Download 
Lastest release
Quick Start
Follow along with example.py
import kirb path = '.' watcher = kirb.Watcher(path)
Watchers can specify File Sets, which are a group of files to watch for changes. A File Set must first specify an output file (where the build output is written) and a list a files to monitor (relative to the Watcher's path)
watcher.add_file_set('styles.css', ['css/a.css', 'css/b.css'])
To start monitoring for file changes, call:
watcher.start()
Whenever css/a.css or css/b.css are modified, both files will be concatenated (in list order) to styles.css When your program is ready to exit, call watcher.stop()
To immediately apply the operations and callbacks to all File Sets without waiting for a file to change (i.e. may be called without start() ):
watcher.compile()
Mirrored Builds
File Sets can used to create Mirror Sets. Mirror Sets are useful for building variations of the main file set, such as static themes or customer tuned builds. Mirror Sets assume that there is an identical tree structure to a given File Set with files that may be overriden or added in the mirror directory.
my_project.add_mirror_set('styles.css', 'customerA')
Now we are watching 'customerA/css/a.css' and 'customerA/css/b.css' (if they exist), and if not we are continuing to watch the original File Set files (with original order preserved). Output for the Mirror Set will be written to 'customerA/styles.css'. Mirror Sets may also be passed an optional third argument, a list of files for only the Mirror Set to watch.
Modes
add_file_set can be passed a third argument, a dict containing a mode and callbacks for various mode-specific build stages:
watcher.add_file_set(out, js_paths, {'mode' : 'concat', 'onchange' : js_lint})
concat
This mode reads the files in the File Set line by line and writes them to the File Set's out. This is the default mode.
Callbacks
each([out, files]) <list>
Return a list to modify the files used for concatenation. This could be used when watching SCSS files to generate CSS if none is present (assume onchange callback already compiled the modified SCSS file), then return a list of css paths.
line(str) <str>
Called during concatenation with line of original file Return value is written to concatenated final
post(file_path)
Called with string path of concatenated file after concatenation
files_only
This mode calls the each handler then does nothing else. It's useful for spriting, so the spriter could be run then return False
Callbacks
each([out, files])
Same as above.
slurp
This mode reads the contents of all of files into memory, which it then passes to a required each handler. Note this each handler receives different arguments that the other callbacks. It is useful for pickling files together, i.e. to serialize template files into a single JSON object.
Callbacks
each(out, [[file_name_1, file_contents_1], [file_name_2, file_contents_2], ...])
Be careful with this, try to avoid using unless absolutely necessary.
after Watcher.start()
After a Watcher has been started, it will fire the onchange callback (all modes)
Callbacks
onchange(name of changed file)
onchange will be passed a string containing the changed file's path. This is useful for linting, which may be needlessly expensive to run on all files in the File Set. Raise an exception if an error is detected.
Other Options
In addition to mode, add_file_set can also be passed the options: * no_out
Set to False to prevent the File Set from writing its output file. Default to True. Useful for when the File Set is only used to setup Mirror Set builds.
- prefix_chomp This prefix will be removed from any files in the File Set when identifying files to be overwritten in a Mirror Set.
successful builds
Builds marked with
a lock icon are only available via PyPM to users with a current