Start sketching out the first tutorial

doc/tutorial/first-grid.rst

@@ -1,83 +1,41 @@
 Making your first grid
 ======================
 
-TODO:
-
-* Use an actual projection here
-* Simplify the intro tutorial. It should grid using geographic coordinates without projections and no blocking. Then refer to the tutorial.
-* This one produces cartesian grids.
-* Next tutorial will show how to make geographic grids
-
-Intro:
-
-* Load the new carribean or hawaiian bathymetry data data
-* Fit a spline or Cubic to the raw geographic data
-* Make a grid
-* Make a profile
-
-Tutorial structure:
-
-1. Make a grid in Cartesian coordinates using block reduce, projection, and Spline with defaults. Use new bathymetry data.
-2. Make a geographic grid by passing the projection to the gridder. Explain a bit of what is going on. Also use the bathymetry data.
-3. Use weights when we have uncertainties. How to use them in BlockMean. How to use them in Spline. Use GPS vertical data. All geographic.
-3. Make a Chain with blockmean and spline.
-3. Using cross-validation to find the model performance. Use train-test-split. Use block kfold. Use GPS data. Use a chain. All geographic.
-4. Model selection by grid search. Find the best spline damping. Use GPS data. All geographic with a Chain.
-5. Interpolating vectors with Vector and vectorspline. Use all of the above already.
-
-
-How-tos:
-
-* Interpolate large datasets. Use KNearest.
-* Project a grid.
-* Slice points to a given window.
-* Mask points according to distance.
-* Mask points out of the convexhull.
-* Decimate large datasets. Run block reduce on the volcano lidar.
-* Estimate a polinomial trend.
-* Bin statistics. Calculate standard deviation within blocks on volcano lidar. Measure of roughness.
-* Split spatial data by blocks. Run block_split.
-* Split points along rollowing windows.
-
-Explanations:
-
-* The different types of grid and line coordinates.
-* How spline interpolation works. Build the Jacobian and show how to solve the system.
-* Conventions
-
 .. jupyter-execute::
 
    import verde as vd
    import ensaio
    import matplotlib.pyplot as plt
    import numpy as np
    import pandas as pd
+   import pyproj
 
-   path_to_data = ensaio.fetch_bushveld_gravity(version=1)
+   path_to_data = ensaio.fetch_caribbean_bathymetry(version=2)
    data = pd.read_csv(path_to_data)
-   data = data.assign(easting=data.longitude * 111e3, northing=data.latitude * 111e3)
    data
 
 
 .. jupyter-execute::
 
-   spacing = 5e3
-   print(np.median(vd.median_distance((data.easting, data.northing)))
-   elevation = data.height_sea_level_m
-   coordinates = data.easting, data.northing
+   projection = pyproj.Proj(proj="merc", lat_ts=data.latitude.mean())
+   easting, northing = projection(easting, northing)
 
 .. jupyter-execute::
 
-   plt.scatter(coordinates[0], coordinates[1], c=elevation, s=1)
-   plt.colorbar()
-   plt.show()
+   spacing = 10e3
+   blockmedian = vd.BlockReduce(np.median, spacing=spacing)
+   coordinates, bathymetry = blockmedian.filter((easting, northing), data.bathymetry_m)
+   print(bathymetry.size)
 
 .. jupyter-execute::
 
-   gridder = vd.Spline()
-
-   gridder.fit(coordinates, elevation)
+   plt.scatter(coordinates[0], coordinates[1], c=bathymetry, s=1)
+   plt.colorbar()
+   plt.show()
 
-   grid = gridder.grid(spacing=spacing, data_names="elevation")
+.. jupyter-execute::
 
-   grid.elevation.plot()
+   # gridder = vd.Spline()
+   # gridder.fit(coordinates, elevation)
+   # grid = gridder.grid(spacing=spacing, data_names="elevation")
+   # grid.elevation.plot()

env/requirements-docs.txt

@@ -10,4 +10,4 @@ pyproj
 pygmt==0.11.*
 gmt==6.5.*
 ipython
-ensaio
+ensaio>=0.6.0

environment.yml

@@ -37,7 +37,7 @@ dependencies:
     - pygmt==0.11.*
     - gmt==6.5.*
     - ipython
-    - ensaio
+    - ensaio>=0.6.0
     # Style
     - black
     - pathspec
@@ -49,8 +49,5 @@ dependencies:
     - flake8-mutable
     - flake8-rst-docstrings
     - flake8-simplify
+    - flake8-unused-arguments
     - pep8-naming
-    - pip:
-        # Install flake8-unused-arguments through pip
-        # (not available through conda yet)
-        - flake8-unused-arguments