words
stringlengths 0
4.03k
| end_time
float64 0.46
2.95k
| speaker
stringclasses 4
values | start_time
float64 0
2.93k
| session_id
stringclasses 100
values |
|---|---|---|---|---|
way it doesn't crash if you try and load all ten in one it crashes yeah i think probably it's a bit dumb if you can fool it if you load up ten different engines simultaneously it can do that c can't do them can't do them if it thinks it one because then it's about the amount that each | 907.69 | speaker2 | 890.6 | session_0ed514d2 |
if | 908.83 | speaker1 | 908.56 | session_0ed514d2 |
you have to kind of call a new class and then it will do it fine but if you don't i like hate | 912.71 | speaker2 | 908.84 | session_0ed514d2 |
mo | 913.32 | speaker4 | 912.72 | session_0ed514d2 |
yeah | 914.27 | speaker1 | 913.92 | session_0ed514d2 |
says okay oh yeah i've got all this space you do otherwise it goes oh no | 919.36 | speaker2 | 914.28 | session_0ed514d2 |
talk | 922.47 | speaker1 | 922.24 | session_0ed514d2 |
to you now | 923.36 | speaker4 | 922.48 | session_0ed514d2 |
a | 926.07 | speaker2 | 925.92 | session_0ed514d2 |
nice | 926.39 | speaker1 | 926.08 | session_0ed514d2 |
machine | 927 | speaker2 | 926.4 | session_0ed514d2 |
you spent too much time think alrigh well i've been i've been doing some random pop up windows for the speaker characterization which is good so i can fill that in now okay the windows are all there now after lots filling | 942.44 | speaker4 | 928.88 | session_0ed514d2 |
oh they done are they okay cool | 944.51 | speaker2 | 942.68 | session_0ed514d2 |
well you know just the pop a pop up window they pop up is difficult enough because it it basically has to set up like whether ten pop up windows in memory as in already initialize them because there might be ten speakers and all have to have different meet windows and | 962.9 | speaker4 | 944.52 | session_0ed514d2 |
and | 966.39 | speaker2 | 966.08 | session_0ed514d2 |
i'm working on when you click on a topic to see a list of the of the topics of of the meetings that that topic was mentioned okay and i have a a question about the we said we wanted a start screen | 981.93 | speaker4 | 966.4 | session_0ed514d2 |
what | 984.27 | speaker2 | 984 | session_0ed514d2 |
kind | 984.43 | speaker1 | 984.28 | session_0ed514d2 |
of start screen do we want i mean do we want some general blurb about welcome to the browser | 990.8 | speaker4 | 984.44 | session_0ed514d2 |
more letters | 992.6 | speaker1 | 991.84 | session_0ed514d2 |
letsd browser some speech and some music animation' take too long we have | 1,002.27 | speaker2 | 992.72 | session_0ed514d2 |
to while we're debugging probably getting sick of it but or something that just sort of guides them to the most obvious things yeah | 1,010.23 | speaker1 | 1,002.28 | session_0ed514d2 |
so | 1,010.39 | speaker2 | 1,010.24 | session_0ed514d2 |
do we want yeah maybe we could write un like the whatever it's called the top of the window saying welcome to the mi browser | 1,018.36 | speaker4 | 1,010.4 | session_0ed514d2 |
yeah or a switchboard that comes up that's just a blank form like that with some buttons on load me a meeting load me a search load me something else yeah whistle with you | 1,027.86 | speaker2 | 1,018.6 | session_0ed514d2 |
yeah but how i mean at the moment the way is it's it's a drop down menu as you've seen | 1,035.12 | speaker4 | 1,029.24 | session_0ed514d2 |
with | 1,036.63 | speaker2 | 1,036.44 | session_0ed514d2 |
all the and that because it's the dialogu class that you use there | 1,043.75 | speaker4 | 1,036.64 | session_0ed514d2 |
yeah i was just going to ask if we should assume that they'll want to load a meeting first but not necessarily i guess if they're doing a search on the entire | 1,050.62 | speaker1 | 1,044.56 | session_0ed514d2 |
yeah i mean i think they should have two options either load up a meeting that they choose or do a search yeah and that once you've got the results you click on something andt | 1,061.24 | speaker4 | 1,051.56 | session_0ed514d2 |
search | 1,062.04 | speaker1 | 1,061.44 | session_0ed514d2 |
need you to have a meeting loaded before you'll start doing any search | 1,065.16 | speaker2 | 1,062.24 | session_0ed514d2 |
