one does not simply pickle a tree

Ahrrr... It took me 5 hours to figure it out.

Python multiprocessing.Queue is using internally shelve module, which again internally is using cPickle for speed up.  But if something cannot be pickled, when calling Queue.put, you will see some cryptic error message:


File "/opt/dirac/pro/Linux_x86_64_glibc-2.5/lib/python2.6/multiprocessing/queues.py", line 242, in _feed
-    send(obj)
TypeError: expected string or Unicode object, NoneType found

Ugh! Not helping at all, but what to do except start debugging? Taking the advantage of the opportunity I had fixed also some problems here and there (un-shelvable locks or instance methods in several classes), but at the very, very end  I've got this:

pickle.PicklingError: Can't pickle : it's not found as __main__.copyelement

 Huh? What the heck? Where is this one? Find and grep fellows stayed completely  muted, so I've started to dump symbols from *.so under PYTHONPATH, and guess what? I HAVE FOUND IT!

[volhcb13] /opt/dirac/pro > nm Linux_x86_64_glibc-2.5/lib/python2.6/lib-dynload/_elementtree.so | grep copyelement
0000000000209dd8 b elementtree_copyelement_obj



This little bastard was hiding in cElementTree! I've swapped this one with its  python twin (ElementTree). Taa-daam!  Queue.put is working again, not a big deal for me,  XML fragments I'm dealing with are rather small.

5 hours! Five! FIVE HOURS lost...

So, dear children, please remember: when going multiprocessing forget about cElementTree (and perhaps several other modules rewritten in C for speed up) and use pure python implementations.

nth permutation in a lexicographical order

Problem:

Somebody is generating permutations of a given list of elements in a lexicographical order. What is the value of say nth permutation?

Brute force solution is to generate them all, but in that case you'll spend quite a lot of time generating useless permutations - O(n!). So here is the easy way:

clogged pipe

What if? Say you want to execute something in a python sub-process and you want to read and resend all output data, no matter how big they are?

The simplest solution to use pipe is BAD idea - pipe has it's own buffer and what's worse size of this buffer is constant and cannot be changed by using some smart fnctl call any more. Let's run a simple test:



On my laptop I'm not able to send more than 64kB at a time, moreover its completely hanging:



Hmmm... So how can you send more than that? Let's say this way:



Use a simple manager, man! More info here.

There is also possibility  to use shared memory objects (read this), but those are rather for storing not so huge amount of data.

Disclaimer: Before use, read the contents of the package insert or consult your physician or pharmacist because each drug used improperly threatens your life or health.

Revert words in a sentence.

Revert words in a sentence, so " The quick brown fox jumped over the lazy dog." becomes "dog. lazy the over jumped fox brown quick The" .

In C:

In C++:


In py:


52 > 27 > 4!

python graph implementation

And now as I have promised in my previous post. it needs some more work, i.e. dumping to Graphviz format, but apart of that it is working quite well.

metaclass with dynamic creation of properties

Meteclass with dynamic properties creation: Written as helper class for my graph implementation (metaclass for nodes, edges and graph itself).

yet another Observer pattern in python

Observer sniffing a particular attribute from Observable class. This little spy is notified every time watched attribute is changing its value.

Observers are hold in weakref.WeakValueDictionary to avoid crashes in case they would disappear (i.e. being collected and deleted by gc). And here is the source code: