Merge pull request #429 from nasaharvest/Corrective_labels_Uganda_Nor…

…th_2019

Add_Data_Corrective_labels_Uganda_North_2019

data/datasets.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: a6eb46d364ecfbde721d79e5f545dcde.dir
-  size: 756013135
-  nfiles: 64
+- md5: ffdbfcb7eb9735e7b485676b7d9af37a.dir
+  size: 770113901
+  nfiles: 65
   path: datasets
   hash: md5

data/raw.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: fa2e5d606695b1e4acb3e611d0a9d5d4.dir
-  size: 447402768
-  nfiles: 412
+- md5: 288bc2945bb4453d99720766b16b5815.dir
+  size: 447498504
+  nfiles: 413
   path: raw
   hash: md5

data/report.txt

@@ -537,3 +537,11 @@ eo_data_complete    1000
 ✔ training amount: 387, positive class: 1.3%
 ✔ validation amount: 294, positive class: 1.0%
 ✔ testing amount: 319, positive class: 0.3%
+
+
+
+UgandaNorthCorLabel2019 (Timesteps: 24)
+----------------------------------------------------------------------------
+eo_data_complete     2631
+eo_data_duplicate       1
+✔ training amount: 2631, positive class: 44.6%

datasets.py

@@ -680,6 +680,18 @@ def load_labels(self) -> pd.DataFrame:
         return df
 
 
+class UgandaNorthCorLabel2019(LabeledDataset):
+    def load_labels(self) -> pd.DataFrame:
+        raw_folder = raw_dir / "Uganda_North_2019_GEE_labels"
+        df = pd.read_csv(raw_folder / "Uganda_North_2019_GEE_labels.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(2019, 1, 1), date(2020, 12, 31)
+        df[SUBSET] = "training"
+        return df
+
+
 class KenyaCropArea2019(LabeledDataset):
     def load_labels(self) -> pd.DataFrame:
         raw_folder = raw_dir / "Kenya_Crop_Area_2019"
@@ -1637,6 +1649,7 @@ def load_labels(self) -> pd.DataFrame:
     Uganda_NorthCEO2017(),
     Uganda_NorthCEO2020(),
     Uganda_NorthCEO2018(),
+    UgandaNorthCorLabel2019(),
 ]
 
 if __name__ == "__main__":