well we can does it put the object in there yeah you can't | 1,068.43 | speaker1 | 1,065.6 | session_0ed514d2 |
the only search is an night object model and the only time you get one of those if you've loaded an observation | 1,072.83 | speaker2 | 1,068.44 | session_0ed514d2 |
you could probably even that' shouldn't be too hard to put a check in there and if it hasn't loaded then force them to load one yeah | 1,079.39 | speaker1 | 1,072.84 | session_0ed514d2 |
which we but doesn't you could use the inverted file search to return a list of meetings and then use one of those to load and search but you won't can you not get this nom | 1,092.51 | speaker2 | 1,079.4 | session_0ed514d2 |
object | 1,093.12 | speaker4 | 1,092.52 | session_0ed514d2 |
it has to be an observation and even if you go you can go and search the whole corpus from that but you have to have it has to start with something for some bizarre reason the engine there's only one there's only one search method to the search engine the engine class only has i suppose we wouldn't | 1,108.95 | speaker2 | 1,093.64 | session_0ed514d2 |
want to use yeah yeah we wouldn't want to use nxt at the beginning anyway because we don't want to search locally we | 1,114.63 | speaker4 | 1,108.96 | session_0ed514d2 |
want to sech soless | 1,117.72 | speaker2 | 1,114.64 | session_0ed514d2 |
they know what meeting it is well yeah | 1,119.93 | speaker1 | 1,117.8 | session_0ed514d2 |
one no then they would open the meeting first right | 1,122.47 | speaker4 | 1,120.32 | session_0ed514d2 |
so | 1,123.93 | speaker2 | 1,123.8 | session_0ed514d2 |
let's | 1,124.27 | speaker1 | 1,124.04 | session_0ed514d2 |
say they have two options either pick one meeting or search globally to find a meeting that they like | 1,130.78 | speaker4 | 1,124.28 | session_0ed514d2 |
yeah | 1,135.39 | speaker1 | 1,135 | session_0ed514d2 |
or | 1,135.67 | speaker2 | 1,135.4 | session_0ed514d2 |
well the things they may just be looking for a word and you know like and if it shows up ten different meetings then at that point they'll probably want to like do we want to like if it don't by default goes with the first one and they want number five of the ones that are returned i mean then they have to go through that no but why do we want default i mean just well not like to load up like the thing is i think | 1,154.26 | speaker1 | 1,135.68 | session_0ed514d2 |
like | 1,156.23 | speaker4 | 1,156.12 | session_0ed514d2 |
when it loads up it'll load up the transcript window and the whatever other window right now i think just the topic ones although the time is probably more ummm caused by loading the actual data so i'm just thinking if it's kind of like you were saying you have to have a nom object right i mean before you do a search at all yeah but local search yeah we can do global search anything | 1,180.21 | speaker1 | 1,156.24 | session_0ed514d2 |
doesn to | 1,181.99 | speaker4 | 1,181.28 | session_0ed514d2 |
load up anyway we can load the dumb one up it doesn't have any | 1,184.87 | speaker2 | 1,182 | session_0ed514d2 |
yeah but the o the global one feeds into the local so for getting any the usable data it's going to be doing a search on each of those files so it's going to be loading up each of those at a time to get the data yeah the global | 1,197.48 | speaker1 | 1,184.88 | session_0ed514d2 |
inverter file search gives the nom objects basically to the local search | 1,203.55 | speaker4 | 1,198.48 | session_0ed514d2 |
right right but it could be time consuming like if there are ten documents that well yeah that's always | 1,208.67 | speaker1 | 1,203.56 | session_0ed514d2 |
going to be the case with search | 1,209.74 | speaker4 | 1,208.68 | session_0ed514d2 |
but | 1,210.88 | speaker1 | 1,210.84 | session_0ed514d2 |
and it gives one at a time anyway doesn't it because otherwise it'll crush the thing | 1,214.24 | speaker2 | 1,210.88 | session_0ed514d2 |
