Various fixes (#2)

Makefile

@@ -1,12 +1,8 @@
 # List of Python demos (without file extenion) from the repo `demo` to include in the jupyterbook. 
-# These files should be listed in `docs/_toc.yml`
-DEMOS = "assembly_strategies"
 
 # We use --set-kernel with jupytext to make it possible for binder to pick it up
 doc: # Generate Sphinx HTML documentation, including API docs 
-	@for demo in ${DEMOS}; do \
-	jupytext --to=ipynb --set-kernel=python3 demo/$$demo.py --output=docs/$$demo.ipynb ;\
-	done
+	jupytext --pre-commit --sync --set-kernel=python3
 	jupyter book build -W docs
 
 clean-pytest: # Remove output from pytest

demo/assembly_bcs.py

@@ -1,7 +1,7 @@
 # ---
 # jupyter:
 #   jupytext:
-#     formats: ipynb,py:light
+#     cell_metadata_filter: -all
 #     text_representation:
 #       extension: .py
 #       format_name: light
@@ -24,7 +24,9 @@
 # 1. Using matrix-vector products of pre-assembled matrices to compute the RHS
 # 2. Assemble the matrix and vector separately
 # We start by importing the necessary modules
+#
 
+# This is a test to check notebooks
 import matplotlib.pyplot as plt
 import time
 import seaborn
@@ -287,30 +289,31 @@ def create_plot(results: dict, outfile: str):
         plt.figure()
         df_rhs = df[df["side"] == "rhs"]
         plot = seaborn.catplot(data=df_rhs, kind="swarm",  x="label",
-                               y="time (s)", hue="method")
+                               y="time (s)", hue="method", title="RHS assembly")
         plot.set(yscale="log")
         plt.grid()
         plt.savefig(f"{outfile}_rhs.png")
+
         plt.figure()
         df_lhs = df[df["side"] == "lhs"]
         plot = seaborn.catplot(data=df_lhs, kind="swarm",  x="label",
-                               y="time (s)", hue="method")
+                               y="time (s)", hue="method", title="LHS assembly")
         plot.set(yscale="log")
         plt.grid()
         plt.savefig(f"{outfile}_lhs.png")
+
         plt.figure()
         df_total = df[df["side"] == "total"]
         plot = seaborn.catplot(data=df_total, kind="swarm",  x="label",
-                               y="time (s)", hue="method")
+                               y="time (s)", hue="method", title="Total assembly")
         plot.set(yscale="log")
         plt.grid()
         plt.savefig(f"{outfile}_total.png")
 
-# -
-
 
+# -
 # We start by running the comparison for an increasing number of degrees of freedom on a fixed grid.
 if __name__ == "__main__":
-    N = 100
+    N = 40
     results_p = run_parameter_sweep(N, N, N, repeats=3, min_degree=1, max_degree=3)
     create_plot(results_p, "results")

demo/assembly_strategies.py

@@ -1,7 +1,7 @@
 # ---
 # jupyter:
 #   jupytext:
-#     formats: ipynb,py:light
+#     cell_metadata_filter: -all
 #     text_representation:
 #       extension: .py
 #       format_name: light
@@ -187,12 +187,11 @@ def create_plot(results: dict, outfile: str):
         import matplotlib.pyplot as plt
         plt.grid()
         plt.savefig(outfile)
-
-
 # -
 
 # We start by running the comparison for an increasing number of degrees of freedom on a fixed grid.
 
+
 if __name__ == "__main__":
     results_p = run_parameter_sweep(30, 25, 23, repeats=3, min_degree=1, max_degree=4)
     create_plot(results_p, "P_sweep.png")
@@ -203,8 +202,8 @@ def create_plot(results: dict, outfile: str):
 # We note that the run for $P=1$ is relatively small, and therefore run $P=1$ on a larger mesh
 
 if __name__ == "__main__":
-    results_p1 = run_parameter_sweep(100, 100, 80, 3, 1, 1)
+    results_p1 = run_parameter_sweep(50, 40, 45, 3, 1, 1)
     create_plot(results_p1, "P1.png")
 
-# We observe that the run-time of both strategies for $P=1$ are more or less the
+# We observe that the run-time of both strategies for $P = 1$ are more or less the
 # same for larger matrices.

docs/_toc.yml

@@ -6,6 +6,7 @@ parts:
     chapters:
     - file: "splitting_schemes"
     - file: "assembly_strategies"
+    - file: "assembly_bcs"
 
   - caption: Python API
     chapters: