Build documentation in Travis (NeuralEnsemble#335)

* fixed _check_consistency in BinnedSpikeTrain and parallel.ipynb notebook

* added docs build in travis; fixed plt.eventplot(spiketrains)

* don't use conda-forge channel for python 2.7

* install extra requirements in docs build

* conda install pandoc

* travis hangs

* travis doc mpi error

* travis check if error is occurred

* Revert "travis check if error is occurred"

This reverts commit 3e59894.

* reverted travis

.travis.yml

@@ -12,6 +12,7 @@ matrix:
     - name: "conda 2.7"
       python: 2.7
       env: DISTRIB="conda"
+      before_install: sed -i 's/conda-forge/conda/g' requirements/environment.yml
 
     - name: "pip 2.7"
       python: 2.7
@@ -41,6 +42,18 @@ matrix:
       python: 3.8
       env: DISTRIB="pip"
 
+    - name: "docs"
+      python: 3.6
+      env: DISTRIB="conda"
+      before_install: sudo apt install -y libopenmpi-dev openmpi-bin
+      before_script:
+        - conda install -c conda-forge pandoc
+        - pip install -r requirements/requirements-docs.txt
+        - pip install -r requirements/requirements-tutorials.txt
+        - pip install -r requirements/requirements-extras.txt
+        - sed -i -E "s/nbsphinx_execute *=.*/nbsphinx_execute = 'always'/g" doc/conf.py
+      script: cd doc && make html
+
 install:
   - if [[ "${DISTRIB}" == "conda" ]];
     then

doc/tutorials/asset.ipynb

@@ -134,7 +134,7 @@
    ],
    "source": [
     "plt.figure()\n",
-    "plt.eventplot(spiketrains, linewidths=5, linelengths=5)\n",
+    "plt.eventplot([st.magnitude for st in spiketrains], linewidths=5, linelengths=5)\n",
     "plt.xlabel('time [ms]')\n",
     "plt.ylabel('neuron id')\n",
     "plt.title('Raw spikes')\n",
@@ -800,7 +800,7 @@
    ],
    "source": [
     "plt.figure()\n",
-    "plt.eventplot(reordered_sts, linewidths=5, linelengths=5)\n",
+    "plt.eventplot([st.magnitude for st in reordered_sts], linewidths=5, linelengths=5)\n",
     "plt.xlabel('time [ms]')\n",
     "plt.ylabel('reordered neuron id')\n",
     "plt.title('Reconstructed ordering of the neurons (y-axis) with synfire chains');"
@@ -839,7 +839,7 @@
     "spiketrains_ordered = [spiketrains[idx] for idx in ordering_true]\n",
     "\n",
     "plt.figure()\n",
-    "plt.eventplot(spiketrains_ordered, linewidths=5, linelengths=5)\n",
+    "plt.eventplot([st.magnitude for st in spiketrains_ordered], linewidths=5, linelengths=5)\n",
     "plt.xlabel('time [ms]')\n",
     "plt.ylabel('neuron id')\n",
     "plt.title('True (unknown) ordering of the neurons (y-axis)')\n",

doc/tutorials/gpfa.ipynb

@@ -940,6 +940,9 @@
   }
  ],
  "metadata": {
+   "nbsphinx": {
+    "execute": "never"
+  },
   "anaconda-cloud": {},
   "kernelspec": {
    "display_name": "Python 3",

