Python conquors the BBC [was Re: IDLE has suddenly become FAWLTY - so should I be hitting it with a big stick, or what?]

[Twirlip2]

On Wednesday, 27 August 2014 01:21:32 UTC+1, Chris Angelico  wrote:
> On Wed, Aug 27, 2014 at 10:08 AM, Twirlip2 wrote:
> > On Wednesday, 27 August 2014 01:04:18 UTC+1, Steven D'Aprano  wrote:
> >> Twirlip2 wrote:
> >> >
> >> > Since I require Python in order to listen to my beloved radio
> >> > programmes reliably (don't get me started on the subject of the
> >> > thrice-accursed BBC website!), I therefore have IDLE running all the
> >> > time, very probably sometimes for weeks on end.
> 
> >> Well, don't keep us in suspenders, tell us what you use!
> 
> > Sorry, I don't understand. [...]
> 
> [...] But
> the rest of the question is serious. What is it you do with Idle that
> lets you listen to the radio reliably? I'm guessing this is streamed
> over the internet and you're getting around some stupid limitation,
> but I second Steven's request that you share your tricks!

I'm happy to oblige, but: (1) it's past my bedtime, so I'll have to be
brief; (2) the background, concerning the BBC's "Listen Again" facility,
is quite complicated; and (3) my present Python code is simply awful (it
was my first serious Python program, apart from a few short mathematical
routines for computing values of some complex analytic functions), which
is why, during the last week or so, I've been dipping into Chun's book
on Python, reading through Bertrand Meyer's 'Object-Oriented Software
Construction' (2nd ed. 1997), and trying to reorganised my awful code
into classes and smaller modules, so that I can begin to rewrite it in
a more robust way (i.e. (a) it will work more robustly, and (b) it also
won't fall apart as soon as I try to change it).

So, please give me a few weeks to improve my code, before posting it. (I
recently came across somewhere on the Web where you can post code, but I
forget where.)

As for the background:

All I used to need to use was a simple Windows batch file (of course, a
simple Python script would have been better, but there was no real need
for it) which called XMPlay with some WMA audio stream URLs, which were
easily calculable from the day of the week and the time of day at which
the desired programme was transmitted.

But a few months ago, the BBC, in their wisdom, changed all that, in such
a way that it is now logically impossible to calculate the WMA audio stream
URLs from any information available to anyone except a BBC technical insider.

A more sophisticated approach therefore became necessary.

At the Beebotron, where I post under the pseudonym 'Matamore!' (named after
a deluded grandiose lunatic, if you must know!), there was a very long and
convoluted discussion about the problem.  Here is the thread (but you won't
thank me for pointing you to something so rambling!):

<http://beebotron.org/phorum/read.php?5,57006>
"No international streams after 9:30am 01-May-2014"

I followed some of the suggestions made in that thread, by people more
knowledgeable than myself, and set about using Python to automate the
manual procedures they described.  I just hacked the code together [in
one of the two pejorative senses of the verb "to hack"!] - in a single
module, with dozens of global variables, precious little error-checking,
and an idiotic dependence on the BBC not changing any detail of their
website (in spite of the fact that they muck about with it regularly!).

Still, the code works, well enough that I can rely on it for my daily
fix of radio.  It would probably even work for anyone else who tried
it (on Windows, Linux, or just about any other OS). It's a mess, but
it does at least keep local dependencies in a configuration file. (I
had no trouble getting it to run on two different PCs, under both XP
and Win98SE - and, if I recall correctly, also Vista, but I never use
that laptop.)

It just pulls a lot of HTML and XML from the website, and extracts the
addresses of various other pages, and eventually *.WMA streams, and
hands the stream URLs over to XMPlay <http://www.un4seen.com/>.

It 'knows' what pages to visit, because I have manually built up a plain
text file containing a list of (at the moment) 274 BBC radio programmes,
represented by 579 different mnemonic key strings - of which the user
only needs to type in a sufficiently long initial segment to disambiguate.

Anyone else could use the same list, or build up their own, or use mine as
a basis for their own. (I just maintain it using a text editor.  I haven't
[yet] attempted to do any database programming.)

I have plenty of ideas for improving the program, but first I have to
re-organise the present spaghetti code in a more logical fashion.

That's what I've been doing for the last week or so.

(Did I say I'd be brief?  Zzzzz...)