Crop / non-crop labels in China

cropharvest/config.py

@@ -16,7 +16,7 @@
 EXPORT_END_MONTH = 2
 EXPORT_END_DAY = 1
 
-DATASET_VERSION_ID = 7257688
+DATASET_VERSION_ID = 10251170
 DATASET_URL = f"https://zenodo.org/record/{DATASET_VERSION_ID}"
 LABELS_FILENAME = "labels.geojson"
 FEATURES_DIR = "features"
@@ -46,4 +46,4 @@
     ),
 }
 
-TEST_DATASETS = {"Togo": "togo-eval"}
+TEST_DATASETS = {"Togo": "togo-eval", "People's Republic of China": "china-crop"}

process_labels/datasets.py

@@ -211,6 +211,11 @@
         ),
         "externally_contributed": False,
     },
+    "china-crop": {
+        "function": loading_funcs.load_china,
+        "description": "Hand-labelled crop / non crop labels in China",
+        "externally_contributed": False,
+    },
 }
 
 

process_labels/loading_funcs/__init__.py

@@ -15,6 +15,7 @@
 from .canada import load_canada
 from .germany import load_germany
 from .jecam import load_jecam
+from .china import load_china
 
 
 __all__ = [
@@ -44,4 +45,5 @@
     "load_tanzania_ecaas",
     "load_tanzania_ceo",
     "load_jecam",
+    "load_china",
 ]

process_labels/loading_funcs/china.py

@@ -0,0 +1,115 @@
+import pandas as pd
+import geopandas
+from datetime import datetime
+
+from cropharvest.config import EXPORT_END_MONTH, EXPORT_END_DAY
+from cropharvest.columns import RequiredColumns
+
+from .utils import LATLON_CRS
+from ..utils import DATASET_PATH
+
+
+def raw_from_ceo():
+    # this is the code used to go from the CEO (unanonymised) labels
+    # to the csv in `raw_data`. While we keep this here for posterity,
+    # the original data is not uploaded to git
+
+    def combine_csvs(format_name: str, year: int, name_2=None):
+        assert str(year) in format_name
+        if name_2 is not None:
+            df = join_on_agreement(pd.read_csv(format_name), pd.read_csv(name_2))
+        else:
+            df = join_on_agreement(
+                pd.read_csv(format_name.format(set_val=1)),
+                pd.read_csv(format_name.format(set_val=2)),
+            )
+        df["year"] = year
+        return df
+
+    def join_on_agreement(a, b):
+        def clean_df(a):
+            crop_column = None
+            for crop_column_name in [
+                "Does this pixel contain active cropland?",
+                "Does this point fall within active cropland?",
+                "Does this point lie on active cropland?",
+                "Does this point contain active cropland?",
+            ]:
+                if crop_column_name in a.columns:
+                    crop_column = crop_column_name
+            assert crop_column is not None
+            a = a[["lon", "lat", crop_column, "collection_time"]]
+            a = a.rename(columns={crop_column: "crop", "collection_time": "collection_date"})
+            return a
+
+        a = clean_df(a)
+        b = clean_df(b)
+        joined = a.merge(b, on=["lat", "lon"], how="inner", suffixes=("", "_y"))
+        return joined[joined.crop == joined.crop_y][["lon", "lat", "crop", "collection_date"]]
+
+    dfs = []
+    for filename, name_2, year in [
+        (
+            "ceo-Liaoning-2019-April---November-(Set-{set_val})-sample-data-2022-06-27.csv",
+            None,
+            2019,
+        ),
+        (
+            "ceo-Liaoning-2019-April---November-(Set-{set_val})-sample-data-2022-06-27.csv",
+            None,
+            2019,
+        ),
+        (
+            "ceo-Jilin-2017-(Set-2)-sample-data-2021-04-15.csv",
+            "ceo-Jilin-2017-(Set-1)-sample-data-2021-04-19.csv",
+            2017,
+        ),
+        ("ceo-HLJ-2019-(Set-{set_val})---v3-sample-data-2022-01-21.csv", None, 2019),
+        ("ceo-Heilongjiang-2016-(Set-{set_val})---v2-sample-data-2022-01-21.csv", None, 2016),
+        ("ceo-Heilongjiang-2017-(Set-{set_val})---v2-sample-data-2022-01-21.csv", None, 2017),
+        ("ceo-Heilongjiang-2018-(Set-{set_val})-sample-data-2021-10-26.csv", None, 2018),
+        (
+            "ceo-Jilin-2016-(April-November)-(Set-{set_val})-v2-sample-data-2022-06-27.csv",
+            None,
+            2016,
+        ),
+        (
+            "ceo-Jilin-2018-(April---November)-(Set-{set_val})-sample-data-2022-06-27.csv",
+            None,
+            2018,
+        ),
+        ("ceo-Jilin-2019-April---November-(Set-{set_val})-sample-data-2022-06-27.csv", None, 2019),
+        (
+            "ceo-Liaoning-2016-April---November-(Set-{set_val})-sample-data-2022-06-27.csv",
+            None,
+            2016,
+        ),
+        ("ceo-Liaoning-2017-(Set-{set_val})-sample-data-2021-04-09.csv", None, 2017),
+        (
+            "ceo-Liaoning-2018-April---November-(Set-{set_val})-sample-data-2022-06-27.csv",
+            None,
+            2018,
+        ),
+        (
+            "ceo-Liaoning-2019-April---November-(Set-{set_val})-sample-data-2022-06-27.csv",
+            None,
+            2019,
+        ),
+    ]:
+        dfs.append(combine_csvs(filename, year, name_2))
+    return pd.concat(dfs)
+
+
+def load_china() -> geopandas.GeoDataFrame:
+    df = pd.read_csv(DATASET_PATH / "china" / "combined_and_anonymised_points.csv")
+    gdf = geopandas.GeoDataFrame(
+        data=df, geometry=geopandas.points_from_xy(df.lon, df.lat), crs=LATLON_CRS
+    )
+    gdf = df.reset_index(drop=True)
+    gdf[RequiredColumns.IS_CROP] = df.apply(lambda x: 1 if x.crop == "Crop" else 0, axis=1)
+    gdf[RequiredColumns.INDEX] = df.index
+    gdf[RequiredColumns.EXPORT_END_DATE] = df.apply(
+        lambda x: datetime(x.year, EXPORT_END_MONTH, EXPORT_END_DAY), axis=1
+    )
+    gdf[RequiredColumns.COLLECTION_DATE] = pd.to_datetime(df.collection_date)
+    return gdf