self._image_cache[cache_key]
else:
reference_images = self.ee.get("/list?id=%s&starttime=%s&endtime=%s" % (
self.earth_engine_resource,
int(period['start']),
int(period['end'])
))
logging.info(reference_images)
img = [x['id'] for x in reference_images['data']]
self._image_cache[cache_key] = img
return img
def _image_composition(self, image_list):
""" create commands to compose images in google earth engine
ok, i really have NO idea what's going on :)
"""
specs = []
for image in image_list:
name = image.split("_", 2)[-1]
specs.append({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ModisCombiner',
"args": ['MOD09GA_005_' + name, 'MOD09GQ_005_' + name]
});
return specs;
def _baseline_image(self):
return {
'creator': CALL_SCOPE + '/com.google.earthengine.examples.sad.ProdesImage',
'args': ["Xy539pUtkRlazIO1"]
}
def _krig_filter(self, period):
work_month = self.getMidMonth(period['start'], period['end'])
work_year = self.getMidYear(period['start'], period['end'])
end = "%04d%02d" % (work_year, work_month)
filter = [{'property':'Compounddate','equals':int(end)}]
return filter
def _NDFI_image(self, period, long_span=0):
""" given image list from EE, returns the operator chain to return NDFI image """
filter = self._krig_filter(period)
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.NDFIImage',
"args": [{
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.UnmixModis',
"args": [{
"creator": KRIGING,
"args": [ self._MakeMosaic(period, long_span),
{'type':'FeatureCollection','table_id':4468280,
'filter':filter,'mark':str(timestamp())} ]
}]
}]
}
def _change_detection_data(self, reference_period, work_period, polygons=[], cols=5, rows=5):
ndfi_image_1 = self._NDFI_image(reference_period)
ndfi_image_2 = self._NDFI_image(work_period)
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ChangeDetectionData',
"args": [ndfi_image_1,
ndfi_image_2,
self.PRODES_IMAGE,
polygons,
rows,
cols]
}
def _NDFI_change_value(self, reference_period, work_period, polygons, cols=5, rows=5):
""" calc the ndfi change value between two periods inside specified polys
``polygons`` are a list of closed polygons defined by lat, lon::
[
[ [lat, lon], [lat, lon]...],
[ [lat, lon], [lat, lon]...]
]
"""
POLY = []
fields = []
image = self._change_detection_data(reference_period, work_period, [polygons], cols, rows)
return {
"image": json.dumps(image),
"fields": 'ndfiSum'#','.join(fields)
}
def _NDFI_period_image_command(self, period, long_span=0):
""" get NDFI command to get map of NDFI for a period of time """
ndfi_image = self._NDFI_image(period, long_span)
return {
"image": json.dumps(ndfi_image),
"bands": 'vis-red,vis-green,vis-blue',
"gain": 1,
"bias": 0.0,
"gamma": 1.6
}
def _baseline_image_command(self):
baseline_image = self._baseline_image()
return {
"image": json.dumps(baseline_image),
"bands": 'classification',
"min": 0,
"max": 9
}
def _RGB_image_command(self, period):
""" commands for RGB image """
filter = self._krig_filter(period)
return {
"image": json.dumps({
"creator": KRIGING,
"args": [ self._MakeMosaic(period),{'type':'FeatureCollection','table_id':4468280,
'filter':filter,'mark':str(timestamp())} ]
}),
"bands": 'sur_refl_b01,sur_refl_b04,sur_refl_b03',
"gain": 0.1,
"bias": 0.0,
"gamma": 1.6
};
def _MakeMosaic(self, period, long_span=0):
middle_seconds = int((period['end'] + period['start']) / 2000)
this_time = time.gmtime(middle_seconds)
month = this_time[1]
year = this_time[0]
yesterday = date.today() - timedelta(1)
micro_yesterday = time.mktime(yesterday.timetuple()) * 1000000
logging.info("month " + str(month))
logging.info("year " + str(year))
if long_span == 0:
filter = [{'property':'month','equals':month},{'property':'year','equals':year}]
start_time = period['start']
else:
start = "%04d%02d" % (year - 1, month)
end = "%04d%02d" % (year, month)
start_time = period['start'] - 1000 * 60 * 60 * 24 * 365
filter = [{'property':'compounddate','greater_than':start},
{'or': [{'property':'compounddate','less_than':end}, {'property':'compounddate','equals':end}]}]
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.MakeMosaic',
"args": [{"id":"MODIS/MOD09GA","version":micro_yesterday,"start_time":start_time,"end_time":period['end']},
{"id":"MODIS/MOD09GQ","version":micro_yesterday,"start_time":start_time,"end_time":period['end']},
{'type':'FeatureCollection','table_id':1868251,
'filter':filter}, start_time, period['end']]
}
def _SMA_image_command(self, period):
