Fix documentation example of blob returns for case when ll or lp is n…

…ot finite (#399)

docs/tutorials/autocorr.ipynb

@@ -30,6 +30,7 @@
     "%config InlineBackend.figure_format = \"retina\"\n",
     "\n",
     "from matplotlib import rcParams\n",
+    "\n",
     "rcParams[\"savefig.dpi\"] = 100\n",
     "rcParams[\"figure.dpi\"] = 100\n",
     "rcParams[\"font.size\"] = 20"
@@ -258,7 +259,11 @@
     "for n, ax in zip([10, 100, 1000], axes):\n",
     "    nn = int(true_tau * n)\n",
     "    ax.plot(tau / true_tau, f0, \"k\", label=\"true\")\n",
-    "    ax.plot(tau / true_tau, autocorr_func_1d(y[0, :nn])[: window + 1], label=\"estimate\")\n",
+    "    ax.plot(\n",
+    "        tau / true_tau,\n",
+    "        autocorr_func_1d(y[0, :nn])[: window + 1],\n",
+    "        label=\"estimate\",\n",
+    "    )\n",
     "    ax.set_title(r\"$N = {0}\\,\\tau_\\mathrm{{true}}$\".format(n), fontsize=14)\n",
     "    ax.set_xlabel(r\"$\\tau / \\tau_\\mathrm{true}$\")\n",
     "\n",
@@ -307,7 +312,10 @@
     "    nn = int(true_tau * n)\n",
     "    ax.plot(tau / true_tau, f0, \"k\", label=\"true\")\n",
     "    f = np.mean(\n",
-    "        [autocorr_func_1d(y[i, :nn], norm=False)[: window + 1] for i in range(len(y))],\n",
+    "        [\n",
+    "            autocorr_func_1d(y[i, :nn], norm=False)[: window + 1]\n",
+    "            for i in range(len(y))\n",
+    "        ],\n",
     "        axis=0,\n",
     "    )\n",
     "    f /= f[0]\n",
@@ -447,7 +455,9 @@
     "\n",
     "sampler = emcee.EnsembleSampler(32, 3, log_prob)\n",
     "sampler.run_mcmc(\n",
-    "    np.concatenate((np.random.randn(16, 3), 4.0 + np.random.randn(16, 3)), axis=0),\n",
+    "    np.concatenate(\n",
+    "        (np.random.randn(16, 3), 4.0 + np.random.randn(16, 3)), axis=0\n",
+    "    ),\n",
     "    500000,\n",
     "    progress=True,\n",
     ");"
@@ -584,7 +594,9 @@
     "    # Build the GP model\n",
     "    tau = max(1.0, init / thin)\n",
     "    kernel = terms.RealTerm(\n",
-    "        np.log(0.9 * np.var(z)), -np.log(tau), bounds=[(-5.0, 5.0), (-np.log(N), 0.0)]\n",
+    "        np.log(0.9 * np.var(z)),\n",
+    "        -np.log(tau),\n",
+    "        bounds=[(-5.0, 5.0), (-np.log(N), 0.0)],\n",
     "    )\n",
     "    kernel += terms.RealTerm(\n",
     "        np.log(0.1 * np.var(z)),\n",

docs/tutorials/line.ipynb

@@ -30,6 +30,7 @@
     "%config InlineBackend.figure_format = \"retina\"\n",
     "\n",
     "from matplotlib import rcParams\n",
+    "\n",
     "rcParams[\"savefig.dpi\"] = 100\n",
     "rcParams[\"figure.dpi\"] = 100\n",
     "rcParams[\"font.size\"] = 20"
@@ -429,7 +430,9 @@
     "pos = soln.x + 1e-4 * np.random.randn(32, 3)\n",
     "nwalkers, ndim = pos.shape\n",
     "\n",
-    "sampler = emcee.EnsembleSampler(nwalkers, ndim, log_probability, args=(x, y, yerr))\n",
+    "sampler = emcee.EnsembleSampler(\n",
+    "    nwalkers, ndim, log_probability, args=(x, y, yerr)\n",
+    ")\n",
     "sampler.run_mcmc(pos, 5000, progress=True);"
    ]
   },

docs/tutorials/monitor.ipynb

@@ -30,6 +30,7 @@
     "%config InlineBackend.figure_format = \"retina\"\n",
     "\n",
     "from matplotlib import rcParams\n",
+    "\n",
     "rcParams[\"savefig.dpi\"] = 100\n",
     "rcParams[\"figure.dpi\"] = 100\n",
     "rcParams[\"font.size\"] = 20"
@@ -305,7 +306,9 @@
    "source": [
     "new_backend = emcee.backends.HDFBackend(filename)\n",
     "print(\"Initial size: {0}\".format(new_backend.iteration))\n",
-    "new_sampler = emcee.EnsembleSampler(nwalkers, ndim, log_prob, backend=new_backend)\n",
+    "new_sampler = emcee.EnsembleSampler(\n",
+    "    nwalkers, ndim, log_prob, backend=new_backend\n",
+    ")\n",
     "new_sampler.run_mcmc(None, 100)\n",
     "print(\"Final size: {0}\".format(new_backend.iteration))"
    ]
@@ -345,7 +348,9 @@
     "# Rinse, Wash, and Repeat as above\n",
     "coords = np.random.randn(32, 5)\n",
     "nwalkers, ndim = coords.shape\n",
-    "sampler2 = emcee.EnsembleSampler(nwalkers, ndim, log_prob2, backend=run2_backend)\n",
+    "sampler2 = emcee.EnsembleSampler(\n",
+    "    nwalkers, ndim, log_prob2, backend=run2_backend\n",
+    ")\n",
     "\n",
     "# note: this is *not* necessarily the right number of iterations for this\n",
     "# new prior.  But it will suffice  to demonstrate the second backend.\n",

docs/tutorials/moves.ipynb

@@ -25,6 +25,7 @@
     "%config InlineBackend.figure_format = \"retina\"\n",
     "\n",
     "from matplotlib import rcParams\n",
+    "\n",
     "rcParams[\"savefig.dpi\"] = 100\n",
     "rcParams[\"figure.dpi\"] = 100\n",
     "rcParams[\"font.size\"] = 20"
@@ -107,7 +108,11 @@
     "sampler0 = emcee.EnsembleSampler(nwalkers, ndim, logprob)\n",
     "sampler0.run_mcmc(init, 5000)\n",
     "\n",
-    "print(\"Autocorrelation time: {0:.2f} steps\".format(sampler0.get_autocorr_time()[0]))"
+    "print(\n",
+    "    \"Autocorrelation time: {0:.2f} steps\".format(\n",
+    "        sampler0.get_autocorr_time()[0]\n",
+    "    )\n",
+    ")"
    ]
   },
   {
@@ -201,7 +206,11 @@
     ")\n",
     "sampler.run_mcmc(init, 5000)\n",
     "\n",
-    "print(\"Autocorrelation time: {0:.2f} steps\".format(sampler.get_autocorr_time()[0]))\n",
+    "print(\n",
+    "    \"Autocorrelation time: {0:.2f} steps\".format(\n",
+    "        sampler.get_autocorr_time()[0]\n",
+    "    )\n",
+    ")\n",
     "\n",
     "plt.plot(sampler.get_chain()[:, 0, 0], \"k\", lw=0.5)\n",
     "plt.xlim(0, 5000)\n",

docs/tutorials/parallel.ipynb

@@ -22,6 +22,7 @@
     "%config InlineBackend.figure_format = \"retina\"\n",
     "\n",
     "from matplotlib import rcParams\n",
+    "\n",
     "rcParams[\"savefig.dpi\"] = 100\n",
     "rcParams[\"figure.dpi\"] = 100\n",
     "rcParams[\"font.size\"] = 20\n",
@@ -407,7 +408,9 @@
     "    multi_data_time = end - start\n",
     "    print(\"Multiprocessing took {0:.1f} seconds\".format(multi_data_time))\n",
     "    print(\n",
-    "        \"{0:.1f} times faster(?) than serial\".format(serial_data_time / multi_data_time)\n",
+    "        \"{0:.1f} times faster(?) than serial\".format(\n",
+    "            serial_data_time / multi_data_time\n",
+    "        )\n",
     "    )"
    ]
   },
@@ -460,12 +463,16 @@
     "\n",
     "\n",
     "with Pool() as pool:\n",
-    "    sampler = emcee.EnsembleSampler(nwalkers, ndim, log_prob_data_global, pool=pool)\n",
+    "    sampler = emcee.EnsembleSampler(\n",
+    "        nwalkers, ndim, log_prob_data_global, pool=pool\n",
+    "    )\n",
     "    start = time.time()\n",
     "    sampler.run_mcmc(initial, nsteps, progress=True)\n",
     "    end = time.time()\n",
     "    multi_data_global_time = end - start\n",
-    "    print(\"Multiprocessing took {0:.1f} seconds\".format(multi_data_global_time))\n",
+    "    print(\n",
+    "        \"Multiprocessing took {0:.1f} seconds\".format(multi_data_global_time)\n",
+    "    )\n",
     "    print(\n",
     "        \"{0:.1f} times faster than serial\".format(\n",
     "            serial_data_time / multi_data_global_time\n",

docs/tutorials/quickstart.ipynb

@@ -22,6 +22,7 @@
     "%config InlineBackend.figure_format = \"retina\"\n",
     "\n",
     "from matplotlib import rcParams\n",
+    "\n",
     "rcParams[\"savefig.dpi\"] = 100\n",
     "rcParams[\"figure.dpi\"] = 100\n",
     "rcParams[\"font.size\"] = 20"
@@ -300,7 +301,11 @@
     }
    ],
    "source": [
-    "print(\"Mean acceptance fraction: {0:.3f}\".format(np.mean(sampler.acceptance_fraction)))"
+    "print(\n",
+    "    \"Mean acceptance fraction: {0:.3f}\".format(\n",
+    "        np.mean(sampler.acceptance_fraction)\n",
+    "    )\n",
+    ")"
    ]
   },
   {

docs/user/blobs.rst

@@ -69,10 +69,10 @@ To do this, we would update the above example as follows:
     def log_prob(params):
         lp = log_prior(params)
         if not np.isfinite(lp):
-            return -np.inf, -np.inf
+            return -np.inf, -np.inf, -np.inf
         ll = log_like(params)
         if not np.isfinite(ll):
-            return lp, -np.inf
+            return lp, -np.inf, -np.inf
         return lp + ll, lp, np.mean(params)
 
     coords = np.random.randn(32, 3)