Table Of Contents

tg-admin command line tool reference

The tg-admin command line tool provides commands for working with TurboGears projects. You can get a list of the available commands by running:


The basic syntax is:

tg-admin <command> [options]

For most commands, if you add the -h or --help argument, you’ll find a reference to the options for that command.

You can also give the -h or --help option without any command, which will give you this help message for a standard TurboGears installation:

TurboGears X.Y.Z command line interface

Usage: /path/to/tg-admin [options] <command>

    -c CONFIG --config=CONFIG    Config file to use
    -e EGG_SPEC --egg=EGG_SPEC   Run command on given Egg

      i18n  Manage i18n data
      info  Show version info
quickstart  Create a new TurboGears project
     shell  Start a Python prompt with your database available
       sql  Run the database provider manager
   toolbox  Launch the TurboGears Toolbox
    update  Update an existing turbogears project

info – list installed packages with version info

The tg-admin info command will list all packages that are part of TurboGears or are required by it, which are currently installed and shows their version numbers. Also, all packages that are registered for one of TurboGears’ entry points are shown.

quickstart – create a new project

The tg-admin quickstart [options] [NAME] command will give you a new project skeleton so that you can be up and running without wondering where to put things. The quickstart command needs a couple of options and can either get them from the command line or prompt you for them.

The options available for quickstart are:

-s, --sqlalchemy                     use SQLAlchemy rather than SQLObject
-e, --elixir                         use SQLAlchemy Elixir instead of SQLObject
-o, --sqlobject                      use SQLObject instead of SQLAlchemy
-i, --identity                       enable Identity
-p PACKAGE, --package=PACKAGE        package name for the code
-t TEMPLATES, --templates=TEMPLATES  use non-default template
-r REPOS, --svn-repository=REPOS     create project in given SVN repository
--dry-run                            dry run (don't actually do anything)

In a standard installation, TurboGears provides two additional templates:

  • tgbig is a project with a more complex directory structure for when you’re starting a big project.
  • tgwidget serves as a base for your own custom widget distribution.

More quickstart templates can be created as TurboGears extensions and may be found in the CogBin.

Most people will just let quickstart prompt them for the name and the package.

The name of the project is the friendly name that you want to appear on PyPI (if you’re releasing it), and it’s also the name that will be used for the egg (if you package it up that way).

The package name is the filesystem- and Python-friendly version of the project’s name. Since Python package names are generally lower case, the package name should be lowercase as well. If you don’t specify a package name on the command line, quickstart will guess at a suitable name based on the project’s name.

If you want to create the new TurboGears project in an SVN repository, you can specify its URL with the -r option, and quickstart will automatically create your new TurboGears project there, using the svn add command. You will still have to give a final svn commit command in order to check the project in to your SVN repository.

If you use the --dry-run option, quickstart will tell you about the files and directories it would have created without actually creating them.

shell – manipulate your database with ease

The tg-admin shell command will give you a Python interactive shell, just like when you type python from the command line. But, unlike a default python command, this shell is prepopulated with everything in your file and with your database configuration loaded. This makes it really easy to interactively manipulate your database.

Let’s say you have a class in called Person. Let’s say your group is going for the casual amosphere and you’ve decided that all Fredericks will now be called “Fred”:

freds = Person.select_by(first_name="Frederick")
for fred in freds:
  fred.first_name = "Fred"

SQLObject generally defaults to running in autoCommit mode. This means that the update is automatically saved to the database. If you want to do multiple updates in a single transaction, you can use the hub that is defined in the file:

freds = Person.select_by(first_name="Frederick")
for fred in freds:
    fred.first_name = "Fred"
#hub.rollback()  #rollback does work

The tg-admin shell command does not have any command line parameters.

sql – wrapper for sqlobject-admin

SQLObject includes a utility called sqlobject-admin. sqlobject-admin includes a number of handy features, including a command to create a database and another for recording a checkpoint to keep track of database changes.

Running tg-admin sql is equivalent to sqlobject-admin, but the tg-admin sql command will pass a command line parameter specifying the database to use (as pulled from the configuration file). sqlobject-admin also automatically knows where your model objects are stored, because of metadata put into Project.egg-info/sqlobject.txt when you run quickstart.

The tg-admin sql wrapper makes it possible to create your database just by running tg-admin sql create.

Beyond these default parameters, tg-admin sql just wraps sqlobject-admin. For further information about the commands and options available there, look at the documentation for sqlobject-admin itself.

If you are using SQLAlchemy instead of SQLObject, the command will provide a similar functionality, though less sophisticated. Further information about the subcommands available here can be listed with tg-admin sql help.

Note that you can pass the location of the configuration file to tg-admin. So in order to create the database specified in your productive configuration, you would run tg-admin -c prod.cfg sql create.

i18n – I18N administration tool

The tg-admin i18n command helps you to manage your i18n data. It works similarly to admi18n command from the toolbox but operates via command-line.

The following sub-commands available:

  • add: Creates a message catalog for specified locale
  • collect: Scan source files to gather translatable strings to a .pot file
  • merge: Sync message catalog in different languages with master .pot file
  • compile: Compile message catalog (.po to .mo)
  • clean: Delete backups and compiled files

Collecting strings to translate

Command tg-admin i18n collect runs through all of the project’s source file to find strings marked for translations and write them to a so-called “master .pot file”, which in most cases is locales/messages.pot.

If file exists its content is updated with new data and backup is made, just in case.

Adding new locales

To add another locale to your project run tg-admin i18n add command and supply appropriate locale code, like es or uk_UA. This will create locale’s directory and locale’s message catalog file which contains translations.

Compiling catalog data

Command tg-admin i18n compile compile every locale’s message catalog file into machine efficient .mo file.

Keeping translations up to date

Command tg-admin i18n merge copies newly-added message strings from master .pot file to locales’ message catalogs, keeping existing tranlsations intact.

Note: you’ll need to run tg-admin i18n collect first to grab new strings from the source files. Once you’ve added new translations run tg-admin i18n compile to actually see the changes.

toolbox – through the web toolkit

This starts up the Toolbox. The options available are:

--version                   show program's version number and exit
-h, --help                  show this help message and exit
-n, --no-open               don't open browser automatically
-c HOST, --add-client=HOST  allow the client ip address specified to connect to
                            the toolbox

By default, the toolbox allows only the localhost to connect but more host ip addresses can be added using the -c option, which accepts hostnames in standard CIDR format.

So if your IP address was, you could allow all the machines in 192.168.0.* to connect by passing -c, 10.0.* would be -c, and * would be -c (not recommended).