Clean dependencies

README.md

@@ -31,46 +31,3 @@ https://nlmod.readthedocs.io/.
 <p align="center">
   <img src="docs/_static/logo_10000_2.png" width="256"/>
 </p>
-
-## Installation
-
-Install the module with pip:
-
-`pip install nlmod`
-
-`nlmod` has the following required dependencies: 
-
-* `flopy`
-* `xarray`
-* `netcdf4`
-* `rasterio`
-* `rioxarray`
-* `affine`
-* `geopandas`
-* `owslib`
-* `hydropandas`
-* `shapely`
-* `pyshp`
-* `rtree`
-* `matplotlib`
-* `dask`
-* `colorama`
-* `joblib`
-* `bottleneck`
-
-There are some optional dependecies, only needed (and imported) in a single method.
-Examples of this are `geocube`, `rasterstats` (both used in nlmod.util.zonal_statistics),
-`h5netcdf` (used for hdf5 files backend in xarray), `scikit-image`
-(used in nlmod.read.rws.calculate_sea_coverage).
-To install `nlmod` with the optional dependencies use:
-
-`pip install nlmod[full]`
-
-When using pip the dependencies are automatically installed. Some dependencies are
-notoriously hard to install on certain platforms. Please see the
-[dependencies](https://github.com/ArtesiaWater/hydropandas#dependencies) section of the
-`hydropandas` package for more information on how to install these packages manually.
-
-## Getting started
-
-Start with the Jupyter Notebooks in the examples folder. These notebooks illustrate how to use the `nlmod` package.

docs/getting_started.rst

@@ -103,27 +103,28 @@ potential solutions.
 - flopy
 - xarray
 - netcdf4
-- rasterio
 - rioxarray
-- affine
 - geopandas
-- owslib
-- hydropandas
-- shapely
-- pyshp
-- rtree
 - matplotlib
 - dask
-- colorama
-- joblib
+- requests
+- scipy
 - bottleneck
 
 On top of that there are some optional dependecies:
 
 - geocube (used in nlmod.util.zonal_statistics)
-- rasterstats  (used in nlmod.util.zonal_statistics)
-- h5netcdf (used for the hdf5 backend of xarray)
+- rasterstats (used in nlmod.util.zonal_statistics)
+- contextily (nlmod.plot.add_background_map)
 - scikit-image (used in nlmod.read.rws.calculate_sea_coverage)
-
-These dependencies are only needed (and imported) in a single method or function.
+- py7zr (used in nlmod.read.bofek.download_bofek_gdf)
+- joblib (used in nlmod.cache)
+- colorama (used in nlmod.util.get_color_logger)
+- tqdm (used for showing progress in long-running methods)
+- hydropandas (used in nlmod.read.knmi and nlmod.read.bro)
+- owslib (used in nlmod.read.ahn.get_latest_ahn_from_wcs)
+- pyshp (used in nlmod.grid.refine)
+- h5netcdf (used in nlmod.read.knmi_data_platform)
+
+These dependencies are only needed (and imported) in a single module or method.
 They can be installed using ``pip install nlmod[full]`` or ``pip install -e .[full]``.

nlmod/cache.py

@@ -10,7 +10,6 @@
 import pickle
 
 import flopy
-import joblib
 import numpy as np
 import pandas as pd
 import xarray as xr
@@ -462,7 +461,14 @@ def decorator(*args, cachedir=None, cachename=None, **kwargs):
 
             if pickle_check:
                 # add dataframe hash to function arguments dic
-                func_args_dic["_pklz_hash"] = joblib.hash(cached_pklz)
+                try:
+                    import joblib
+
+                    func_args_dic["_pklz_hash"] = joblib.hash(cached_pklz)
+                except ImportError:
+                    logger.warning(
+                        "joblib is not installed, cannot add dataframe hash to function arguments"
+                    )
 
                 # check if cache was created with same function arguments as
                 # function call
@@ -487,7 +493,14 @@ def decorator(*args, cachedir=None, cachename=None, **kwargs):
             # add dataframe hash to function arguments dic
             with open(fname_cache, "rb") as f:
                 temp = pickle.load(f)
-            func_args_dic["_pklz_hash"] = joblib.hash(temp)
+            try:
+                import joblib
+
+                func_args_dic["_pklz_hash"] = joblib.hash(temp)
+            except ImportError:
+                logger.warning(
+                    "joblib is not installed, cannot add dataframe hash to function arguments"
+                )
 
