update for loop doc and api

brainpy/math/controls.py

@@ -785,6 +785,10 @@ def _body_fun(op):
     if not isinstance(static_vals, (tuple, list)):
       static_vals = (static_vals, )
     new_vals = body_fun(*static_vals)
+    if new_vals is None:
+      new_vals = tuple()
+    if not isinstance(new_vals, tuple):
+      new_vals = (new_vals, )
     return [v.value for v in dyn_vars], new_vals
 
   def _cond_fun(op):

docs/tutorial_building/build_conductance_neurons.ipynb

@@ -36,9 +36,7 @@
    ]
   },
   {
-   "cell_type": "code",
-   "execution_count": null,
-   "outputs": [],
+   "cell_type": "markdown",
    "source": [
     "On the other hand, simplified models do not care about the physiological features of neurons but mainly focus on how to reproduce the exact spike timing. Therefore, they are more simplified and maybe not biologically explicable.\n",
     "\n",
@@ -47,21 +45,19 @@
    "metadata": {
     "collapsed": false,
     "pycharm": {
-     "name": "#%%\n"
+     "name": "#%% md\n"
     }
    }
   },
   {
-   "cell_type": "code",
-   "execution_count": null,
-   "outputs": [],
+   "cell_type": "markdown",
    "source": [
     "## Building an ion channel"
    ],
    "metadata": {
     "collapsed": false,
     "pycharm": {
-     "name": "#%%\n"
+     "name": "#%% md\n"
     }
    }
   },
@@ -436,17 +432,6 @@
    "source": [
     "By combining different ion channels, we can get different types of conductance-based neuron models easily and straightforwardly. To see all predifined channel models in BrainPy, please click [here](../apis/dyn.rst)."
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {