Skip to content

Commit

Permalink
style: apply black formatter
Browse files Browse the repository at this point in the history
  • Loading branch information
@dfguerrerom
dfguerrerom committed Nov 12, 2021
1 parent 6014672 commit 5dd2370
Show file tree
Hide file tree
Showing 9 changed files with 150 additions and 147 deletions.
225 changes: 111 additions & 114 deletions component/model/model.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,31 +21,31 @@ class MgciModel(Model):
# output parameters
year = Unicode("", allow_none=True).tag(sync=True)
source = Unicode("", allow_none=True).tag(sync=True)

# Custom
custom_lulc = Bool(allow_none=True).tag(sync=True)
# LCLU Classes (came from the reclassify model)
lulc_classes = Dict(allow_none=True).tag(sync=True)
rsa = Bool(False, allow_none=True).tag(sync=True)
rsa = Bool(False, allow_none=True).tag(sync=True)

# Observation variables
kapos_done = Int(0).tag(sync=True)
task_file = Unicode('', allow_none=True).tag(sync=True)

task_file = Unicode("", allow_none=True).tag(sync=True)

# Results
summary_df = Any(allow_none=True).tag(sync=True)

@su.need_ee
def __init__(self, aoi_model):
"""
Parameters:
Dashboard
---------
results_file (str): file containing a task .csv file with name and task_id
"""

self.kapos_image = None
Expand All @@ -54,7 +54,7 @@ def __init__(self, aoi_model):

# Styled results dataframe
self.mgci_report = None

# Save the GEE reduce to region json proces
self.reduced_process = None

Expand Down Expand Up @@ -102,72 +102,71 @@ def get_kapos(self):
.where(aoi_dem.gte(300).And(aoi_dem.lt(1000)).And(local_range.gt(300)), 6)
.selfMask()
)

# # Get the areas inside the other classes but which doesn't belong to them.
# inverse = kapos_1_6.unmask().Not().eq(1).selfMask()

# # Add the inverse band to the 1-6 classes
# all_kapos = kapos_1_6.addBands(inverse).reduce(ee.Reducer.Max())

# # Create neighborhoods, reduce them and update the mask
# # It will get the inverse areas with a border (the surrounding class)
# reduced_neighbors = (
# all_kapos.neighborhoodToBands(ee.Kernel.plus(1))
# .reduce(ee.Reducer.Max())
# .updateMask(inverse)
# )

# # Separate and clump individual 'objects'
# inverse_connected = inverse.connectedComponents({
# connectedness: ee.Kernel.plus(1),
# maxSize: 1024
# }).updateMask(inverse)

# # Calculate its area
# connected_size = inverse_connected.select('labels')
# .connectedPixelCount({
# maxSize: 1024,
# eightConnected: False
# })

# # Filter them by its area
# connected_area = (
# ee.Image.pixelArea()
# .addBands(connected_size)
# .lte(25000000)
# .eq(1)
# .selfMask()
# .select('labels')
# )

# # Remove the objects which are greather than the given area
# inverse_connected = inverse_connected.updateMask(connected_area)

# # reduce by the maximum value from the reduced_neighbors(with borders).
# reduced_neighbors = reduced_neighbors.addBands(
# inverse_connected.select('labels')
# )

# result = reduced_neighbors.reduceConnectedComponents({
# reducer: ee.Reducer.max(),
# labelBand: 'labels'
# })

# # Merge all classes
# self.kapos_image = kapos_1_6.addBands(result).reduce(ee.Reducer.max())



# # Get the areas inside the other classes but which doesn't belong to them.
# inverse = kapos_1_6.unmask().Not().eq(1).selfMask()

# # Add the inverse band to the 1-6 classes
# all_kapos = kapos_1_6.addBands(inverse).reduce(ee.Reducer.Max())

# # Create neighborhoods, reduce them and update the mask
# # It will get the inverse areas with a border (the surrounding class)
# reduced_neighbors = (
# all_kapos.neighborhoodToBands(ee.Kernel.plus(1))
# .reduce(ee.Reducer.Max())
# .updateMask(inverse)
# )

# # Separate and clump individual 'objects'
# inverse_connected = inverse.connectedComponents({
# connectedness: ee.Kernel.plus(1),
# maxSize: 1024
# }).updateMask(inverse)

# # Calculate its area
# connected_size = inverse_connected.select('labels')
# .connectedPixelCount({
# maxSize: 1024,
# eightConnected: False
# })

# # Filter them by its area
# connected_area = (
# ee.Image.pixelArea()
# .addBands(connected_size)
# .lte(25000000)
# .eq(1)
# .selfMask()
# .select('labels')
# )

# # Remove the objects which are greather than the given area
# inverse_connected = inverse_connected.updateMask(connected_area)

# # reduce by the maximum value from the reduced_neighbors(with borders).
# reduced_neighbors = reduced_neighbors.addBands(
# inverse_connected.select('labels')
# )

# result = reduced_neighbors.reduceConnectedComponents({
# reducer: ee.Reducer.max(),
# labelBand: 'labels'
# })

# # Merge all classes
# self.kapos_image = kapos_1_6.addBands(result).reduce(ee.Reducer.max())

def reduce_to_regions(self):
"""
Reduce land use/land cover image to kapos regions using planimetric or real
surface area
Attributes:
lulc (ee.Image, categorical): Input image to reduce
kapos (ee.Image, categorical): Input region
aoi (ee.FeatureCollection, ee.Geometry): Region to reduce image
Return:
GEE Dicionary process (is not yet executed), with land cover class area
GEE Dicionary process (is not yet executed), with land cover class area
per kapos mountain range
"""

Expand All @@ -180,10 +179,8 @@ def reduce_to_regions(self):
raise Exception(
"Please go to the vegetation descriptor layer and reclassify an image"
)

lulc = self.vegetation_image.select(
[self.vegetation_image.bandNames().get(0)]
)

lulc = self.vegetation_image.select([self.vegetation_image.bandNames().get(0)])

aoi = self.aoi_model.feature_collection.geometry()

Expand All @@ -201,9 +198,9 @@ def reduce_to_regions(self):
.max(lulc.projection().nominalScale())
.getInfo()
)

self.reduced_process = (
image_area.divide(param.UNITS['sqkm'][0])
image_area.divide(param.UNITS["sqkm"][0])
.updateMask(lulc.mask().And(self.kapos_image.mask()))
.addBands(lulc)
.addBands(self.kapos_image)
Expand All @@ -218,82 +215,83 @@ def reduce_to_regions(self):
}
)
)

def task_process(self):
"""
Send the task to the GEE servers and process it in background. This will be
neccessary when the process is timed out.
"""

# Create an unique name (to search after in Drive)
unique_preffix = su.random_string(4).upper()
filename = f'{unique_preffix}_{self.aoi_model.name}_{self.year}.csv'

task = ee.batch.Export.table.toDrive(**{
'collection': ee.FeatureCollection(
filename = f"{unique_preffix}_{self.aoi_model.name}_{self.year}.csv"

task = ee.batch.Export.table.toDrive(
**{
"collection": ee.FeatureCollection(
[ee.Feature(None, self.reduced_process)]
),
'description': Path(filename).stem,
'fileFormat': 'CSV'
})

"description": Path(filename).stem,
"fileFormat": "CSV",
}
)

task.start()

# Create a file containing the task id to track when the process is done.
task_id_file = (param.TASKS_DIR / filename)
task_id_file.write_text(f'{filename}, {task.id}')

task_id_file = param.TASKS_DIR / filename
task_id_file.write_text(f"{filename}, {task.id}")

return filename, task.id, str(task_id_file)

def download_from_task_file(self, task_file):
"""Download result from task file"""

gdrive = cs.GDrive()

# Read and get the first row (it shouldn't be mroeo than one)
filename, task_id = pd.read_csv(task_file, header=None).values.tolist()[0]

# Check if the task is completed
task = gdrive.get_task(task_id.strip())

if task.state == "COMPLETED":

tmp_result_file = Path(param.TASKS_DIR, f'tmp_{filename}')
tmp_result_file = Path(param.TASKS_DIR, f"tmp_{filename}")
gdrive.download_file(filename, tmp_result_file)
return tmp_result_file

elif task.state == "FAILED":
raise Exception(f"The task {Path(filename).stem} failed.")

else:
raise SepalWarning(
f"The task '{Path(filename).stem}' state is: {task.state}."
)



def extract_summary_from_result(self, from_task=False):
"""
From the gee result dictionary, extract the values and give a proper
format in a pd.DataFrame.
Args:
from_task (bool, optional): Wheter the extraction will be from a results
from_task (bool, optional): Wheter the extraction will be from a results
file coming from a task or directly from the fly.
"""

if from_task:
if not self.task_file:
raise Exception("You have to download and select a task file.")

