Skip to content

Latest commit

 

History

History
115 lines (79 loc) · 4.1 KB

INSTALL.md

File metadata and controls

115 lines (79 loc) · 4.1 KB
Raw

System Requirements

In order to install c3bottles, you will need:

  • Flask (python-flask)

  • Flask-SQLAlchemy (python-flask-sqlalchemy)

  • Flask-Login >= 0.3.0 (versions <0.3.0 will not work since a breaking change has been done in the API) (python-flask-login)

  • Flask-WTF (python-flaskext.wtf)

  • CairoSVG (python-cairosvg)

  • PyPDF (python-pypdf)

  • Python-QRCode (python-qrcode)

  • a WSGI-capable webserver (e.g. Apache)

  • some SQL server supported by SQLAlchemy (the author uses PostgreSQL but others should work, too)

  • ImageMagick, GDAL and the gdal2tiles.py script to generate the map tiles (imagemagick, gdal-bin and python-gdal)

  • the Node.js package manager (npm)

  • on Debian, Node.js is fucked up, so the legacy symlink is needed for npm etc. (node-legacy)

Installation

  1. Make sure all the dependencies are installed. On Debian using Apache, you can do:

    $ apt-get install python-flask python-flask-sqlalchemy \
                  python-flask-login python-flaskext.wtf \
                  python-cairosvg python-pypdf python-qrcode \
                  libapache2-mod-wsgi imagemagick \
                  gdal-bin python-gdal npm node-legacy

  2. Copy the files into some directory readable by the web server. You can clone the repository from Github:

    $ git clone https://github.com/der-michik/c3bottles.git

  3. Fetch all the dependencies and build everything:

    $ npm install
$ npm run build

  4. Create a configuriation file config.py. You will find a template for the configuration in the file config.default.py. c3bottles will not work if no config.py with the required settings is present. You have to configure at least a database URI and a secret key.

  5. Configure your database accordingly. The user for c3bottles needs full access to the database. If you use SQLite, the web server needs write access to the directory containing the *.db file and to the file itself.

  6. Initialize the database using the Python interpreter:

    $ cd /path/to/c3bottles
$ python
>>> from controller import db, load_config
>>> load_config()
>>> db.create_all()

  7. Configure the users needed and their passwords in model/user.py as needed. To generate password hashes, you can do the following:

    $ python
>>> from werkzeug.security import generate_password_hash
>>> generate_password_hash("foo")

  8. Configure your webserver accordingly to run the WSGI application. Apache needs something like this to run c3bottles as the document root of a host:

    WSGIScriptAlias / /path/to/c3bottles/wsgi.py
WSGIApplicationGroup %{GLOBAL}
Alias /static /path/to/c3bottles/static

Map

To be able to use the drop point map, you first have to generate the map tiles. The normal build task already does that with the map delivered with the package. If you replaced the map and just want to regenerate the tiles using the defaults, you can do this with:

$ npm run build:map

The standard source file for the map is static/img/map.png. For tile generation, the image has to be a square whose height/width is a power of 2. It is useful to resize and enlarge the image to the next power of 2. This can be done with ImageMagick's convert utility as follows, assuming that the image has a largest dimension between 8192 and 16384 and a white background:

$ cd /path/to/c3bottles/static/img
$ convert map.png -background white -compose Copy -gravity center \
                  -resize 16384x16384 -extent 16384x16384 map_sq.png

An image size of 16384x16384 allows a maximum zoom level of 6 which is the default in the JavaScript code of the map. If your image is larger or smaller, the maxZoom setting in static/js/map.js has to be adapted accordingly (i.e. to 5 for an 8192x8192 image).

Once you have a square source image with a power of 2 in size, you can generate the tiles as follows:

$ cd /path/to/c3bottles/static/img
$ gdal_translate -of vrt map_sq.png map_sq.vrt
$ gdal2tiles.py -w none -p raster map_sq.vrt tiles