From da0077f31bc015a6cb679c7f7bbc22f6d4030bfd Mon Sep 17 00:00:00 2001
From: Jason Vander Heiden <jason.vanderheiden@yale.edu>
Date: Fri, 20 Dec 2024 11:09:22 -0800
Subject: [PATCH] Add conda environment to installation instructions

---
 docs/_static/environment.yaml | 13 +++++++++++++
 docs/install.rst              | 20 ++++++++++++++++++++
 2 files changed, 33 insertions(+)
 create mode 100644 docs/_static/environment.yaml

diff --git a/docs/_static/environment.yaml b/docs/_static/environment.yaml
new file mode 100644
index 0000000..5ae05b5
--- /dev/null
+++ b/docs/_static/environment.yaml
@@ -0,0 +1,13 @@
+name: scimilarity
+channels:
+  - nvidia
+  - pytorch
+  - bioconda
+  - conda-forge
+dependencies:
+  - python=3.10
+  - ipykernel
+  - ipython
+  - ipywidgets
+  - leidenalg
+  - pip
\ No newline at end of file
diff --git a/docs/install.rst b/docs/install.rst
index 90a1371..87bfcfe 100644
--- a/docs/install.rst
+++ b/docs/install.rst
@@ -29,6 +29,13 @@ and installed as follows:
     high memory requirements. To make queries, you will need at least 64 GB of
     system RAM.
 
+.. warning::
+
+    If your environment has sufficient memory but loading the model or making
+    kNN queries crashes, that may be due to older versions of dependencies such
+    as hnswlib or numpy. We recommend using either using the Docker container
+    or Conda environment described below.
+
 .. note::
 
     A GPU is not necessary for most applications, but model training will
@@ -41,6 +48,19 @@ You can download the following pretrained models for use with SCimilarity from
 Zenodo:
 https://zenodo.org/records/10685499
 
+Conda environment setup
+--------------------------------------------------------------------------------
+
+To install the SCimilarity API in a [Conda](https://docs.conda.io) environment
+we recommend this environment setup:
+
+:download:`Download environment file <_static/environment.yaml>`
+
+.. literalinclude:: _static/environment.yaml
+  :language: YAML
+
+Followed by installing the ``scimilarity`` package via ``pip``, as above.
+
 Using the SCimilarity Docker container
 --------------------------------------------------------------------------------