python - Ruffus pipeline with internal inputs -

i'd create pipeline ruffus package python , struggling simplest concepts. 2 tasks should executed 1 after other. second task depends on output of first task. in ruffus documentation designed import/export from/to external files. i'd handle internal data types dictionaries.

the problem @follows doesn't take inputs , @transform doesn't take dicts. missing something?

def task1():     # generate dict     properties = {'status': 'original'}     return properties  @follows(task1) def task2(properties):     # update dict     properties['status'] = 'updated'     return properties 

eventually pipeline should combine set of functions in class update class object on go.

you should use ruffus decorators when there input/output files. example, if task1 generates file1.txt , input task2, generates file2.txt write pipeline follows:

@originate('file1.txt') def task1(output):     open(output,'w') out_file:         # write stuff out_file  @follows(task1) @transform(task1, suffix('1.txt'),'2.txt') def task2(input_,output):     open(input_) in_file, open(output,'w') out_file:         # read stuff in_file , write stuff out_file 

if want take dictionary input, don't need ruffus, can order code appropriately (as run sequentially) or call task1 in task2:

def task1():     properties = {'status': 'original'}     return properties  def task2():     properties = task1()     properties['status'] = 'updated'     return properties