From faf9dc90cec76a1f78c9b0504aa1f0605288b016 Mon Sep 17 00:00:00 2001
From: bstabler <benstabler@yahoo.com>
Date: Tue, 9 Nov 2021 13:47:21 -0800
Subject: [PATCH] update getting started notebook for conda

---
 .../notebooks/getting_started.ipynb           | 30 +++++++++++++++----
 1 file changed, 25 insertions(+), 5 deletions(-)

diff --git a/activitysim/examples/example_mtc/notebooks/getting_started.ipynb b/activitysim/examples/example_mtc/notebooks/getting_started.ipynb
index 83940aa185..2a03e452ab 100644
--- a/activitysim/examples/example_mtc/notebooks/getting_started.ipynb
+++ b/activitysim/examples/example_mtc/notebooks/getting_started.ipynb
@@ -30,7 +30,7 @@
    "source": [
     "# Install ActivitySim\n",
     "\n",
-    "The first step is to install activitysim from [pypi](https://pypi.org/project/activitysim/) (the Python package index).  It also installs dependent packages such as [tables](https://pypi.org/project/tables/) for reading/writing HDF5, [openmatrix](https://pypi.org/project/OpenMatrix/) for reading/writing OMX matrix, and [pyyaml](https://pypi.org/project/PyYAML/) for yaml settings files."
+    "The first step is to install activitysim from [conda forge](https://anaconda.org/conda-forge/activitysim).  This also installs dependent packages such as [tables](https://pypi.org/project/tables/) for reading/writing HDF5, [openmatrix](https://pypi.org/project/OpenMatrix/) for reading/writing OMX matrix, and [pyyaml](https://pypi.org/project/PyYAML/) for yaml settings files.  The command below also creates an asim [conda environment](https://conda.io/projects/conda/en/latest/user-guide/concepts/environments.html) just for activitysim."
    ]
   },
   {
@@ -47,7 +47,26 @@
    },
    "outputs": [],
    "source": [
-    "!pip install activitysim"
+    "!conda create -n asim python=3.9 activitysim -c conda-forge --override-channels"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "colab_type": "text",
+    "id": "5Tid_70tVBlc"
+   },
+   "source": [
+    "# Activate the Environment"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!conda activate asim"
    ]
   },
   {
@@ -9355,6 +9374,7 @@
     "```\n",
     "num_processes: 2\n",
     "chunk_size: 0\n",
+    "chunk_training_mode: disabled\n",
     "\n",
     "multiprocess_steps:\n",
     "  - name: mp_initialize\n",
@@ -9370,7 +9390,7 @@
     "\n",
     "```\n",
     "\n",
-    "In brief, `num_processes` specifies the number of processors to use and a `chunk_size` of `0` means ActivitySim is free to use all the available RAM if needed.  The `multiprocess_steps` specifies the beginning, middle, and end steps in multiprocessing.  The `mp_initialize` step is single processed because there is no `slice` setting.  It starts with the `initialize_landuse` submodel and runs until the submodel identified by the next multiprocess submodel starting point, `school_location`.  The `mp_households` step is multiprocessed and the households and persons tables are sliced and allocated to processes using the chunking settings.  The rest of the submodels are run multiprocessed until the final multiprocess step.  The `mp_summarize` step is single processed because there is no `slice` setting and it writes outputs.  See [multiprocessing](https://activitysim.github.io/activitysim/core.html#multiprocessing) and [chunk_size](https://activitysim.github.io/activitysim/core.html#chunk) for more information.  "
+    "In brief, `num_processes` specifies the number of processors to use and a `chunk_size` of `0` plus a `chunk_training_mode` of `disabled` means ActivitySim is free to use all the available RAM if needed.  The `multiprocess_steps` specifies the beginning, middle, and end steps in multiprocessing.  The `mp_initialize` step is single processed because there is no `slice` setting.  It starts with the `initialize_landuse` submodel and runs until the submodel identified by the next multiprocess submodel starting point, `school_location`.  The `mp_households` step is multiprocessed and the households and persons tables are sliced and allocated to processes using the chunking settings.  The rest of the submodels are run multiprocessed until the final multiprocess step.  The `mp_summarize` step is single processed because there is no `slice` setting and it writes outputs.  See [multiprocessing](https://activitysim.github.io/activitysim/core.html#multiprocessing) and [chunk_size](https://activitysim.github.io/activitysim/core.html#chunk) for more information.  "
    ]
   },
   {
@@ -9629,7 +9649,7 @@
    "toc_visible": true
   },
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "Python 3 (ipykernel)",
    "language": "python",
    "name": "python3"
   },
@@ -9643,7 +9663,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.10"
+   "version": "3.9.7"
   }
  },
  "nbformat": 4,