search

kiyo-masui / burst_search

Program to perform real-time and off-line searches for fast radio bursts in GBT GUPPI data.
GNU General Public License v2.0
7 stars 6 forks source link

Fixed bug in timestamp used in *png and *npy filenames. #36

Closed liamconnor closed 8 years ago

kiyo-masui commented 8 years ago

Cn you strip out the extra t_unix = time.time() on line 208?

liamconnor commented 8 years ago

Sorry, why is it extra, I don't see another time.time()? I use it to identify the creation-time and make multiple copies if the same png npy files in case I need to search the data again.

kiyo-masui commented 8 years ago

Oh I see, sorry I missed that further down.

Instead of changing the naming scheme to include a not-entirely-relevant search time, could we instead just check for a previously existing file and append an integer?

liamconnor commented 8 years ago

OK, will do that next. What do you think about my removing the whole DM5-20, 20-100, etc. and just writing the actual dispersion measure into the filename? Much easier to skim through when triggers are pouring in.

kiyo-masui commented 8 years ago

Yeah, that sounds good. Rounding to an integer.

Kiyo

On Mon, Jun 6, 2016 at 10:17 AM liamconnor notifications@github.com wrote:

OK, will do that next. What do you think about my removing the whole DM5-20, 20-100, etc. and just writing the actual dispersion measure into the filename? Much easier to skim through when triggers are pouring in.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/kiyo-masui/burst_search/pull/36#issuecomment-224025259, or mute the thread https://github.com/notifications/unsubscribe/AAYXnx2nhPclPC9m_24D6JL7oaTDgMoZks5qJFY-gaJpZM4Isxxu .

liamconnor commented 8 years ago

Made the discussed changes. Now writes a %02d integer as a trigger version number.

kiyo-masui commented 8 years ago

Whoops, looks like you might have clobbered some lines I added further down in process_real_time().

liamconnor commented 8 years ago

Sorry, should be replaced.

kiyo-masui commented 8 years ago

Sorry to nit-pic, but in the file diff, the very last change moves the location of the logging output down a line. The location of this actually matters a bit since the call to get_next_block can block while it waits for data from another thread. Can you move this back? Then I'll merge.

liamconnor commented 8 years ago

Sorry, forgot about this. Does this PR make sense now?