filter = self._krig_filter(period)
return {
"image": json.dumps({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.UnmixModis',
"args": [{
"creator": KRIGING,
"args": [self._MakeMosaic(period), {'type':'FeatureCollection','table_id':4468280,'filter':filter}]
}]
}),
"bands": 'gv,soil,npv',
"gain": 256,
"bias": 0.0,
"gamma": 1.6
};
def _RGB_streched_command(self, period, polygon, sensor, bands):
filter = self._krig_filter(period)
if(sensor=="modis"):
""" bands in format (1, 2, 3) """
bands = "sur_refl_b0%d,sur_refl_b0%d,sur_refl_b0%d" % bands
return {
"image": json.dumps({
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.StretchImage",
"args":[{
"creator":"ClipToMultiPolygon",
"args":[
{
"creator":KRIGING,
"args":[ self._MakeMosaic(period), {'type':'FeatureCollection','table_id':4468280,
'filter':filter}]
},
polygon]},
["sur_refl_b01","sur_refl_b02","sur_refl_b03","sur_refl_b04","sur_refl_b05"],
2
]
}),
"bands": bands
}
else:
three_months = timedelta(days=90)
work_period_end = self.work_period['end']
work_period_start = self.work_period['start'] - 7776000000 #three_months
yesterday = date.today() - timedelta(1)
micro_yesterday = time.mktime(yesterday.timetuple()) * 1000000
landsat_bands = ['10','20','30','40','50','70','80','61','62']
creator_bands =[{'id':id, 'data_type':'float'} for id in landsat_bands]
bands = "%d,%d,%d" % bands
return {
"image": json.dumps({
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.StretchImage",
"args":[{
"creator":"LonLatReproject",
"args":[{
"creator":"SimpleMosaic",
"args":[{
"creator":"LANDSAT/LandsatTOA",
"input":{"id":"LANDSAT/L7_L1T","version":micro_yesterday},
"bands":creator_bands,
"start_time": work_period_start, #131302801000
"end_time": work_period_end }] #1313279999000
},polygon, 30]
},
landsat_bands,
2
]
}),
"bands": bands
}
class NDFI(object):
""" ndfi info for a period of time
"""
# hardcoded data for request
PRODES_IMAGE = {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ProdesImage',
"args": ["PRODES_2009"]
};
MODIS_BANDS = [
'sur_refl_b01_250m', 'sur_refl_b02_250m', 'sur_refl_b03_500m',
'sur_refl_b04_500m', 'sur_refl_b06_500m', 'sur_refl_b07_500m'];
THIS_POLY = None
def __init__(self, ee_res, last_period, work_period):
self.last_period = dict(start=last_period[0],
end=last_period[1])
self.work_period = dict(start=work_period[0],
end=work_period[1])
self.earth_engine_resource = ee_res
self.ee = EarthEngine(settings.EE_TOKEN)
self._image_cache = {}
def paint_deforestation(self, asset_id, month, year):
year_str = "%04d" % (year)
#end = "%04d%02d" % (year, month)
return {
"type": "Image", "creator": "Paint", "args": [asset_id,
{
"table_id": int(settings.FT_TABLE_ID), "type": "FeatureCollection",
"filter":[{"property":"type","equals":7}, {"property":"asset_id","contains":year_str}]},
4]
}
def mapid2_cmd(self, asset_id, polygon=None, rows=5, cols=5):
year_msec = 1000 * 60 * 60 * 24 * 365
month_msec = 1000 * 60 * 60 * 24 * 30
six_months_ago = self.work_period['end'] - month_msec * 6
one_month_ago = self.work_period['end'] - month_msec
last_month = time.gmtime(int(six_months_ago / 1000))[1]
last_year = time.gmtime(int(six_months_ago / 1000))[0]
previous_month = time.gmtime(int(one_month_ago / 1000))[1]
previous_year = time.gmtime(int(one_month_ago / 1000))[0]
work_month = self.getMidMonth(self.work_period['start'], self.work_period['end'])
work_year = self.getMidYear(self.work_period['start'], self.work_period['end'])
end = "%04d%02d" % (work_year, work_month)
start = "%04d%02d" % (last_year, last_month)
previous = "%04d%02d" % (previous_year, previous_month)
start_filter = [{'property':'compounddate','greater_than':start},{'property':'compounddate','less_than':end}]
deforested_asset = self.paint_deforestation(asset_id, work_month, work_year)
json_cmd = {"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.GetNDFIDelta","args": [
self.last_period['start'] - year_msec,
self.last_period['end'],
self.work_period['start'],
self.work_period['end'],
"MODIS/MOD09GA",
"MODIS/MOD09GQ",
{'type':'FeatureCollection','table_id': 1868251, 'mark': str(timestamp()), 'filter':start_filter},
{'type':'FeatureCollection','table_id': 1868251, 'mark': str(timestamp()),
'filter':[{"property":"month","equals":work_month},{"property":"year","equals":work_year}]},
{'type':'FeatureCollection','table_id': 4468280, 'mark': str(timestamp()),
'filter':[{"property":"Compounddate","equals":int(previous)}]},
