Developer Documentation

Overall Design

DevAssistant consists of several parts:

Core
Core of DevAssistant is written in Python. It is responsible for interpreting Yaml Assistants and it provides an API that can be used by any consumer for the interpretation.
CL Interface
CL interface allows users to interact with DevAssistant on commandline; it consumes the Core API.
GUI
(work in progress) GUI allows users to interact with Developer Assistant from GTK based GUI; it consumes the Core API.
Assistants
Assistants are Yaml files with special syntax and semantics (defined in Yaml Assistant Reference). They are indepent of the Core, therefore any software distribution can carry its own assistants and drop them into the directory from where DevAssistant loads them - they will be loaded on next invocation. Note, that there is also a possibility to write assistants in Python, but this is no longer supported and will be removed in near future.

Assistants

Internally, each assistant is represented by instance of devassistant.yaml_assistant.YamlAssistant. Instances are constructed by DevAssistant in runtime from parsed yaml files. Each assistant can have zero or more subassistants. This effectively forms a tree-like structure. For example:

       MainAssistant
       /           \
      /             \
    Python          Ruby
    /   \            / \
   /     \          /   \
Django  Flask    Rails Sinatra

This structure is defined by filesystem hierarchy as explained in Assistants Loading Mechanism

Each assistant can optionally define arguments that it accepts (either on commandline, or from GUI). For example, you can run the leftmost path with:

$ da crt python [python assistant arguments] django [django assistant arguments]

If an assistant has any subassistants, one of them must be used. E.g. in the example above, you can’t use just Python assistant, you have to choose between Django and Flask. If Django would get a subassistant, it wouldn’t be usable on its own any more, etc.

Assistant Roles

The crt in the above example means, that we’re running an assistant that creates a project.

There are four assistant roles:

creator (crt in short)
creates new projects
modifier (mod in short)
modifies existing projects
preparer (prep in short)
prepares environment for development of upstream projects
task (task in short)
performs arbitrary tasks not related to a specific project

The main purpose of having roles is separating different types of tasks. It would be confusing to have e.g. python django assistant (that creates new project) side-by-side with eclipse assistant (that registers existing project into Eclipse).

Contributing

If you want to contribute (bug reporting, new assistants, patches for core, improving documentation, ...), please use our Github repo:

If you have DevAssistant installed (version 0.8.0 or newer), there is a fair chance that you have devassistant preparer. Just run da prep devassistant and it will checkout our sources and do all the boring stuff that you’d have to do without DevAssistant.

If you don’t have DevAssistant installed, you can checkout the sources like this (just copy&paste this to get the job done):

git clone https://github.com/bkabrda/devassistant
# get the official set of assistants
cd devassistant
git submodule init
git submodule update

You can find list of core Python dependencies in file requirements.txt. If you want to write and run tests (you should!), install dependencies from requirements-devel.txt:

pip install -r requirements-devel.txt

On top of that, you’ll need polkit for requesting root privileges for dependency installation etc. If you want to play around with GUI, you have to install pygobject, too (see how hard this is compared to da prep devassistant?)

Read the Docs v: v0.8.0
Versions
latest
v0.8.0
v0.7.0
v0.6.1
v0.5.0
Downloads
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.