[R] get compressed data via a socket connection

[Prof Brian Ripley]

?gzcon may help, depending exactly what you mean by `zlib-compressed 
data'. If not, its code will provide you with a prototype to work on. 
There are also private entry points in connections.c to (de)compress 
blocks of data to/from a raw vector which again may be a useful statring 
point for ideas.

There is no public API to the connections code.

On Wed, 8 Nov 2006, Simon Penel wrote:

> Dear R developers
>
> I am currently working on the seqinR package.  The seqinR package
> allows  a  remote access to  biological databases via a socket connection.
> We are using the  functions  socketConnection, writeLines and  readLines
> to open the socket, send request to the server  and receive response
> from the
> server respectively.
> Recently, a new function implemented in the socket server allows the
> server to send zlib-compressed data and we would like to use this
> function in R ( seqinR).
> Thus I have 2 questions about this.
> I did not found any R function dedicated to  handle compressed data from
> a socket connection.  My first question is :  Is it possible that this
> function actually exist or
> that it will exist in the future ?
>
> Here there is an example of the R fonction we use, the first time with
> no compression, the second one with compression.
>
> Type 'demo()' for some demos, 'help()' for on-line help, or
> 'help.start()' for an HTML browser interface to help.
> Type 'q()' to quit R.
>
> > library(seqinr)  # loading seqinr
> > choosebank("hoverprot") # choose the database ( this operation will
> open the socketanknameSocket$socket )
> > query("toto","k=btg@")  # do a query ( it means : give me the
> sequences associated to the keywords begining with btg, save the list in
> toto)
> > socket <- banknameSocket$socket  # here I save the socket in "socket"
> >
> request="extractseqs&lrank=2&format=\"fasta\"&operation=\"simple\"&zlib=F"
> # writing a request
> > #( it means : give me the sequence information of the list of rank 2
> ( toto) in fasta format, no compressed
> > writeLines(request, socket, sep = "\n")  # put the request into the
> socket
> > seq <- readLines(socket , n = -1) # read the answer of the socket
> server: it works, seq object contains the socket answer with sequence data
>
> >
> request="extractseqs&lrank=2&format=\"fasta\"&operation=\"simple\"&zlib=T"
> # now I try with zlib compression on
> > writeLines(request, socket, sep = "\n") #
> > seq <- readLines(socket , n = -1) #  it doesno work
> Warning message:
> incomplete final line found by readLines on '->pbil.univ-lyon1.fr:5558'
> >
>
>
> Finaly  I  begin to create a C interface to get the compressed data from
> the server.  However I am wondering how I can pass a opened socket as
> argument
> to the C interface, especially which R type defined in C  I should  use
> to handle this.  Should I  use the structure Rconn defined in
> Rconnection.h ? It seems to me that I may finally have to write a
> modified version of the C function readLines, but I think this not the
> best solution. This my second question....
>
>
> I hope this mail is not irrelevant,  thanks for any help about these
> questions
> All the best,
>
> Simon
>
>
>
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595