Merge pull request #64 from knaaptime/simplify_deps

Simplify deps

.travis.yml

@@ -32,7 +32,6 @@ install:
 
 script:
   - if [[ $TRAVIS_JOB_NAME == python-* ]]; then
-      python -c "import os; from tobler.data import store_rasters; store_rasters(os.getcwd())"; 
       travis_wait 45 pytest --cov tobler ;
     fi
 

environment.yml

@@ -3,7 +3,6 @@ channels:
   - conda-forge
   - defaults
 dependencies:
-  - python-dateutil<=2.8.0
   - jupyterlab
   - numpy
   - geopandas
@@ -15,6 +14,5 @@ dependencies:
   - scikit-learn
   - scipy
   - libpysal
-  - quilt3 >=3.1.11
   - tqdm
   - pip

requirements.txt

@@ -5,7 +5,5 @@ rasterio
 scipy
 statsmodels
 rasterstats
-python-dateutil<=2.8.0
-quilt3==3.1.8
 libpysal
 tqdm

requirements_tests.txt

@@ -14,4 +14,5 @@ pytest-mpl
 pytest-cov
 twine
 xgboost
-shap >=0.33
+shap >=0.33
+quilt3 >=3.1.11

tobler/__init__.py

@@ -5,6 +5,5 @@
 
 """
 from . import area_weighted
-from . import data
 from . import dasymetric
 from . import model

tobler/area_weighted/vectorized_raster_interpolation.py

@@ -23,7 +23,6 @@
 
 import statsmodels.formula.api as smf
 from statsmodels.genmod.families import Poisson, Gaussian, NegativeBinomial
-from ..data import fetch_quilt_path
 
 __all__ = [
     "getFeatures",
@@ -70,7 +69,6 @@ def fast_append_profile_in_gdf(geodataframe, raster_path, force_crs_match=True):
     """
 
     _check_presence_of_crs(geodataframe)
-    raster_path = fetch_quilt_path(raster_path)
     if force_crs_match:
         with rasterio.open(raster_path) as raster:
             # raster =
@@ -237,7 +235,6 @@ def return_weights_from_xgboost(
     2) The pixel value, usually, ranges from 0 to 255. That is why the default of 'n_pixels_option_values' is 256.
     3) The returning weights represent the average of the Shapley's values from each feature.
     """
-    raster_path = fetch_quilt_path(raster_path)
     try:
         import xgboost as xgb
         import shap
@@ -378,7 +375,7 @@ def create_non_zero_population_by_pixels_locations(
         )
 
     else:
-        with rasterio.open(fetch_quilt_path(raster)) as raster:
+        with rasterio.open(raster) as raster:
             with warnings.catch_warnings():
                 warnings.simplefilter("ignore")
                 geodataframe_projected = geodataframe.to_crs(crs=raster.crs.data)
@@ -527,7 +524,6 @@ def calculate_interpolated_population_from_correspondence_table(
 
     final_geodataframe = geodataframe.copy()[["geometry"]]
     pop_final = np.empty(len(geodataframe))
-    raster = fetch_quilt_path(raster)
     with rasterio.open(raster) as raster:
 
         pbar = tqdm(total=len(geodataframe), desc="Estimating target polygon values")

tobler/dasymetric/masked_area_interpolate.py

@@ -23,9 +23,7 @@ def masked_area_interpolate(
     target_df : geopandas.GeoDataFrame
         target geometries that will form the new representation of the input data
     raster : str
-        path to raster file that contains ancillary data. 
-        alternatively a user can pass `ncld_2001` or `nlcd_2011` to use built-in data from the
-        National Land Cover Database
+        path to raster file that contains ancillary data
     codes : list of ints
         list of pixel values that should be considered part of the mask (the default is None).
         If no codes are passed, this defaults to  [21, 22, 23, 24] which are the developed land use
@@ -44,14 +42,14 @@ def masked_area_interpolate(
     Returns
     -------
     geopandas.GeoDataFrame
-        GeoDataFrame with geometries matching the target_df and extensive and intensive 
+        GeoDataFrame with geometries matching the target_df and extensive and intensive
         variables as the columns
 
     """
     if not codes:
         codes = [21, 22, 23, 24]
     if not raster:
-        raster = 'nlcd_2011'
+        raise IOError('You must pass the path to a raster that can be read with rasterio')
 
     if not tables:
         tables = area_tables_raster(

tobler/model/models.py

@@ -17,7 +17,7 @@
 def glm_pixel_adjusted(
     source_df=None,
     target_df=None,
-    raster="nlcd_2011",
+    raster=None,
     raster_codes=None,
     variable=None,
     formula=None,
@@ -37,10 +37,9 @@ def glm_pixel_adjusted(
         geodataframe containing source original data to be represented by another geometry
     target_df : geopandas.GeoDataFrame, required
         geodataframe containing target boundaries that will be used to represent the source data
-    raster : str, required (default="nlcd_2011")
+    raster : str, required
         path to raster file that will be used to input data to the regression model.
         i.e. a coefficients refer to the relationship between pixel counts and population counts.
-        Defaults to 2011 NLCD
     raster_codes : list, required (default =[21, 22, 23, 24])
         list of integers that represent different types of raster cells.
         Defaults to [21, 22, 23, 24] whichare considered developed land types in the NLCD
@@ -60,6 +59,8 @@ def glm_pixel_adjusted(
     """
     if not raster_codes:
         raster_codes = [21, 22, 23, 24]
+    if not raster:
+        raise IOError('You must provide the path to a raster that can be read with rasterio')
 
     # build weights from raster and vector data
     weights = return_weights_from_regression(

tobler/tests/test_pysal_integration.py

@@ -1,8 +1,6 @@
 """lightweight test for pysal metapckage that functions import."""
 
 def test_imports():
-    import quilt3
     from tobler.dasymetric import masked_area_interpolate
     from tobler.area_weighted import area_interpolate
-    from tobler.data import store_rasters
     from tobler.model import glm, glm_pixel_adjusted