A component of the Ender CLI, used to build a dependency graph of installed packages given a list of parent packages.
Similar to the npm ls command, this package understands Ender-specific packages by using ender-package-util to interpret the package.json files (which may contain Ender-specific overrides for some root keys).
The DependencyGraph object returned by this package can be used to iterate over packages in the dependency graph. This is used for package-bundling in the Ender CLI.
The ender-js package (or alternative client-lib, supplied with the 'client-lib' option), where it exists in the list of packages, will be automatically shifted to the top of the dependency graph so it is always processed first.
For more information check out http://ender.jit.su
enderDependencyGraph() called as a function, will construct a DependencyGraph object for the given list of packages, working in the current working directory, and return it to via the callback function.
The options object may contain an optional 'client-lib' key that will override the default 'ender-js'. The client-lib will be automatically shifted to the begining of the graph so that it will always be processed first.
Each node in the graph takes the form:
{
"packageJSON": {} // the package.json data, interpreted by ender-package-util
, "parents": [] // an array of parent names/paths, useful for locating the package on disk
, "dependencies": {} // any child-nodes of this node
}Any dependencies in the tree, including root packages, that are not found on disk, will be identified by replacing the node object with the sring: 'missing'.
create() is mainly for internal use but it's also useful for testing with dummy data. It will return a DependencyGraph structure given the graphData generated by scanning the filesystem and the package.json files. The graphData may also be constructed manually for unit testing purposes.
getClientPackageName() is a simple utility to work out the client-lib from the given options. By default it is 'ender-js' but the 'client-lib' property in the options object may override this value.
archyTree() will take a list of packages and a DependencyGraph object and return an archy-compatible tree out of the data. Each node takes the following structure:
{
"label": "" // the name of the package
, "version": x.y.z // the semver for the packake
, "description": "" // the package description from package.json
, "nodes": [] // child nodes of this package
}archy only cares about the 'label' and 'nodes' properties, the rest are there to allow for non-archy output methods.
By providing true for the third argument to archyTree(), preparePretty, you will receive back a string, already passed through archy, complete with versions, descriptions and colouring, ready for printing to stdout.
If you install with npm install ender-dependency-graph -g (why would you?) then you'll get an ender-dependency-graph executable that you can run with a list of packages. It will scan your node_modules directory, understand the Ender-specific dependency structure and pretty-print a tree for you.
Contributions are more than welcome! Just fork and submit a GitHub pull request! If you have changes that need to be synchronized across the various Ender CLI repositories then please make that clear in your pull requests.
Ender Dependency Graph uses Buster for unit testing. You'll get it (and a bazillion unnecessary dependencies) when you npm install in your cloned local repository. Simply run npm test to run the test suite.
Ender Dependency Graph is Copyright (c) 2012 @rvagg, @ded, @fat and other contributors. It is licenced under the MIT licence. All rights not explicitly granted in the MIT license are reserved. See the included LICENSE file for more details.