{'type':'FeatureCollection','table_id': 4468280, 'mark': str(timestamp()),
'filter':[{"property":"Compounddate","equals":int(end)}]},
deforested_asset,
polygon,
rows,
cols]
}
logging.info("GetNDFIDelta")
logging.info(json_cmd)
return json_cmd
def getMidMonth(self, start, end):
middle_seconds = int((end + start) / 2000)
this_time = time.gmtime(middle_seconds)
return this_time[1]
def getMidYear(self, start, end):
middle_seconds = int((end + start) / 2000)
this_time = time.gmtime(middle_seconds)
return this_time[0]
def mapid2(self, asset_id):
cmd = {
"image": json.dumps(self.mapid2_cmd(asset_id)),
"format": 'png'
}
return self._execute_cmd('/mapid', cmd)
def freeze_map(self, asset_id, table, report_id):
"""
"""
base_image = {"creator": CALL_SCOPE + "/com.google.earthengine.examples.sad.ProdesImage", "args":[asset_id]};
remapped = {"algorithm": "Image.remap", "image":base_image,
"from":[0,1,2,3,4,5,6,7,8,9], "to":[0,1,2,3,4,5,6,2,3,9]}
def_image = self.paint_call(remapped, int(report_id), table, 7)
selected_def = {"algorithm": "Image.select", "input": def_image,
"bandSelectors":["remapped"]}
deg_image = self.paint_call(selected_def, int(report_id), table, 8)
renamed_image = {"algorithm": "Image.select", "input": deg_image,
"bandSelectors":["remapped"], "newNames":["classification"]}
clipped_image = {"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.AddBB",
"args":[renamed_image, asset_id, "classification"]}
map_image = {"algorithm": "Image.addBands", "dstImg": asset_id, "srcImg": clipped_image,
"names": ["classification"], "overwrite": True}
cmd = {"value": json.dumps(map_image)}
return self._execute_cmd('/create', cmd)
def paint_call(self, current_asset, report_id, table, value):
return {
"algorithm": "Image.paint", "image": current_asset,
"featureCollection": {
"table_id": int(table), "type": "FeatureCollection",
"filter":[{"property":"report_id","equals":int(report_id)},
{"property":"type", "equals": value}]},
"color": value
}
def rgbid(self):
""" return params to access NDFI rgb image """
# get map id from EE
params = self._RGB_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def smaid(self):
""" return params to access NDFI rgb image """
# get map id from EE
params = self._SMA_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def ndfi0id(self):
# get map id from EE set long_span=1
params = self._NDFI_period_image_command(self.last_period, 1)
return self._execute_cmd('/mapid', params)
def baseline(self, asset_id):
params = self._baseline_image_command(asset_id)
return self._execute_cmd('/mapid', params)
def rgb0id(self):
quarter_msec = 1000 * 60 * 60 * 24 * 90
last_start = self.last_period['start']
last_period = dict(start=last_start - quarter_msec,
end=self.last_period['end'])
params = self._RGB_image_command(last_period)
return self._execute_cmd('/mapid', params)
def ndfi1id(self):
# get map id from EE
params = self._NDFI_period_image_command(self.work_period)
return self._execute_cmd('/mapid', params)
def rgb_strech(self, polygon, sensor, bands):
# this is an special call, the application needs to call /value
# before call /mapid in order to google earthn engine makes his work
cmd = self._RGB_streched_command(self.work_period, polygon, sensor, bands)
del cmd['bands']
if (sensor=="modis"):
cmd['fields'] = 'stats_sur_refl_b01,stats_sur_refl_b02,stats_sur_refl_b03,stats_sur_refl_b04,stats_sur_refl_b05'
else:
cmd['fields'] = 'stats_30,stats_20,stats_10'
self._execute_cmd('/value', cmd)
cmd = self._RGB_streched_command(self.work_period, polygon, sensor, bands)
return self._execute_cmd('/mapid', cmd)
def _get_polygon_bbox(self, polygon):
lats = [x[0] for x in polygon]
lngs = [x[1] for x in polygon]
max_lat = max(lats)
min_lat = min(lats)
max_lng = max(lngs)
min_lng = min(lngs)
return ((min_lat, max_lat), (min_lng, max_lng))
def _execute_cmd(self, url, cmd):
params = "&".join(("%s=%s"% v for v in cmd.iteritems()))
return self.ee.post(url, params)
def ndfi_change_value(self, asset_id, polygon, rows=5, cols=5):
img = self.mapid2_cmd(asset_id, polygon, rows, cols)
cmd = {
"image": json.dumps(img),
"fields": 'ndfiSum'#','.join(fields)
}
return self._execute_cmd('/value', cmd)
def ndfi_change_value_old(self, polygons, rows=5, cols=5):
""" return how much NDFI has changed in the time period
``polygons`` are a list of closed polygons defined by lat, lon::
[
[ [lat, lon], [lat, lon]...],
[ [lat, lon], [lat, lon]...]
