This might be happening because "gr.file_sink" is not closing the passed file
handle until its destructed. I checked "gr_file_sink_base"'s source and various
functions like open, close and unbuffered writes are supported. Try playing
around with these(I haven't used these things yet).
http://www.reynwar.net/gnuradio/sphinx/gr/sink_blk.html (Check gr.file_sink
section)
---
Regards
Sreeraj Rajendran
http://home.iitb.ac.in/~rsreeraj
________________________________
From: Nazmul Islam <mnis...@winlab.rutgers.edu>
To: GNURadio Discussion List <discuss-gnuradio@gnu.org>
Sent: Monday, 6 August 2012 8:52 PM
Subject: Re: [Discuss-gnuradio] Dynamic flow graph control
Hi,
I am trying to simplify my previous email for clarification. I am looking to
resolve the following issue:
I have two flowgraphs:
1. USRP Source --> File Sink # self.file_sink =
gr.file_sink(gr.sizeof_gr_complex*1, "Data.dat")
2. USRP Source --> Null Sink
I have defined both of them inside a class. I use unlock --> disconnect -->
connect --> lock to switch from one flow graph to the other. When I connect
USRP-->Null, I work on the stored data of "Data.dat" file. Thereafter, I want
to clear the data of the "Data.dat" file. I use the following line of code to
obtain my goal:
open("Data.dat",'w').close()
However, this command does not seem to erase the data of "Data.dat" file, i.e.,
when I run flow graph 1 again, the previous file does not get erased. The file
size of "Data.dat" keeps growing with each run of flow graph.
How can I erase the contents of the file sink before I switch to flow graph #1?
Any suggestion will be very appreciated. Sorry for sending two emails on the
same matter.
Thanks,
Nazmul
On Sun, Aug 5, 2012 at 6:46 PM, Nazmul Islam <mnis...@winlab.rutgers.edu> wrote:
Hello,
>
>I have an application where I need to collect the data for 5 seconds from the
>USRP source and do offline processing for 5 seconds in a repeated manner. I am
>planning to use the lock & unlock features of the gnuradio to obtain my goals.
>The major part of my code is given below:
>
>
>class SingleSource(grc_wxgui.top_block_gui):
>
>
>self.source =
>.....
> # Source block
>self.throttle =
>......
> # throttle
>self.file_sink = gr.file_sink(gr.sizeof_gr_complex*1, "Data.dat") #
>File Sink
>self.null_sink =
>............
># Null sink
>
>self.connect((self.source, 0), (self.throttle,
>0))
>self.connect((self.throttle, 0), (self.sink, 0))
>
> def reconf1(self):
> self.disconnect((self.gr_throttle_0, 0),(self.gr_file_sink_0,
>0)) # Disconnecting the throttle & the file sink
> self.connect((self.gr_throttle_0, 0), (self.gr_null_sink_0,
>0)) # re-connecting the throttle & the null sink
>
> def reconf2(self):
> self.disconnect((self.gr_throttle_0, 0),(self.gr_null_sink_0,
>0)) # disconnecting the throttle & the null sink
> self.connect((self.gr_throttle_0, 0),(self.gr_file_sink_0,
>0)) # re-connecting the throttle & the file sink
>
>................................
>
>
>if __name__ == '__main__':
>
> tb = SingleSource()
> var = 1
>
>tb.start()
> # Flow graph starts
> for loop in range(0,2):
>
>sleep(5) #
>Collects the data in the file sink
>
>tb.lock()
>
>tb.reconf1() #
>Flowgraph gets locked & reconfigured and unlocked
>
>tb.unlock() # Now,
>data is going to the null sink. I don't need this data
> sleep(5)
>
>................. # Offline
>processing
> open("SineData.dat",'w').close() # I
>WANT TO ERASE THE DATA IN THE FILE SINK AFTER PROCESSING
>
> # SO THAT THE FILE SIZE DOES NOT BIGGER
> tb.lock()
> tb.reconf2() # File sink
>is getting reconnected so that I can collect the data in the next run
>
> tb.unlock()
>
>Now, the problem is: the file size of "SineData.dat" should not grow with each
>loop since I am using open("SineData.dat",'w').close() . In each new loop, I
>expect the file to only contain the latest information and not the previous
>ones. However, I am finding that the file size keeps growing with each loop,
>i.e., the open("SineData.dat",'w').close() command is not working.
>
>Any suggestion will be very appreciated.
>
>Thanks,
>
>Nazmul
>
>
>--
>Muhammad Nazmul Islam
>
>Graduate Student
>Electrical & Computer Engineering
>Wireless Information & Networking Laboratory
>Rutgers, USA.
>
>
--
Muhammad Nazmul Islam
Graduate Student
Electrical & Computer Engineering
Wireless Information & Networking Laboratory
Rutgers, USA.
_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