doc/tutorials/parallel.ipynb

@@ -10,6 +10,7 @@
     "\n",
     "`elephant.parallel` module provides a simple interface to parallelize multiple calls to any user-specified function. The typical use case is calling a function many times with different parameters.\n",
     "\n",
+    "\n",
     "## Available executors\n",
     "\n",
     "`elephant.parallel` has 3 interfaces to choose from, depending whether the user has a laptop/PC or the computation is being done on a cluster machine with many nodes and MPI installed.\n",
@@ -22,6 +23,7 @@
     "\n",
     "All listed above classes have the same API and can be used interchangeably.\n",
     "\n",
+    "\n",
     "## How to use\n",
     "\n",
     "Let's say you want to call some function `my_function()` for each element in a list `iterables_list` like so:\n",
@@ -32,7 +34,11 @@
     "\n",
     "(eq. 2) `results = Executor().execute(my_function, iterables_list)`,\n",
     "\n",
-    "where `Executor` can be any of the available executors listed above. For more information about parallel executors in Python refer to https://docs.python.org/3/library/concurrent.futures.html."
+    "where `Executor` can be any of the available executors listed above. For more information about parallel executors in Python refer to https://docs.python.org/3/library/concurrent.futures.html.\n",
+    "\n",
+    "**Note**. To successfully run this notebook, `mpi4py` package should be installed in either of ways:\n",
+    "* `conda install -c conda-forge mpi4py`\n",
+    "* `pip install mpi4py`, if you have manually installed OpenMPI with `sudo apt install libopenmpi-dev openmpi-bin`\n"
    ]
   },
   {
@@ -235,7 +241,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "This is because both MPI executors require `-m mpi4py.futures` flag while running python scripts:\n",
+    "This is because MPI executors require `-m mpi4py.futures` flag while running python scripts:\n",
     "\n",
     "```\n",
     "mpiexec -n numprocs python -m mpi4py.futures pyfile [arg] ...\n",
@@ -281,6 +287,49 @@
     "time_hist = ProcessPoolExecutor().execute(my_custom_function, binsize_list)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[AnalogSignal with 1 channels of length 100; units dimensionless; datatype float64 \n",
+       " sampling rate: 10.0 1/s\n",
+       " time: 0.0 s to 10.0 s,\n",
+       " AnalogSignal with 1 channels of length 43; units dimensionless; datatype float64 \n",
+       " sampling rate: 4.375 1/s\n",
+       " time: 0.0 s to 9.82857142857143 s,\n",
+       " AnalogSignal with 1 channels of length 28; units dimensionless; datatype float64 \n",
+       " sampling rate: 2.7999999999999994 1/s\n",
+       " time: 0.0 s to 10.000000000000002 s,\n",
+       " AnalogSignal with 1 channels of length 20; units dimensionless; datatype float64 \n",
+       " sampling rate: 2.0588235294117645 1/s\n",
+       " time: 0.0 s to 9.714285714285715 s,\n",
+       " AnalogSignal with 1 channels of length 16; units dimensionless; datatype float64 \n",
+       " sampling rate: 1.627906976744186 1/s\n",
+       " time: 0.0 s to 9.82857142857143 s,\n",
+       " AnalogSignal with 1 channels of length 13; units dimensionless; datatype float64 \n",
+       " sampling rate: 1.346153846153846 1/s\n",
+       " time: 0.0 s to 9.657142857142858 s,\n",
+       " AnalogSignal with 1 channels of length 11; units dimensionless; datatype float64 \n",
+       " sampling rate: 1.1475409836065573 1/s\n",
+       " time: 0.0 s to 9.585714285714285 s,\n",
+       " AnalogSignal with 1 channels of length 10; units dimensionless; datatype float64 \n",
+       " sampling rate: 1.0 1/s\n",
+       " time: 0.0 s to 10.0 s]"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "time_hist"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -294,7 +343,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 11,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -307,14 +356,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 12,
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "722 ms ± 43.9 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
+      "873 ms ± 35.9 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
      ]
     }
    ],
@@ -325,7 +374,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": 13,
    "metadata": {},
    "outputs": [
     {
@@ -358,7 +407,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.7"
+   "version": "3.8.3"
   }
  },
  "nbformat": 4,

doc/tutorials/statistics.ipynb

@@ -154,7 +154,7 @@
    "outputs": [],
    "source": [
     "plt.figure(figsize=(8, 3))\n",
-    "plt.eventplot([spiketrain1, spiketrain2], linelengths=0.75, color='black')\n",
+    "plt.eventplot([spiketrain1.magnitude, spiketrain2.magnitude], linelengths=0.75, color='black')\n",
     "plt.xlabel('Time (ms)', fontsize=16)\n",
     "plt.yticks([0,1], labels=[\"spiketrain1\", \"spiketrain2\"], fontsize=16)\n",
     "plt.title(\"Figure 1\");"
@@ -475,7 +475,7 @@
    "outputs": [],
    "source": [
     "plt.figure(dpi=150)\n",
-    "plt.eventplot(spiketrain_list, linelengths=0.75, linewidths=0.75, color='black')\n",
+    "plt.eventplot([st.magnitude for st in spiketrain_list], linelengths=0.75, linewidths=0.75, color='black')\n",
     "plt.xlabel(\"Time, s\")\n",
     "plt.ylabel(\"Neuron id\")\n",
     "plt.xlim([0, 1]);"

elephant/conversion.py

@@ -633,12 +633,17 @@ def _check_consistency(self, spiketrains, binsize, num_bins, t_start,
                 raise ValueError(
                     'too many / too large time bins. Some spike trains are '
                     'not defined in the ending time')
-        if num_bins != int((
+        # account for rounding errors in the reference num_bins
+        num_bins_test = ((
             (t_stop - t_start).rescale(
-                binsize.units) / binsize).magnitude):
+                binsize.units) / binsize).magnitude)
+        if _detect_rounding_errors(num_bins_test, tolerance=self.tolerance):
+            num_bins_test += 1
+        num_bins_test = int(num_bins_test)
+        if num_bins != num_bins_test:
             raise ValueError(
                 "Inconsistent arguments t_start (%s), " % t_start +
-                "t_stop (%s), binsize (%d) " % (t_stop, binsize) +
+                "t_stop (%s), binsize (%s) " % (t_stop, binsize) +
                 "and num_bins (%d)" % num_bins)
         if num_bins - int(num_bins) != 0 or num_bins < 0:
             raise TypeError(

postBuild

@@ -1,5 +1,4 @@
-pip install .[tutorials]
-
 # Install conda version of mpi4py for mybinder service
-conda install mpi4py
+conda install -c conda-forge mpi4py
 
+pip install .[tutorials]

requirements/environment.yml

@@ -1,5 +1,8 @@
 name: elephant
 
+channels:
+  - conda-forge  # required by mpi4py
+
 dependencies:
   - python>=3.6
   - pip
@@ -10,3 +13,5 @@ dependencies:
   - pandas
   - scikit-learn
   - statsmodels
+  - pip:
+     - -r file:requirements.txt

requirements/requirements-tutorials.txt

@@ -1,3 +1,4 @@
 # Packages required to execute jupyter notebook tutorials
 matplotlib>=3.1.0
-h5py>=2.9.0
+h5py>=2.9.0
+nixio==1.5.0b3