Big upload

croplands/io.py

@@ -0,0 +1,37 @@
+import xarray as xr
+from pathlib import Path
+import numpy as np
+from affine import Affine
+from rasterio import CRS
+
+
+def from_spatial_to_profile(zarr):
+  profile = {}
+  profile["driver"] = "GTiff"
+  profile["dtype"] =  zarr.dtypes["band_data"].name
+  profile["width"] = zarr.dims["x"]
+  profile["height"] = zarr.dims["y"]
+  profile["count"] = zarr.dims["band"]
+  profile["crs"] = CRS.from_epsg(zarr["spatial_ref"].attrs["crs_wkt"].split(",")[-1].split("\"")[1])
+  vals = zarr["spatial_ref"].attrs["GeoTransform"].split(" ")
+  profile["transform"] = Affine(float(vals[1]),float(vals[2]),float(vals[0]),float(vals[4]),float(vals[5]),float(vals[3]))
+  return profile
+
+
+def read_zarr(
+    path: Path
+) -> np.ndarray:
+    """Read a zarr file using xarray.
+
+    Args:
+        path (Path): Path to the zarr file.
+    Returns:
+        np.ndarray: zarr data as numpy array.
+    """
+    data = xr.open_zarr(path)["band_data"]
+    return data.values
+
+def read_zarr_profile(path:Path):
+    data = xr.open_zarr(path)
+    profile = from_spatial_to_profile(data)
+    return profile

croplands/polygonize.py

@@ -0,0 +1,59 @@
+import numpy as np
+import rasterio
+import geopandas as gpd
+import cv2
+from shapely.geometry import shape,Polygon
+from shapely.validation import make_valid
+from collections import defaultdict
+
+
+def onehot(components: np.ndarray) -> np.ndarray:
+    oh = np.zeros((components.max() + 1, *components.shape), dtype=np.uint8)
+    for i in range(oh.shape[0]):
+        oh[i][components == i] = 1
+    if 0 in np.unique(components):
+        oh = oh[1:]
+    return oh
+
+def apply_transform(polygon: Polygon, transform: rasterio.Affine) -> Polygon:
+    return Polygon([transform * c for c in polygon.exterior.coords])
+
+def polygonize_raster(
+    data_img:np.ndarray,
+    raster: np.ndarray,
+    tolerance: float,
+    transform: rasterio.transform.Affine = None,
+    crs = None,
+    pixel_threshold: int = 25,
+) -> gpd.GeoDataFrame:
+    data = defaultdict(list)
+    onehot_raster = onehot(raster)
+    if onehot_raster.shape[0] == 0:
+        return None
+    for i in range(onehot_raster.shape[0]):
+        mask = onehot_raster[i]
+        if mask.sum() < pixel_threshold:
+            continue
+
+        contours, _ = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
+        for contour in contours:
+            perimeter = cv2.arcLength(contour, True)
+            area = cv2.contourArea(contour)
+            approx = cv2.approxPolyDP(contour, tolerance * perimeter, True)
+            contour = approx.squeeze()
+            if contour.shape[0] < 3:
+                continue
+            poly = shape({"type": "Polygon", "coordinates": [contour]})
+            if transform is not None:
+                poly = apply_transform(poly, transform)
+
+            most_freq_class = np.argmax(np.bincount(data_img[onehot_raster[i].astype(bool)]))
+            data["geometry"].append(poly)
+            data["component"].append(i)
+            data["class"].append(most_freq_class)
+            data["area"].append(area)
+
+    gdf = gpd.GeoDataFrame(data, crs=crs)
+    gdf["geometry"] = gdf["geometry"].buffer(9.0)
+    gdf["geometry"] = gdf["geometry"].apply(make_valid)
+    return gdf 

croplands/utils.py

@@ -0,0 +1,37 @@
+import numpy as np
+
+def normalize_s2(img: np.ndarray, factor: int = 10_000) -> np.ndarray:
+    """Normalizes the given S2 L2A image dividing the raw reflectance values by
+    the given factor, then clipping the values to the range [0, 1].
+
+    Args:
+        img (np.ndarray): The image to normalize.
+        factor (int, optional): The factor to divide the raw reflectance values
+            by. Defaults to 10_000.
+
+    Returns:
+        np.ndarray: The normalized image.
+    """
+    return np.clip(img / float(factor), 0, 1)
+
+
+def impute_nan(arr):
+    nan_times = np.where(np.isnan(arr).any(-1).any(-1).any(-1))[0]
+
+    for month in nan_times:
+      i = 1
+      if month < 4:
+        while np.isnan(arr[month]).any():
+          if month+i>=6:
+            arr[month][np.isnan(arr[month])] = 0.0
+          else:
+            arr[month][np.isnan(arr[month])] = arr[month+i][np.isnan(arr[month])]
+            i+=1
+      elif month>=4:
+        while np.isnan(arr[month]).any():
+          if month-i<0:
+            arr[month][np.isnan(arr[month])] = 0.0
+          else:
+            arr[month][np.isnan(arr[month])] = arr[month-i][np.isnan(arr[month])]
+            i+=1
+    return arr