# Dowload from file
result_file = self.download_from_task_file(self.task_file)
result_ = cs.read_from_task(result_file)
result_file.unlink()

else:
result_ = self.reduced_process.getInfo()

class_area_per_kapos = {}
for group in result_["groups"]:

Expand All @@ -319,12 +317,12 @@ def extract_summary_from_result(self, from_task=False):

# kapos classes are the rows and lulc are the columns
df = pd.DataFrame.from_dict(class_area_per_kapos, orient="index")

# Create the kapos class even if is not present in the dataset.
for kapos_range in list(range(1,7)):
for kapos_range in list(range(1, 7)):
if not kapos_range in df.index:
df.loc[kapos_range]=0
df.loc[kapos_range] = 0

df.sort_index(inplace=True)
df["green_area"] = df[param.GREEN_CLASSES].sum(axis=1)
df["krange_area"] = df[param.DISPLAY_CLASSES].sum(axis=1)
Expand Down Expand Up @@ -357,7 +355,7 @@ def get_report(self):

# The following format respects
# https://github.com/dfguerrerom/sepal_mgci/issues/23

if self.summary_df is None:
raise Exception(cm.dashboard.alert.no_summary)

Expand All @@ -367,19 +365,18 @@ def get_report(self):
vegetation_names = {k: v[0] for k, v in self.lulc_classes.items()}
vegetation_columns = list(vegetation_names.values())


# Create the base columns for the statistics dataframe
stats_df = self.summary_df.copy()
# Replace 0 vals with "N" according with requirements.
stats_df.replace(0,NO_VALUE, inplace=True)
stats_df.replace(0, NO_VALUE, inplace=True)
stats_df.rename(columns=vegetation_names, inplace=True)
stats_df[INDICATOR] = INDICATOR_NUM
stats_df[GEOAREANAM] = cs.get_geoarea(self.aoi_model)[0]
stats_df[GEOAREACODE] = cs.get_geoarea(self.aoi_model)[1]
stats_df[TIMEPERIOD] = self.year
stats_df[TIMEDETAIL] = self.year
stats_df[SOURCE] = self.source
stats_df[NATURE] = CUSTOM_ if self.custom_lulc==True else GLOBAL_
stats_df[NATURE] = CUSTOM_ if self.custom_lulc == True else GLOBAL_
stats_df[REPORTING] = REPORTING_VALUE
stats_df[MOUNTAINCLASS] = "C" + stats_df.mgci.index.astype(str)
stats_df = stats_df.reset_index()
Expand All @@ -401,13 +398,13 @@ def get_mgci_report():
mgci_df[SERIESCOD] = SERIESCOD_GRNCVI

return mgci_df[BASE_COLS]
#ER_MTN_GRNCOV

# ER_MTN_GRNCOV
def get_green_cov_report():
"""Returns df with green cover and total mountain area for every mountain
range"""

unit = param.UNITS['sqkm'][1]
unit = param.UNITS["sqkm"][1]

green_area_df = pd.merge(base_df, stats_df, how="right")
green_area_df[VALUE] = stats_df.green_area
Expand Down Expand Up @@ -445,7 +442,7 @@ def get_land_cov_report():
)
# Merge dataframes

unit = param.UNITS['sqkm'][1]
unit = param.UNITS["sqkm"][1]

landcov_df = pd.merge(base_df, melt_df, how="right")
landcov_df[SERIESDESC] = SERIESDESC_TTL.format(unit=unit)
Expand Down
Loading

0 comments on commit 5dd2370

Please sign in to comment.