implement some more memory optimizations for article collection

[frnsys]

using a generator for membrane.collector.collect helped a lot, but the ponder process is still killed for too much memory consumption (albeit after much longer than it used to be).

so there are still some optimizations which can happen to make this not such a ridiculous memory strain.

take a look at:
http://guillaume.segu.in/blog/code/487/optimizing-memory-usage-in-python-a-case-study/

for profiling memory usage, look at heapy:
https://stackoverflow.com/questions/110259/which-python-memory-profiler-is-recommended

potentially use namedtuples?
https://docs.python.org/2/library/collections.html#collections.namedtuple

[frnsys]

dug into this a for a few days and am having a hard time narrowing down what this is. for now chunking out these large tasks into smaller, distributed ones (smaller so that they have shorter process lifespans) might be a good solution, see #112