]
for performance reassons bbox greater than 3 degrees are now allowed
"""
# some assertions
for p in polygons:
bbox = self._get_polygon_bbox(p)
if bbox[0][1] - bbox[0][0] > 3.5 or bbox[1][1] - bbox[1][0] > 3.5:
#raise Exception("polygon bbox size must be less than 3 degrees")
pass
cmd = self._NDFI_change_value(
self.last_period,
self.work_period,
polygons,
rows,
cols
)
return self._execute_cmd('/value', cmd)
def _images_for_period(self, period):
cache_key = "%d-%d" %(period['start'], period['end'])
if cache_key in self._image_cache:
img = self._image_cache[cache_key]
else:
reference_images = self.ee.get("/list?id=%s&starttime=%s&endtime=%s" % (
self.earth_engine_resource,
int(period['start']),
int(period['end'])
))
logging.info(reference_images)
img = [x['id'] for x in reference_images['data']]
self._image_cache[cache_key] = img
return img
def _image_composition(self, image_list):
""" create commands to compose images in google earth engine
ok, i really have NO idea what's going on :)
"""
specs = []
for image in image_list:
name = image.split("_", 2)[-1]
specs.append({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ModisCombiner',
"args": ['MOD09GA_005_' + name, 'MOD09GQ_005_' + name]
});
return specs;
def _baseline_image(self, asset_id):
classification = {"algorithm":"Image.select",
"input":{"type":"Image", "id":asset_id},
"bandSelectors":["classification"]}
mask = {"algorithm":"Image.eq",
"image1":classification,
"image2":{"algorithm":"Constant","value":4}}
image = {"algorithm":"Image.mask", "image":classification, "mask":mask}
return image
def _krig_filter(self, period):
work_month = self.getMidMonth(period['start'], period['end'])
work_year = self.getMidYear(period['start'], period['end'])
end = "%04d%02d" % (work_year, work_month)
filter = [{'property':'Compounddate','equals':int(end)}]
return filter
def _NDFI_image(self, period, long_span=0):
""" given image list from EE, returns the operator chain to return NDFI image """
filter = self._krig_filter(period)
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.NDFIImage',
"args": [{
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.UnmixModis',
"args": [{
"creator": KRIGING,
"args": [ self._MakeMosaic(period, long_span),
{'type':'FeatureCollection','table_id':4468280,
'filter':filter,'mark':str(timestamp())} ]
}]
}]
}
def _change_detection_data(self, reference_period, work_period, polygons=[], cols=5, rows=5):
ndfi_image_1 = self._NDFI_image(reference_period)
ndfi_image_2 = self._NDFI_image(work_period)
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.ChangeDetectionData',
"args": [ndfi_image_1,
ndfi_image_2,
self.PRODES_IMAGE,
polygons,
rows,
cols]
}
def _NDFI_change_value(self, reference_period, work_period, polygons, cols=5, rows=5):
""" calc the ndfi change value between two periods inside specified polys
``polygons`` are a list of closed polygons defined by lat, lon::
[
[ [lat, lon], [lat, lon]...],
[ [lat, lon], [lat, lon]...]