             # pickle function arguments
             with open(fname_pickle_cache, "wb") as fpklz:
@@ -555,7 +568,9 @@ def _same_function_arguments(func_args_dic, func_args_dic_cache):
         if item is None:
             # Value of None type is always None so the check happens in previous if statement
             pass
-        elif isinstance(item, (numbers.Number, bool, str, bytes, list, tuple, pathlib.PurePath)):
+        elif isinstance(
+            item, (numbers.Number, bool, str, bytes, list, tuple, pathlib.PurePath)
+        ):
             if item != func_args_dic_cache[key]:
                 if key.endswith("_hash") and isinstance(item, str):
                     logger.info(

nlmod/dims/grid.py

@@ -27,9 +27,9 @@
 from shapely.affinity import affine_transform
 from shapely.geometry import Point, Polygon
 from shapely.ops import unary_union
-from tqdm import tqdm
 
 from .. import cache, util
+from ..util import tqdm
 from .layers import (
     fill_nan_top_botm_kh_kv,
     get_first_active_layer,

nlmod/gwf/surface_water.py

@@ -6,7 +6,6 @@
 import pandas as pd
 import xarray as xr
 from shapely.geometry import Polygon
-from tqdm import tqdm
 
 from ..cache import cache_pickle
 from ..dims.grid import (
@@ -17,7 +16,7 @@
 )
 from ..dims.layers import get_idomain
 from ..read import bgt, waterboard
-from ..util import extent_to_polygon, gdf_intersection_join, zonal_statistics
+from ..util import extent_to_polygon, gdf_intersection_join, zonal_statistics, tqdm
 
 logger = logging.getLogger(__name__)
 

nlmod/gwf/wells.py

@@ -4,9 +4,9 @@
 import geopandas as gpd
 import numpy as np
 import pandas as pd
-from tqdm import tqdm
 
 from ..dims.grid import gdf_to_grid
+from ..util import tqdm
 
 logger = logging.getLogger(__name__)
 

nlmod/read/administrative.py

@@ -2,7 +2,6 @@
 
 from . import waterboard, webservices
 from .. import cache
-from . import waterboard, webservices
 
 
 def get_municipalities(*args, **kwargs):

nlmod/read/ahn.py

@@ -17,12 +17,10 @@
 from rasterio.io import MemoryFile
 from requests.exceptions import HTTPError
 from rioxarray.merge import merge_arrays
-from tqdm import tqdm
-
 from .. import NLMOD_DATADIR, cache
 from ..dims.grid import get_extent
 from ..dims.resample import structured_da_to_ds
-from ..util import extent_to_polygon, get_ds_empty
+from ..util import extent_to_polygon, get_ds_empty, tqdm
 from .webservices import arcrest, wcs
 
 logger = logging.getLogger(__name__)

nlmod/read/bofek.py

@@ -7,9 +7,9 @@
 
 import geopandas as gpd
 import requests
-from tqdm import tqdm
 
-from nlmod import cache, util
+from .. import cache, util
+from ..util import tqdm
 
 logger = logging.getLogger(__name__)
 

nlmod/read/bro.py

@@ -1,7 +1,6 @@
 import logging
 import warnings
 
-import hydropandas as hpd
 import numpy as np
 import pandas as pd
 
@@ -186,6 +185,13 @@ def _get_bro_within_extent(extent, name, ignore_max_obs, epsg, **kwargs):
     hpd.ObsCollection
         _description_
     """
+    try:
+        import hydropandas as hpd
+    except ImportError as exc:
+        raise ImportError(
+            "hydropandas is required for nlmod.read.bro.download_bro_groundwater(), "
+            "please install it using 'pip install hydropandas'"
+        ) from exc
     return hpd.read_bro(
         extent, name=name, ignore_max_obs=ignore_max_obs, epsg=epsg, **kwargs
     )

nlmod/read/brt.py

@@ -11,9 +11,8 @@
 import pandas as pd
 import requests
 from shapely.geometry import LineString, MultiPolygon, Point, Polygon
-from tqdm import tqdm
 
-from ..util import extent_to_polygon
+from ..util import extent_to_polygon, tqdm
 
 logger = logging.getLogger(__name__)
 

nlmod/read/knmi.py

@@ -2,18 +2,15 @@
 import logging
 import warnings
 
-import hydropandas as hpd
 import numpy as np
 import pandas as pd
-from hydropandas.io import knmi as hpd_knmi
 
 from .. import cache, util
 from ..dims.grid import get_affine_mod_to_world, is_structured, is_vertex
 from ..dims.layers import get_first_active_layer
 from ..dims.base import get_ds
 from ..dims.shared import get_area
 
-
 logger = logging.getLogger(__name__)
 
 
@@ -325,6 +322,7 @@ def _get_locations_vertex(ds):
     locations = pd.DataFrame(
         index=icell2d_active, data={"x": x, "y": y, "layer": layer}
     )
+    hpd = _import_hydropandas()
     locations = hpd.ObsCollection(locations)
 
     return locations
@@ -354,7 +352,7 @@ def _get_locations_structured(ds):
         affine = get_affine_mod_to_world(ds)
         x, y = affine * (x, y)
     layers = [fal.data[row, col] for row, col in zip(rows, columns)]
-
+    hpd = _import_hydropandas()
     locations = hpd.ObsCollection(
         pd.DataFrame(
             data={"x": x, "y": y, "row": rows, "col": columns, "layer": layers}
@@ -364,6 +362,17 @@ def _get_locations_structured(ds):
     return locations
 
 
+def _import_hydropandas(method="nlmod.read.knmi.download_knmi()"):
+    try:
+        import hydropandas as hpd
+    except ImportError as exc:
+        raise ImportError(
+            f"hydropandas is required for {method}, "
+            "please install it using 'pip install hydropandas'"
+        ) from exc
+    return hpd
+
+
 @cache.cache_pickle
 def download_knmi(
     ds=None,
@@ -478,19 +487,19 @@ def get_locations(ds, oc_knmi=None, most_common_station=False):
         locations = _get_locations_vertex(ds)
     else:
         raise ValueError("gridtype should be structured or vertex")
+    _import_hydropandas(method="nlmod.read.knmi.get_locations()")
+    from hydropandas.io import knmi
 
     if oc_knmi is not None:
-        locations["stn_RD"] = hpd_knmi.get_nearest_station_df(
+        locations["stn_RD"] = knmi.get_nearest_station_df(
             locations, stations=oc_knmi.loc[oc_knmi["meteo_var"] == "RD"]
         )
-        locations["stn_EV24"] = hpd_knmi.get_nearest_station_df(
+        locations["stn_EV24"] = knmi.get_nearest_station_df(
             locations, stations=oc_knmi.loc[oc_knmi["meteo_var"] == "EV24"]
         )
     else:
-        locations["stn_RD"] = hpd_knmi.get_nearest_station_df(locations, meteo_var="RD")
-        locations["stn_EV24"] = hpd_knmi.get_nearest_station_df(
-            locations, meteo_var="EV24"
-        )
+        locations["stn_RD"] = knmi.get_nearest_station_df(locations, meteo_var="RD")
+        locations["stn_EV24"] = knmi.get_nearest_station_df(locations, meteo_var="EV24")
 
     if most_common_station:
         if is_structured(ds):
@@ -530,6 +539,8 @@ def _download_knmi_at_locations(locations, start=None, end=None):
     stns_ev24 = locations["stn_EV24"].unique()
 
     # get knmi data stations closest to any grid cell
+    hpd = _import_hydropandas()
+
     olist = []
     for stnrd in stns_rd:
         o = hpd.PrecipitationObs.from_knmi(

nlmod/read/knmi_data_platform.py

@@ -11,7 +11,7 @@
 import xarray as xr
 from numpy import arange, array, ndarray
 from pandas import Timedelta, Timestamp
-from tqdm import tqdm
+from ..util import tqdm
 
 logger = logging.getLogger(__name__)
 

nlmod/read/rws.py

@@ -8,10 +8,10 @@
 import numpy as np
 import xarray as xr
 from rioxarray.merge import merge_arrays
-from tqdm import tqdm
 
-from nlmod import NLMOD_DATADIR, cache, dims, util
-from nlmod.read.webservices import arcrest
+from .. import NLMOD_DATADIR, cache, dims, util
+from ..util import tqdm
+from .webservices import arcrest
 
 logger = logging.getLogger(__name__)
 

nlmod/read/webservices.py

@@ -7,14 +7,12 @@
 import pandas as pd
 import requests
 import rioxarray
-from owslib.wcs import WebCoverageService
 from rasterio import merge
 from rasterio.io import MemoryFile
 from requests.exceptions import HTTPError
 from shapely.geometry import MultiPolygon, Point, Polygon
-from tqdm import tqdm
 
-# from owslib.wfs import WebFeatureService
+from ..util import tqdm
 
 logger = logging.getLogger(__name__)
 
@@ -174,11 +172,6 @@ def arcrest(
 
                 assert "exceededTransferLimit" not in data, "exceededTransferLimit"
                 data["features"] = features
-
-                df = pd.DataFrame(
-                    [feature["attributes"] for feature in data["features"]]
-                )
-
     return gdf
 
 
@@ -536,6 +529,8 @@ def _download_wcs(extent, res, url, identifier, version, fmt, crs):
         f"- download wcs between: x ({str(extent[0])}, {str(extent[1])}); "
         f"y ({str(extent[2])}, {str(extent[3])})"
     )
+    from owslib.wcs import WebCoverageService
+
     wcs = WebCoverageService(url, version=version)
     if identifier is None:
         identifiers = list(wcs.contents)