Merge pull request #399 from nasaharvest/Add-data-UGA_2022_GEE

Created new dataset

data/datasets.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: d70a2e3e7b64b45737ab212c869b76e1.dir
-  size: 693418938
-  nfiles: 54
+- md5: add693a42b38c020b895e8e7c855210d.dir
+  size: 709637820
+  nfiles: 55
   path: datasets
   hash: md5

data/raw.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: 03ecf7c484499852ddf368170fa7201e.dir
-  size: 444911361
-  nfiles: 393
+- md5: 7b5c138e97f360400e6b3fc080d896e2.dir
+  size: 445019716
+  nfiles: 394
   path: raw
   hash: md5

data/report.txt

@@ -450,3 +450,10 @@ eo_data_complete    1000
 ✔ training amount: 387, positive class: 21.4%
 ✔ validation amount: 294, positive class: 15.3%
 ✔ testing amount: 319, positive class: 14.1%
+
+
+
+UgandaNorthCorLabel2022 (Timesteps: 24)
+----------------------------------------------------------------------------
+eo_data_complete    2975
+✔ training amount: 2975, positive class: 51.6%

datasets.py

@@ -540,6 +540,18 @@ def load_labels(self) -> pd.DataFrame:
         return df
 
 
+class UgandaNorthCorLabel2022(LabeledDataset):
+    def load_labels(self) -> pd.DataFrame:
+        raw_folder = raw_dir / "Uganda_North_2022_GEE_labels"
+        df = pd.read_csv(raw_folder / "UGA_2022_labels_from_GEE.csv")
+        df.rename(columns={"lat": LAT, "long": LON}, inplace=True)
+        df = df.drop_duplicates(subset=[LAT, LON]).reset_index(drop=True)
+        df[CLASS_PROB] = (df["class_probability"] == 1).astype(int)
+        df[START], df[END] = date(2022, 1, 1), date(2023, 12, 31)
+        df[SUBSET] = "training"
+        return df
+
+
 datasets: List[LabeledDataset] = [
     CustomLabeledDataset(
         dataset="geowiki_landcover_2017",
@@ -1327,6 +1339,7 @@ def load_labels(self) -> pd.DataFrame:
     Uganda_NorthCEO2022(),
     Uganda_NorthCEO2021(),
     UgandaNorthCEO2019(),
+    UgandaNorthCorLabel2022(),
 ]
 
 if __name__ == "__main__":