]
"""
POLY = []
fields = []
image = self._change_detection_data(reference_period, work_period, [polygons], cols, rows)
return {
"image": json.dumps(image),
"fields": 'ndfiSum'#','.join(fields)
}
def _NDFI_period_image_command(self, period, long_span=0):
""" get NDFI command to get map of NDFI for a period of time """
ndfi_image = self._NDFI_image(period, long_span)
return {
"image": json.dumps(ndfi_image),
"bands": 'vis-red,vis-green,vis-blue',
"gain": 1,
"bias": 0.0,
"gamma": 1.6
}
def _baseline_image_command(self, asset_id):
baseline_image = self._baseline_image(asset_id)
return {
"image": json.dumps(baseline_image)
}
def _RGB_image_command(self, period):
""" commands for RGB image """
filter = self._krig_filter(period)
return {
"image": json.dumps({
"creator": KRIGING,
"args": [ self._MakeMosaic(period),{'type':'FeatureCollection','table_id':4468280,
'filter':filter,'mark':str(timestamp())} ]
}),
"bands": 'sur_refl_b01,sur_refl_b04,sur_refl_b03',
"gain": 0.1,
"bias": 0.0,
"gamma": 1.6
};
def _MakeMosaic(self, period, long_span=0):
middle_seconds = int((period['end'] + period['start']) / 2000)
this_time = time.gmtime(middle_seconds)
month = this_time[1]
year = this_time[0]
yesterday = date.today() - timedelta(1)
micro_yesterday = time.mktime(yesterday.timetuple()) * 1000000
logging.info("month " + str(month))
logging.info("year " + str(year))
if long_span == 0:
filter = [{'property':'month','equals':month},{'property':'year','equals':year}]
start_time = period['start']
else:
start = "%04d%02d" % (year - 1, month)
end = "%04d%02d" % (year, month)
start_time = period['start'] - 1000 * 60 * 60 * 24 * 365
filter = [{'property':'compounddate','greater_than':start},
{'or': [{'property':'compounddate','less_than':end}, {'property':'compounddate','equals':end}]}]
return {
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.MakeMosaic',
"args": [{"id":"MODIS/MOD09GA","version":micro_yesterday,"start_time":start_time,"end_time":period['end']},
{"id":"MODIS/MOD09GQ","version":micro_yesterday,"start_time":start_time,"end_time":period['end']},
{'type':'FeatureCollection','table_id':1868251,
'filter':filter}, start_time, period['end']]
}
def _SMA_image_command(self, period):
filter = self._krig_filter(period)
return {
"image": json.dumps({
"creator": CALL_SCOPE + '/com.google.earthengine.examples.sad.UnmixModis',
"args": [{
"creator": KRIGING,
"args": [self._MakeMosaic(period), {'type':'FeatureCollection','table_id':4468280,'filter':filter}]
}]
}),
"bands": 'gv,soil,npv',
"gain": 256,
"bias": 0.0,
"gamma": 1.6
};
def _RGB_streched_command(self, period, polygon, sensor, bands):
filter = self._krig_filter(period)
if(sensor=="modis"):
""" bands in format (1, 2, 3) """
bands = "sur_refl_b0%d,sur_refl_b0%d,sur_refl_b0%d" % bands
return {
"image": json.dumps({
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.StretchImage",
"args":[{
"creator":"ClipToMultiPolygon",
"args":[
{
"creator":KRIGING,
"args":[ self._MakeMosaic(period), {'type':'FeatureCollection','table_id':4468280,
'filter':filter}]
},
polygon]},
["sur_refl_b01","sur_refl_b02","sur_refl_b03","sur_refl_b04","sur_refl_b05"],
2
]
}),
"bands": bands
}
else:
three_months = timedelta(days=90)
work_period_end = self.work_period['end']
work_period_start = self.work_period['start'] - 7776000000 #three_months
yesterday = date.today() - timedelta(1)
micro_yesterday = time.mktime(yesterday.timetuple()) * 1000000
landsat_bands = ['10','20','30','40','50','70','80','61','62']
creator_bands =[{'id':id, 'data_type':'float'} for id in landsat_bands]
bands = "%d,%d,%d" % bands
return {
"image": json.dumps({
"creator":CALL_SCOPE + "/com.google.earthengine.examples.sad.StretchImage",
"args":[{
"creator":"LonLatReproject",
"args":[{
"creator":"SimpleMosaic",
"args":[{
"creator":"LANDSAT/LandsatTOA",
"input":{"id":"LANDSAT/L7_L1T","version":micro_yesterday},
"bands":creator_bands,
"start_time": work_period_start, #131302801000
"end_time": work_period_end }] #1313279999000
},polygon, 30]
},
landsat_bands,
2
]
}),
"bands": bands
}