A series of bash scripts to build Pelias on CentOS 6.7.

A full overview of Pelias installation docs can be found at the official Pelias GitHub page here.

$ git clone https://github.com/ktjaco/pelias-bash-centos
$ cd pelias-bash-centos

1. Install Elasticsearch

This bash script installs Elasticsearch. The elasticsearch.sh script also configures Elasticsearch settings for best optimization. Edit the ES_HEAP_SIZE in the elasticsearch.sh script to 28 for full planet imports. Note that ES_HEAP_SIZE should not be more than 50% of your available RAM or more than 32GB.

$ sudo nano scripts/elasticsearch.sh

    # Should look like this
    sudo sed -i 's/export ES_HEAP_SIZE/export ES_HEAP_SIZE=28g/g' /etc/init.d/elasticsearch

$ sudo chmod +x scripts/elasticsearch.sh
$ ./scripts/elasticsearch.sh

2. Install Node

This bash script installs Node v0.12 and Oracle-Java 7 needed for Pelias.

$ sudo chmod +x scripts/node.sh
$ ./scripts/node.sh

3. JSON Configuration

Before running import processes, edit the pelias.json file to point to the appropriate file names and data paths of your desired datasets. The pelias.json assumes that Pelias data and code are located in the /home/user/pelias directory. Elasticsearch hosts will have to be changed if it is not on localhost.

$ sudo nano config/pelias.json

Elasticsearch:

{
  "esclient": {
  "hosts": [{
    "host": "localhost",
    "port": 9200
  }]
}

Imports:

{
"imports": {
    "geonames": {
      "datapath": "/home/user/pelias/geonames/data",
      "adminLookup": false
    },
    "openstreetmap": {
      "datapath": "/home/user/pelias/openstreetmap/data",
      "adminLookup": false,
      "leveldbpath": "/tmp/leveldb/",
      "import": [{
        "filename": "planet-latest.osm.pbf"
      }]
    },
    "openaddresses": {
      "datapath": "/home/user/pelias/openaddresses/data",
      "files": []
    },
    "whosonfirst": {
      "datapath": "/home/user/pelias/whosonfirst/wof_data"
    }
  }
}

4. Import

Before running the import, edit the bash script variables with your desired datasets. This example is using Toronto OSM and OA data and Canada GeoNames data.

OSM=planet-latest.osm.pbf

OA=openaddr-collected-global.zip

GN=allCountries.zip
GN_IMPORT=all

Run the build.sh script to begin the import.

$ sudo chmod +x scripts/build.sh
$ ./scripts/build.sh

5. Service

To start the Pelias server at http:localhost:3100/v1 simply run sudo service pelias start.

$ sudo mv service/pelias /etc/init.d/
$ sudo service pelias { start | stop | restart | status }