Installation instruction with ansible

Author: ozagord

.gitignore

@@ -2,3 +2,5 @@ __pycache__/
 *pyc
 .dropbox.attr
 ansible/setup.retry
+my.key
+my.key.pub

ansible/hosts

@@ -1,2 +1,2 @@
-[tryminvar]
-172.23.10.154
+[minvarmachine]
+172.23.10.194

ansible/setup-host.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+ansible-playbook setup.yml -i hosts --key-file=path_to_my_private_key -v

ansible/setup-hosts.sh

@@ -1 +1 @@
-172.23.10.154
+172.23.10.194

ansible/setup.retry

@@ -1,22 +1,18 @@
 ---
-########################
-# Test if MinVar works #
-########################
+#######################################
+# This provides everything for MinVar #
+#######################################
 
-- hosts: tryminvar
+- hosts: minvarmachine
   remote_user: ubuntu
   become: yes
 
   tasks:
 
-    - hostname:
-        name: tryminvar
-
     - name: Ensure sysadmin SSH keys are authorized for user 'ubuntu'
       authorized_key: user="ubuntu" key="{{ item }}"
       with_file:
-        - ~/.ssh/oct2016.key.pub
-
+        - my.key.pub
 
     # - name: Upgrade all installed packages to latest version
     #   apt:
@@ -30,6 +26,7 @@
       with_items:
         - acl
         - automake
+        - bedtools
         - build-essential
         - bwa
         - cmake
@@ -47,7 +44,6 @@
         - pkg-config
         - python-dev
         - python-setuptools
-        - r-cran-ggplot2
         - seqtk
         - tabix
         - unzip
@@ -156,6 +152,18 @@
         owner: ubuntu
         mode: 0775
 
+    - name: Download pandoc packages
+      get_url:
+          url: https://github.com/jgm/pandoc/releases/download/1.19.2.1/pandoc-1.19.2.1-1-amd64.deb
+          dest: /tmp/pandoc-1.19.2.1-1-amd64.deb
+      register: get_pandoc
+
+    - name: Install pandoc
+      shell: |
+        if [ ! -x "$(command -v pandoc)" ]; then
+           dpkg -i /tmp/pandoc-1.19.2.1-1-amd64.deb
+        fi
+
     - name: Download miniconda
       get_url:
         url: https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
@@ -171,7 +179,7 @@
             hash -r && \
                 /opt/miniconda/bin/conda config --set always_yes yes --set changeps1 no && \
                 /opt/miniconda/bin/conda update -q conda
-            /opt/miniconda/bin/conda install -q Biopython pandas seaborn scipy
+            /opt/miniconda/bin/conda install -q Biopython pandas
         fi
       args:
           executable: /bin/bash
@@ -194,3 +202,10 @@
         fi
       args:
           executable: /bin/bash
+
+    - name: Add directories to PATH
+      lineinfile: dest=/home/{{item.user}}/.bash_profile
+                  line="export PATH=$PATH:/usr/local/edirect:/usr/local/vcflib:/opt/miniconda/bin"
+                  create=yes
+      with_items:
+        - { user: "ubuntu" }

ansible/setup.yml

@@ -343,7 +343,7 @@ def phase_mutations(muts, frame, bam_file):
                         pass
                 d_here = {'wt': 'XYZ', 'pos': t[0], 'mut': k, 'freq': freq_here}
                 pm = pm.append(d_here, ignore_index=True)
-                logging.info('%s %f %d', % (k, freq_here, t[0]))
+                logging.info('%s %f %d' % (k, freq_here, t[0]))
         logging.info('<------>')
     muts = muts.reset_index()
     pm = pm.reset_index()

minvar/annotate.py

@@ -1,4 +1,4 @@
-## Installation
+## Installation (work in progress)
 
 A condensed list of dependencies
 
@@ -11,6 +11,12 @@ A condensed list of dependencies
 - bwa
 - lofreq
 
+### Install GATK
+
+GATK is not (yet?) included in the instructions below because you need
+to register online [here](https://software.broadinstitute.org/gatk/download/)
+in order to download it. Then, install it in `/usr/local/GATK`. MinVar expects
+to find the Java archive file `/usr/local/GATK/GenomeAnalysisTK.jar`.
 
 ### Setting up with Ansible
 
@@ -23,11 +29,39 @@ its dependencies on a dedicated machine. The files therein define an
 
 Ansible is a deployment tool that allows an automatic provisioning of machines
 on the cloud (it can be used on AWS, DigitalOcean, Google Cloud Platform etc.)
-The user install ansible on their own local machine (this can be your old laptop),
+The user installs ansible on a local machine (this can be your old laptop),
 defines ansible commands in specific files and uses them to set up a remote
 machine.
 
 In the following we will assume that you have installed ansible on your local
-laptop and you want to set up MinVar on a remote machine running Ubuntu Linux.
+laptop and you want to set up MinVar on a remote machine running Linux Ubuntu 16.04.
+The access to this machine is provided by private-public SSH key pair that must be
+set up. Good instructions for this task can be found in this [help](https://help.ubuntu.com/community/SSH/OpenSSH/Keys).
+
+#### How to proceed
+
+1. Install ansible on your machine (on Mac OS X you can do it via the package
+   manager [brew](https://brew.sh) with `brew install ansible`),
+2. clone MinVar from GitHub with `git clone https://github.com/ozagordi/MinVar.git`
+   or download/unzip it,
+3. move to the directory `ansible` in the cloned repository and identify the file
+   `hosts`. This file contains two lines `[minvarmachine]` and a fake ip address.
+   Adapt to the name of the machine you want to setup and its address,
+4. in the same directory edit the file `setup-hosts.sh`: adapt
+   `--key-file=path_to_your_private_key` to point to your personal *private* key,
+5. copy your personal *public* key into `my.key`,
+6. run `./setup-hosts.sh`.
+
+#### What can go wrong
+
+We assumed that you have an Ubuntu 16.04 available. Most of the stuff will work
+on Ubuntu 14.04, but you need to switch to a newer version of Java provided by
+Oracle in order to run `picard`. You can find
+[here](https://www.digitalocean.com/community/tutorials/how-to-install-java-with-apt-get-on-ubuntu-16-04)
+a good tutorial on this.
+
+### If you don't want to use ansible
 
-https://www.digitalocean.com/community/tutorials/how-to-install-java-with-apt-get-on-ubuntu-16-04
+The ansible [playbook](https://github.com/ozagordi/MinVar/blob/master/ansible/setup.yml)
+reads almost as plain English. You can manually copy the instructions from there
+and install what you need.