Translating Javascript programs to python.

[Vyz]

Yes, this is a trimmed down version of padma, a generic indic
transliteration tool
Thanks for comments though.

Terry Reedy wrote:
> "Vyz" <vyzasatya at gmail.com> wrote in message
> news:1156371324.186510.273930 at 75g2000cwc.googlegroups.com...
> > Its a module to transliterate from telugu language  written in roman
> > script into native unicode. right now its running in a browser window
> > at www.lekhini.org I Intend to translate it into python so that I can
> > integrate into other tools I have. I am able to pass arguments and get
> > output from the script also would be OK. or how about ways to wrap
> > these javascript functions with python.
>
> Leaving aside the code the manipulated the display and user interaction,
> the code should be pretty straightforward logic (if-else statements) and
> table lookups, so translation to Python should be straightforward also.
>
> I checked parser.js.  I don't know javascript but it looks to me like a
> mixture of C and Python.  The for loop headers have to be rewritten, and
> the switch changed to if-elif.  What looks different is the attachment as
> attributes of method functions to functions rather than classes.
>
> As for 'wrapping': can you get a standard javascript interpreter?  If so,
> you could possibly adjust the js so you can pipe a roman string to the js
> program and have it pipe back the telegu unicode version.
>
> >> > I have a script with hundreds of lines of javascript spread accross 7
> >> > files. Is there any tool out there to automatically or
> >> > semi-automatically translate the code into python.
>
> unicode.js is mostly a few hundred verbose lines like
>
> Unicode.codePoints[Padma.lang_TELUGU].letter_PHA  = "\u0C2B";
>
> that setup the translation dict.  Because the object model is different, I
> suspect that these all need to be changed, but, I also suspect, in a
> mechanical way.
>
> If one were starting in Python, one might either just define a dict more
> compactly like
>   TEL_uni = {letter_PHA:"\u0C2B", ...}
> *or* probably better, use the builtin unicodedata module as much as
> possible.
>
> >>> import unicodedata as u
> >>> pha = u.name(u'\u0c2b')
> >>> pha
> 'TELUGU LETTER PHA'
> >>> u.lookup(pha)
> u'\u0c2b'
>
> I don't know what you do with js statement like this:
> Unicode.toPadma[Unicode.codePoints[Padma.lang_TELUGU].misc_VIRAMA +
> Unicode.codePoints[Padma.lang_TELUGU].letter_KA] = Padma.vattu_KA;
> where a constant seems to be assigned to a sum.  But whatever these do
> might correspond to the u.normalize function.
>
> This appears to be based on a generic Indian-script transliteration program
> (Padma), so there may be functions not really needed for Telegu.  (I am
> familiar with Devanagri but know nothing of Telegu and its script except
> that it is Dravidian rather than Indo-European-Sanskritic.)
> 
> Good luck.
> 
> Terry Jan Reedy