Lab Meeting, Spring 2015, Week 17
Awe- and Aw-inspiring readings (or just stuff you think deserves a mention)
What we did over last week
Florian
Chigusa
Andrew
- Rewrote my code for processing HIT status notifications from Amazon SQS. The old version would take several seconds even if the queue was empty. The new version seems to take no more than a second even if there are dozens of messages, and milliseconds if it's empty.
Started captioning Florian's Rutgers talk on the research page while I babysat Scott's MTurk experiments. The audio is pretty bad and if we want people to watch, it would help if you could make out more of what's being said. I've made it about 25 minutes in. In case anyone cares, I'm using AegisSub, which I'm liking so far.
- Hunted down video stimuli from our Yucatec experiments, converted it from AVI to MP4 (AVI = 3.7GB, MP4 = 230MB), and uploaded it to Box for the person who asked for them at Cornell.
- Rewrote my SPR conversion script before converting Scott's results (120 workers)
- With Olga, attempted (with no luck) to find our letter of approval for our Mexico field work, since suddenly the RSRB wants it 7 years later.
- Wrote the most evil Python function I've ever done. You'd think a Perl programmer wrote it:
def parse_region_file(region_file):
with open(region_file, 'r') as rfile:
splitreg = [x.strip().split('|') for x in rfile.readlines()]
regdict = dict([(int(reg[0]), reg[1].split(';')) if reg[1] != '' else (int(reg[0]), ['NA:0,0', ]) for reg in splitreg])
return {k: dict([(x[0], map(int, x[1].split(','))) for x in [y.split(':') for y in v]]) for k, v in regdict.items()}
Olga
Esteban
Dave
Sarah
- Practiced talk on SET for ETAP3 in Kur-Tan lab
- Edited outline of dissertation to reflect feedback obtained from SET presentation in Kur-Tan
- Met with Anne Pier to discuss the appropriate formatting of the edf output file for the new Gotta eye-tracking study
- Updated last of the files needed by my RA to finish the SET speech act annotations
- Outlined a plan of action for consolidating the annotations for the SET corpus.
Dan
Amanda
Zach
Linda