BLD: begin porting ANOVA chapter (#11)

* MAINT: update toc function

* BLD: building

* BUG: add d as a good name to pylint

* MAINT: remove patsy as req

.pylintrc

@@ -94,7 +94,7 @@ evaluation=10.0 - ((float(5 * error + warning + refactor + convention) / stateme
 bad-functions=map,filter,input
 
 # Good variable names which should always be accepted, separated by a comma
-good-names=a,b,c,f,i,j,k,df,x,y,y2,_,fig,ax
+good-names=a,b,c,d,f,i,j,k,df,x,y,y2,_,fig,ax
 
 # Bad variable names which should always be refused, separated by a comma
 bad-names=foo,bar,baz,toto,tutu,tata

Makefile

@@ -16,7 +16,7 @@ venv:  # Set up Python virtual environment.
 	python -m venv ${VENV_PATH}; \
 	source ${VENV_PATH}/bin/activate; \
 	pip install -U pip; \
-	pip install -r requirements.txt; \
+	pip install -r requirements-dev.txt; \
 	deactivate; \
 	)
 	@printf "\n\nVirtual environment created! \033[1;34mRun \`source ${VENV_PATH}/bin/activate\` to activate it.\033[0m\n\n\n"

plots.py

@@ -182,3 +182,54 @@ def dummy_coding_plot():
     ax.legend(fontsize="large")
 
     return fig, ax
+
+
+def one_way_anova_plot():
+    a = np.random.normal(0, 1, 20)
+    b = np.random.normal(-2, 1, 20)
+    c = np.random.normal(3, 1, 20)
+    d = np.random.normal(1.5, 1, 20)
+
+    df = pd.DataFrame()
+    df["y"] = np.concatenate([a, b, c, d])
+    df["group_2"] = np.concatenate(
+        [np.zeros_like(b)] + [np.ones_like(b)] + 2 * [np.zeros_like(b)]
+    )
+    df["group_3"] = np.concatenate(
+        2 * [np.zeros_like(c)] + [np.ones_like(c)] + [np.zeros_like(c)]
+    )
+    df["group_4"] = np.concatenate(3 * [np.zeros_like(d)] + [np.ones_like(d)])
+
+    res = smf.ols("y ~ 1 + group_2 + group_3 + group_4", df).fit()
+    beta0, beta1, beta2, beta3 = res.params
+
+    fig, ax = plt.subplots(figsize=[10, 8])
+    ax.scatter(0 * np.ones_like(a), a, color="k")
+    ax.scatter(1 * np.ones_like(b), b, color="k")
+    ax.scatter(2 * np.ones_like(c), c, color="k")
+    ax.scatter(3 * np.ones_like(d), d, color="k")
+
+    ax.axhline(beta0, color="b", label=r"$\beta_0$ (group 1 mean)")
+
+    ax.plot([0.7, 1.3], 2 * [beta0 + beta1], color="navy")
+    ax.plot(
+        [0, 1],
+        [beta0, beta0 + beta1],
+        color="r",
+        label=r"$\beta_1, \beta_2, ...$ (slopes/differences to $\beta_0$)",
+    )
+
+    ax.plot(
+        [1.7, 2.3],
+        2 * [beta0 + beta2],
+        color="navy",
+        label=r"$\beta_0+\beta_1, \beta_0+\beta_2 ...$ (group 2, 3 ... means)",
+    )
+    ax.plot([1, 2], [beta0, beta0 + beta2], color="r")
+
+    ax.plot([2.7, 3.3], 2 * [beta0 + beta3], color="navy")
+    ax.plot([2, 3], [beta0, beta0 + beta3], color="r")
+
+    ax.legend(fontsize="large")
+
+    return fig, ax

requirements-dev.txt

@@ -0,0 +1,5 @@
+-r requirements.txt
+black==19.3b0
+nbdime==1.0.6
+nbinteract==0.2.4
+pylint==2.3.1

requirements.txt

@@ -1,11 +1,6 @@
-black==19.3b0
 jupyter==1.0.0
 matplotlib==3.1.0
-nbdime==1.0.6
-nbinteract==0.2.4
 numpy==1.16.4
 pandas==0.24.2
-patsy==0.5.1
-pylint==2.3.1
 scipy==1.3.0
 statsmodels==0.10.0

utils.py

@@ -58,7 +58,7 @@ def tabulate_results(test_values, ols_results, names, coeff="x"):
     return table
 
 
-def generate_toc(notebook="tests-as-linear.ipynb"):
+def generate_toc(notebook="tests-as-linear.ipynb", max_header_levels=2):
     """
     Generates a table of contents in Markdown.
 
@@ -70,6 +70,9 @@ def generate_toc(notebook="tests-as-linear.ipynb"):
     ----------
     notebook : str
         Path to notebook for which to generate a table of contents.
+    max_header_levels : int
+        Maximum number of header levels to show in table of contents (i.e. the
+        depth of headers to display).
 
     Returns
     -------
@@ -83,7 +86,7 @@ def generate_toc(notebook="tests-as-linear.ipynb"):
     for cell in cells:
         if cell["cell_type"] == "markdown":
             for line in cell["source"]:
-                match = re.search(r"^#+ ", line)
+                match = re.search(r"^[#]{{1,{0}}} ".format(max_header_levels), line)
                 if match:
                     level = len(line) - len(line.lstrip("#"))
                     link = line.strip(" #\n").replace(" ", "-")