philandstuff/standalone-mapit.org

## standalone-mapit.org

      
    Raw
  

              standalone-mapit.org
            
          
    mapit

standing up a mapit instance

Setting up a standalone mapit instance is documented at
  http://mapit.poplus.org/docs/self-hosted/install/ but there are some
  issues.  This summarizes what I needed to do:
# install postgis
brew install postgis

# set up db
createuser mapit
createdb mapit
psql -d mapit -c 'CREATE EXTENSION postgis;'

# clone the mapit repo
git clone git://github.com/mysociety/mapit.git
cd mapit

# set up standalone django project
cp conf/general.yml-example conf/general.yml
# edit according to documentation above.  in particular, you want these settings:
# MAPIT_DB_HOST: localhost
# MAPIT_DB_PORT: 5432
# AREA_SRID: 4326
# -- leave everything else alone

# install deps into a virtualenv:
virtualenv .venv
source .venv/bin/activate
pip install -e .

# compile sass (may require installing the rubygem)
bin/mapit_make_css

# initialize db for django app
./manage.py migrate

# start the django server:
./manage.py runserver
parking lot: documentation issues to report upstream


  ./manage.py migrate mapit doesn’t work if you haven’t already
    run ./manage.py migrate

importing postcodes from the postcode register

The docs for importing data are quite good.  To import data from the
  postcode register, you only need two commands:
curl -O http://postcode.openregister.org/all.csv
./manage.py mapit_import_postal_codes --coord-field-lat 3 --code-field 2 --strip --header-row all.csv
Explanation:

  --coord-field-lat says which column the latitude will be in.
    (You can specify a seprate column number for longitude, but it
    will assume it’s the next column after latitude if you don’t)
  --code-field says which column the postcode will be in.
  --strip says that the postcode field contains spaces that
    mapit will want to strip (ie “SW1A 0AA” -> “SW1A0AA”)
  --header-row says that the csv has a header row which needs to
    be skipped

Caveat: the all.csv only returns 1000 records at the moment.
To check it worked, visit http://localhost:8000/ and search for “AB15
  9NT” which should be the first postcode in all.csv.  It should show
  you that postcode’s location (north deeside road) on openstreetmap.
To investigate: streaming updates after initial import.
importing boundaries

We don’t have an equivalent register for this at the moment.