feat: Run mypy in strict mode

Can't pull out image type because of
<python/mypy#7178>.

flooding/Sentinel2_Water_Extraction/Sentinel2_Water_Extraction.ipynb

@@ -100,13 +100,43 @@
     "\n",
     "# suppress warnings - raised by external rasterio library\n",
     "import warnings\n",
+    "from datetime import date\n",
+    "from typing import List, Optional, Tuple, TypedDict\n",
     "\n",
+    "import geopandas as gpd\n",
     "import matplotlib.pyplot as plt\n",
     "import numpy as np\n",
+    "import numpy.typing as npt\n",
     "import rasterio\n",
+    "from affine import Affine\n",
+    "from rasterio.crs import CRS\n",
     "from rasterio.plot import show\n",
     "from rio_tiler.io import COGReader\n",
     "from satsearch import Search\n",
+    "from satstac import ItemCollection\n",
+    "from shapely.geometry import shape\n",
+    "from typing_extensions import NotRequired\n",
+    "\n",
+    "if sys.version_info >= (3, 9):\n",
+    "    from typing import Annotated\n",
+    "else:\n",
+    "    from typing_extensions import Annotated\n",
+    "\n",
+    "\n",
+    "class Image(TypedDict):\n",
+    "    \"\"\"Image info being passed around this code.\"\"\"\n",
+    "\n",
+    "    date: date\n",
+    "    src_crs: str\n",
+    "    rgb: npt.NDArray[np.float64]\n",
+    "    mndwi: float\n",
+    "    transform_window: Affine\n",
+    "    image_id: str\n",
+    "    water_mask: NotRequired[gpd.GeoDataFrame]\n",
+    "    water: NotRequired[gpd.GeoDataFrame]\n",
+    "    water_area: int\n",
+    "    area: NotRequired[int]\n",
+    "\n",
     "\n",
     "warnings.filterwarnings(\"ignore\")\n",
     "\n",
@@ -221,7 +251,13 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "def range_request(image_url, crs, bbox, width=None, height=None):\n",
+    "def range_request(\n",
+    "    image_url: str,\n",
+    "    crs: CRS,\n",
+    "    bbox: Annotated[List[float], 4],\n",
+    "    width: Optional[int] = None,\n",
+    "    height: Optional[int] = None,\n",
+    ") -> Tuple[npt.NDArray[np.float64], Affine]:\n",
     "    \"\"\"Request and read just the required pixels from the COG.\"\"\"\n",
     "    with COGReader(image_url) as image:\n",
     "        img = image.part(bbox, width=width, dst_crs=crs, height=height, nodata=-9999)\n",
@@ -242,7 +278,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "def image_search(bbox, date_range):\n",
+    "def image_search(bbox: Annotated[List[float], 4], date_range: str) -> ItemCollection:\n",
     "    \"\"\"Use SatSearch to find all Sentinel-2 images that meet our criteria.\"\"\"\n",
     "    # Note, we are not querying cloud cover and accepting\n",
     "    # all images irrelevant of cloud metadata.\n",
@@ -276,25 +312,25 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "def export_raster(image, image_tpye=\"rgb\"):\n",
+    "def export_raster(image: Image) -> None:\n",
     "    \"\"\"Export GeoTiffs for use in a GIS.\"\"\"\n",
     "    raster_output = os.path.join(output_directory, f\"{image['image_id']}.tif\")\n",
-    "    number_of_bands, height, width = image[image_tpye].shape\n",
+    "    number_of_bands, height, width = image[\"rgb\"].shape\n",
     "\n",
     "    profile = {\n",
     "        \"driver\": \"GTiff\",\n",
     "        \"count\": number_of_bands,\n",
     "        \"height\": height,\n",
     "        \"width\": width,\n",
     "        \"crs\": image[\"src_crs\"],\n",
-    "        \"dtype\": image[image_tpye].dtype,\n",
+    "        \"dtype\": image[\"rgb\"].dtype,\n",
     "        \"transform\": image[\"transform_window\"],\n",
     "        \"nodata\": -9999,\n",
     "        \"photometric\": \"RGB\",\n",
     "    }\n",
     "\n",
     "    with rasterio.open(raster_output, \"w\", **profile) as dst:\n",
-    "        dst.write(image[image_tpye])"
+    "        dst.write(image[\"rgb\"])"
    ]
   },
   {
@@ -317,8 +353,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "images = []\n",
-    "\n",
+    "images: List[Image] = []\n",
     "# Iterate over all observations meeting our search criteria\n",
     "items = image_search(bbox, date_range)\n",
     "for item in items:\n",
@@ -327,7 +362,7 @@
     "    swir = item.asset(\"swir16\")[\"href\"]\n",
     "    rgb = item.asset(\"visual\")[\"href\"]\n",
     "    scl = item.asset(\"SCL\")[\"href\"]\n",
-    "    date = item.date.strftime(\"%d/%m/%Y\")\n",
+    "    date_ = item.date.strftime(\"%d/%m/%Y\")\n",
     "    crs = f\"EPSG:{item.properties['proj:epsg']}\"\n",
     "\n",
     "    # Streamed pixels within bbox\n",
@@ -355,14 +390,12 @@
     "    # Store the data for further processing\n",
     "    images.append(\n",
     "        {\n",
-    "            \"date\": date,\n",
+    "            \"date\": date_,\n",
     "            \"src_crs\": crs,\n",
     "            \"rgb\": rgb_subset,\n",
     "            \"mndwi\": mndwi_subset,\n",
     "            \"transform_window\": transform_window,\n",
     "            \"image_id\": item.id,\n",
-    "            \"water_mask\": None,\n",
-    "            \"water\": None,\n",
     "            \"water_area\": 0,\n",
     "        }\n",
     "    )"
@@ -459,14 +492,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import geopandas as gpd\n",
-    "from shapely.geometry import shape\n",
-    "\n",
     "sieve_threshold = 10\n",
     "\n",
     "for image in images:\n",
     "    # Mask out all values below the mndwi threshold (these are not water)\n",
-    "    water_mask = np.ma.masked_less(image[\"mndwi\"], mndwi_threshold)\n",
+    "    water_mask = np.ma.masked_less(image[\"mndwi\"], mndwi_threshold)  # type: ignore[no-untyped-call]\n",
     "\n",
     "    # Extract the polygons from the mask\n",
     "    water_polygons = rasterio.features.shapes(\n",

pyproject.toml

@@ -3,15 +3,19 @@ profile = "black"
 
 [tool.mypy]
 show_error_codes = true
+strict = true
 
 [[tool.mypy.overrides]]
 module = [
+    "affine", # https://github.com/rasterio/affine/issues/74
     "geopandas", # https://github.com/geopandas/geopandas/issues/1974
     "matplotlib", # https://github.com/matplotlib/matplotlib/issues/20504
     "matplotlib.pyplot", # https://github.com/matplotlib/matplotlib/issues/20504
     "rasterio", # https://github.com/rasterio/rasterio/issues/2322
+    "rasterio.crs", # https://github.com/rasterio/rasterio/issues/2322
     "rasterio.plot", # https://github.com/rasterio/rasterio/issues/2322
-    "satsearch", # Seems to be dead
+    "satsearch", # https://github.com/sat-utils/sat-search/issues/131
+    "satstac", # https://github.com/sat-utils/sat-stac/issues/72
     "shapely.geometry", # https://github.com/shapely/shapely/issues/721
 ]
 ignore_missing_imports = true