Technical Posts
VisiMatik will occasionally publish here articles of scientific or technological interest. These may include software development discussions (e.g. life cycle management, configuration management, programming languages, C++, C#, python, Java, R) as well as articles relating to application areas (e.g. Life Sciences, bioinformatics, large datasets, algorithm complexity).
Generators and Code Inspection
Frequently, programming languages are seen as incarnations of
programming metaphores. Java is object-oriented programming, Pascal
is structured programming and LISP is functional programming.
Naturally, this is most certainly not the right way to look at it.
But the point I want to make is that one cannot look this way at
Python or at C++. They are eclectic. Python is an imperative language
equipped with all essential elements of functional language.
Importantly, it permits, nay, encourages combining these two paradigms.
This article discusses the concept of generators, partly as a
convenient bridge between the two paradigms. The example also
continues the exploration of Python as a dynamic language - this time
in a very literal way. Module inspect is used to extract useful
information from the source code of a running program.
[more...]
[source code]
Python, Pytables and meta-classes
I like to think that programming languages are means of expression of
though. Thus, it is difficult for a programmer not to develop a
preference for one language or another. At the same time, since the
choice will depend chiefly on what needs to be said (and I want to
say a great many things), I can't really have a favorite for all
seasons. But I found that the language I like most consistently is
Python. It is highly expressive and well structured at the same time.
It encourages elegance and clarity. It is modern. A good example of
what I mean is reflection: reflection is available in several modern
languages and it is served there well. But where Python stands out,
in my opinion, is the natural way of representing the access to the
execution model into the language itself.
This article gives an example of use of reflection and Python's
metaclass idiom. The presented code is a minor extension to PyTables,
a package for managing hierarchical datasets from Python. PyTables
itself and builds upon other well-know libraries: HDF5 and numarray.
HDF5 is a library and a binary data format very well suited for
organizing large volumes of numerical data and is chiefly intended
for scientific applications.
What PyTables lacked from my perspective, was a sufficient data
definition layer. Something similar in spirit to the DDL (Data
Definition Language) of SQL. I show how an appropriate mechanism can
be easily devised through a proper application of reflection and the
metaclasses. While the example is quite specific to the toolset I
use, I believe it to be quite educational and thus applicable in
other contexts.
[more...]
[source code]