bumped to version 0.5.8

- todo on seed based propagation of universities ...

rdftools/__version__.py

@@ -1,4 +1,4 @@
 __author__ = 'basca'
 
-version = (0, 5, 7)
+version = (0, 5, 8)
 str_version = '.'.join(['%s' % v for v in version])

rdftools/datagen/base.py

@@ -1,12 +1,3 @@
-"""
-an Lubm data generator that follows several data modes of distribution
-
-data mode 1: each uni is distributed to one host
-data mode 2: horizontal partitioning of all data (based on stars)
-data mode 3: randoly distribute a seed of resources to hosts, propagate from that point on (keep seed specific stuff on the same host)
-data mode 4: choose a distribution (normal) for a university and distribute the data in the university to some machines given that uni
-see substitution smapling
-"""
 import os
 import sh
 from abc import ABCMeta, abstractmethod

rdftools/datagen/lubm_horizontal.py

@@ -37,6 +37,11 @@ def get_results(self, *args, **kwargs):
  return self.site_index
 
 
+"""
+distribution process:
+
+1) horizontal partitioning of all data (based on stars)
+"""
 class LubmHorizontal(LubmGenerator):
  def __init__(self, output_path, sites, permutation='s', **kwargs):
  super(LubmHorizontal, self).__init__(output_path, sites, **kwargs)

rdftools/datagen/lubm_seed_propagation.py

@@ -2,6 +2,13 @@
 
 __author__ = 'basca'
 
+"""
+distribution process:
+
+1) randoly distribute a seed of resources to hosts,
+2) propagate from that point on (keep seed specific stuff on the same host)
+"""
 class LubmSeedPropagation(LubmGenerator):
- def _create_distribution(self, universities_rdf):
- pass
+ def _create_distribution(self, universities_rdf, **kwargs):
+ #TODO: implement me!
+ raise NotImplementedError

rdftools/datagen/lubm_uni2many.py

@@ -36,6 +36,12 @@
 is_valid_distribution = lambda distro: np.sum(distro) == 1.0
 
 
+"""
+distribution process:
+1) choose a distribution (normal) for a university
+2) distribute the data of the university to some machines given that uni
+obs: see substitution smapling
+"""
 class LubmUni2Many(LubmGenerator):
  def __call__(self, p=None):
  if p is None:

rdftools/datagen/lubm_uni2one.py

@@ -5,6 +5,11 @@
 __author__ = 'basca'
 
 
+"""
+distribution process:
+
+1) each uni is distributed to one host
+"""
 class LubmUni2One(LubmGenerator):
  def _create_distribution(self, universities_rdf, **kwargs):
  sites_index = np.random.random_integers(0, self._num_sites-1, len(universities_rdf))