examples/at_si_sk_2021_380.zarr/.zattrs

@@ -0,0 +1 @@
+{}

examples/at_si_sk_2021_380.zarr/.zgroup

@@ -0,0 +1,3 @@
+{
+    "zarr_format": 2
+}

examples/at_si_sk_2021_380.zarr/.zmetadata

@@ -0,0 +1,154 @@
+{
+    "metadata": {
+        ".zattrs": {},
+        ".zgroup": {
+            "zarr_format": 2
+        },
+        "band/.zarray": {
+            "chunks": [
+                10
+            ],
+            "compressor": {
+                "blocksize": 0,
+                "clevel": 5,
+                "cname": "lz4",
+                "id": "blosc",
+                "shuffle": 1
+            },
+            "dtype": "<i8",
+            "fill_value": null,
+            "filters": null,
+            "order": "C",
+            "shape": [
+                10
+            ],
+            "zarr_format": 2
+        },
+        "band/.zattrs": {
+            "_ARRAY_DIMENSIONS": [
+                "band"
+            ]
+        },
+        "band_data/.zarray": {
+            "chunks": [
+                1,
+                1,
+                128,
+                128
+            ],
+            "compressor": {
+                "blocksize": 0,
+                "clevel": 5,
+                "cname": "lz4",
+                "id": "blosc",
+                "shuffle": 1
+            },
+            "dtype": "<f4",
+            "fill_value": "NaN",
+            "filters": null,
+            "order": "C",
+            "shape": [
+                6,
+                10,
+                128,
+                128
+            ],
+            "zarr_format": 2
+        },
+        "band_data/.zattrs": {
+            "AREA_OR_POINT": "Area",
+            "_ARRAY_DIMENSIONS": [
+                "time",
+                "band",
+                "y",
+                "x"
+            ],
+            "coordinates": "spatial_ref",
+            "long_name": "stackstac-9f31e5920d53418737c37d57d00f49d8",
+            "resolution": 10.0,
+            "spec": "RasterSpec(epsg=32633, bounds=(389120.0, 5345280.0, 409600.0, 5365760.0), resolutions_xy=(10.0, 10.0))"
+        },
+        "spatial_ref/.zarray": {
+            "chunks": [],
+            "compressor": null,
+            "dtype": "<i8",
+            "fill_value": null,
+            "filters": null,
+            "order": "C",
+            "shape": [],
+            "zarr_format": 2
+        },
+        "spatial_ref/.zattrs": {
+            "GeoTransform": "389115.0 10.0 0.0 5365765.0 0.0 -10.0",
+            "_ARRAY_DIMENSIONS": [],
+            "crs_wkt": "PROJCS[\"WGS 84 / UTM zone 33N\",GEOGCS[\"WGS 84\",DATUM[\"WGS_1984\",SPHEROID[\"WGS 84\",6378137,298.257223563,AUTHORITY[\"EPSG\",\"7030\"]],AUTHORITY[\"EPSG\",\"6326\"]],PRIMEM[\"Greenwich\",0,AUTHORITY[\"EPSG\",\"8901\"]],UNIT[\"degree\",0.0174532925199433,AUTHORITY[\"EPSG\",\"9122\"]],AUTHORITY[\"EPSG\",\"4326\"]],PROJECTION[\"Transverse_Mercator\"],PARAMETER[\"latitude_of_origin\",0],PARAMETER[\"central_meridian\",15],PARAMETER[\"scale_factor\",0.9996],PARAMETER[\"false_easting\",500000],PARAMETER[\"false_northing\",0],UNIT[\"metre\",1,AUTHORITY[\"EPSG\",\"9001\"]],AXIS[\"Easting\",EAST],AXIS[\"Northing\",NORTH],AUTHORITY[\"EPSG\",\"32633\"]]",
+            "false_easting": 500000.0,
+            "false_northing": 0.0,
+            "geographic_crs_name": "WGS 84",
+            "grid_mapping_name": "transverse_mercator",
+            "horizontal_datum_name": "World Geodetic System 1984",
+            "inverse_flattening": 298.257223563,
+            "latitude_of_projection_origin": 0.0,
+            "longitude_of_central_meridian": 15.0,
+            "longitude_of_prime_meridian": 0.0,
+            "prime_meridian_name": "Greenwich",
+            "projected_crs_name": "WGS 84 / UTM zone 33N",
+            "reference_ellipsoid_name": "WGS 84",
+            "scale_factor_at_central_meridian": 0.9996,
+            "semi_major_axis": 6378137.0,
+            "semi_minor_axis": 6356752.314245179,
+            "spatial_ref": "PROJCS[\"WGS 84 / UTM zone 33N\",GEOGCS[\"WGS 84\",DATUM[\"WGS_1984\",SPHEROID[\"WGS 84\",6378137,298.257223563,AUTHORITY[\"EPSG\",\"7030\"]],AUTHORITY[\"EPSG\",\"6326\"]],PRIMEM[\"Greenwich\",0,AUTHORITY[\"EPSG\",\"8901\"]],UNIT[\"degree\",0.0174532925199433,AUTHORITY[\"EPSG\",\"9122\"]],AUTHORITY[\"EPSG\",\"4326\"]],PROJECTION[\"Transverse_Mercator\"],PARAMETER[\"latitude_of_origin\",0],PARAMETER[\"central_meridian\",15],PARAMETER[\"scale_factor\",0.9996],PARAMETER[\"false_easting\",500000],PARAMETER[\"false_northing\",0],UNIT[\"metre\",1,AUTHORITY[\"EPSG\",\"9001\"]],AXIS[\"Easting\",EAST],AXIS[\"Northing\",NORTH],AUTHORITY[\"EPSG\",\"32633\"]]"
+        },
+        "x/.zarray": {
+            "chunks": [
+                128
+            ],
+            "compressor": {
+                "blocksize": 0,
+                "clevel": 5,
+                "cname": "lz4",
+                "id": "blosc",
+                "shuffle": 1
+            },
+            "dtype": "<f8",
+            "fill_value": "NaN",
+            "filters": null,
+            "order": "C",
+            "shape": [
+                128
+            ],
+            "zarr_format": 2
+        },
+        "x/.zattrs": {
+            "_ARRAY_DIMENSIONS": [
+                "x"
+            ]
+        },
+        "y/.zarray": {
+            "chunks": [
+                128
+            ],
+            "compressor": {
+                "blocksize": 0,
+                "clevel": 5,
+                "cname": "lz4",
+                "id": "blosc",
+                "shuffle": 1
+            },
+            "dtype": "<f8",
+            "fill_value": "NaN",
+            "filters": null,
+            "order": "C",
+            "shape": [
+                128
+            ],
+            "zarr_format": 2
+        },
+        "y/.zattrs": {
+            "_ARRAY_DIMENSIONS": [
+                "y"
+            ]
+        }
+    },
+    "zarr_consolidated_format": 1
+}

