Code Reference

class yadage.stages.JsonStage(json, state_provider)

A stage that is defined via the JSON scheduler schemas

class yadage.stages.OffsetStage(rule, offset=None, identifier=None)

A wrapper object around a scoped rule, so that it can be applied from a global p.o.v., i.e. as adage expects its rules.


determin whether the rule is applicable. Evaluated within the offset. :param adageobj: the workflow object


applies a rule within the scope set by offset

Parameters:adageobj – the workflow object
class yadage.stages.ViewStageBase(name, state_provider)

base class for workflow stages operating on workflow views, that may expose only a partial slice of an overall workflow. The class also provides common methods to ease modifying the workflow.

Implementations are required to provide * a ready() method to implement the predicate method * a schedule() method that is called upon apply

yadage.utils.getinit_data(initfiles, parameters)

get initial data from both a list of files and a list of ‘pname=pvalue’ strings as they are passed in the command line <pvalue> is assumed to be a YAML parsable string.


returns a content-based hash of a json-able structure (serialized with WithJsonRefEncoder)

Parameters:jsonable – a json-serializable object
Returns:the hash

generator function to yield leafs items of a JSON-like structure alongside their position in the structure as determined by a JSONPointer.

Parameters:jsonable – a json-serializable object
Returns:tuples (jsonpointer, leaf value)
yadage.utils.leaf_iterator_jsonlike(jsonlike, path=None)

‘ :param jsonlike: a jsonlike object (i.e. nested lists/arrays: leaf-types must not be JSONable) :return: iterator that yields (JSON path, leaf value) tuples

yadage.utils.pointerize(typedleafs, asref=False, stepid=None)

a helper method that replaces leaf nodes in a JSON object with a outputReference objects (~ a JSONPath) pointing to that leaf position useful to track access to leaf nodes later on.

yadage.utils.set_backend(dag, backend, proxymaker=None)

sets backend and proxies for each node in the DAG. proxymaker is a 1-ary function that takes the node object and returns a suitable result proxy

  • dag – the dag object (par of the workflow object)
  • backend – the backend to associate with the nodes
  • proxymaker – function object with signature f(node) to create result proxies