examples/at_si_sk_2021_380.zarr/band/.zarray

@@ -0,0 +1,20 @@
+{
+    "chunks": [
+        10
+    ],
+    "compressor": {
+        "blocksize": 0,
+        "clevel": 5,
+        "cname": "lz4",
+        "id": "blosc",
+        "shuffle": 1
+    },
+    "dtype": "<i8",
+    "fill_value": null,
+    "filters": null,
+    "order": "C",
+    "shape": [
+        10
+    ],
+    "zarr_format": 2
+}

examples/at_si_sk_2021_380.zarr/band/.zattrs

@@ -0,0 +1,5 @@
+{
+    "_ARRAY_DIMENSIONS": [
+        "band"
+    ]
+}

examples/at_si_sk_2021_380.zarr/band_data/.zarray

@@ -0,0 +1,26 @@
+{
+    "chunks": [
+        1,
+        1,
+        128,
+        128
+    ],
+    "compressor": {
+        "blocksize": 0,
+        "clevel": 5,
+        "cname": "lz4",
+        "id": "blosc",
+        "shuffle": 1
+    },
+    "dtype": "<f4",
+    "fill_value": "NaN",
+    "filters": null,
+    "order": "C",
+    "shape": [
+        6,
+        10,
+        128,
+        128
+    ],
+    "zarr_format": 2
+}

examples/at_si_sk_2021_380.zarr/band_data/.zattrs

@@ -0,0 +1,13 @@
+{
+    "AREA_OR_POINT": "Area",
+    "_ARRAY_DIMENSIONS": [
+        "time",
+        "band",
+        "y",
+        "x"
+    ],
+    "coordinates": "spatial_ref",
+    "long_name": "stackstac-9f31e5920d53418737c37d57d00f49d8",
+    "resolution": 10.0,
+    "spec": "RasterSpec(epsg=32633, bounds=(389120.0, 5345280.0, 409600.0, 5365760.0), resolutions_xy=(10.0, 10.0))"
+}