yeah well i'm just thinking of some way we could you know cache the results in in a nice little format that' will make things a bit easier but the thing is i mean if we've got that then it's gonna be needing really much really to load the entire corpus for that meeting if it's trying to show us where those were if it's trying to highlight those in the text transcript or whatever and we're gonna have to have all that other data in there so each time when we have a search window and we have like you know ten different meetings you know the word wireless comes up go to meeting one then has to reload the nonom object next one i mean that's gonna yeah it's not too slow though i think it's not not too bad on okay i think that will be a pretty much a problem it always seems to be slow loading up the first time at least like that's all i lately is just sort of loading it up tested try some else and shut down load it up | 1,263.62 | speaker1 | 1,214.76 | session_0ed514d2 |
so so | 1,266.74 | speaker2 | 1,264.8 | session_0ed514d2 |
just sort of based on that as well yeah but if it doesn't take that long each time then that should be all right too i mean you can | 1,276.27 | speaker1 | 1,266.84 | session_0ed514d2 |
you know the option is to do a search in the way that if you do a global search first as a first step only return the meetings as in not even the meetings but the name of the meetings so you know say you want to search for the word language then it gives you as a result all the meetings that contain the word language and then you can decide well i want to search on this and this and this meeting or only on this meeting or | 1,301.21 | speaker4 | 1,277.04 | session_0ed514d2 |
make it strings for as long as possible and then i things when they actually but it needs to has to search | 1,308.36 | speaker2 | 1,302.88 | session_0ed514d2 |
and that way if they know specifically which meeting it is they'll save the time because i'm sure they don't want to have the extra loading time either if they can avoid it definitely so or they could check mark against the ones they want to check so | 1,321.98 | speaker1 | 1,310.48 | session_0ed514d2 |
did we think about better names for the meetings | 1,328.18 | speaker2 | 1,322.48 | session_0ed514d2 |
well i think they were in the text we can probably do that ourselves and just sort of b d b you don't just do a there is | 1,335.72 | speaker1 | 1,329.88 | session_0ed514d2 |
yeah in one of the papers they have names | 1,340.92 | speaker4 | 1,336.28 | session_0ed514d2 |
do they do they understanding or like the working group that it's part of okay that probably even better | 1,349.55 | speaker1 | 1,342.12 | session_0ed514d2 |
understanding or something | 1,350.18 | speaker2 | 1,349.56 | session_0ed514d2 |
yeah | 1,350.66 | speaker4 | 1,350.52 | session_0ed514d2 |
well | 1,351.27 | speaker1 | 1,351.08 | session_0ed514d2 |
that's that's a working group is it i love that i want to see the meetings about they even better understanding | 1,357.92 | speaker2 | 1,351.28 | session_0ed514d2 |
yeah | 1,359.53 | speaker4 | 1,359.2 | session_0ed514d2 |
super accurate so' | 1,361.87 | speaker1 | 1,360 | session_0ed514d2 |
yeah | 1,362.15 | speaker2 | 1,361.88 | session_0ed514d2 |
i've put that in already too | 1,363.55 | speaker4 | 1,362.16 | session_0ed514d2 |
okay that's cool | 1,365.4 | speaker2 | 1,364.36 | session_0ed514d2 |
so yeah again concretely to the startu startup wbraries window so | 1,380.43 | speaker4 | 1,368 | session_0ed514d2 |
who would have the ummm idfs i meaning the dfs the document frequencies for each word | 1,387.08 | speaker2 | 1,380.44 | session_0ed514d2 |
do you actually need the frequencies yeah | 1,390.83 | speaker1 | 1,387.48 | session_0ed514d2 |
to do what ks talking about to do the topic label | 1,395.11 | speaker2 | 1,390.84 | session_0ed514d2 |
write that right you don't okay yeah how is | 1,397.23 | speaker1 | 1,395.12 | session_0ed514d2 |
that i mean do you think you would have have time for that as well | 1,400.74 | speaker4 | 1,397.24 | session_0ed514d2 |
well thing is if it's being done somebody's done the key | 1,403.59 | speaker1 | 1,401.44 | session_0ed514d2 |
so the idfs the dfs already would be not too | 1,407.03 | speaker2 | 1,403.6 | session_0ed514d2 |
probably there was a class she was trying to get to work on friday they did have all that and it was sort of built in and just had some trouble getting it running properly but it had sort of all that standard sort of stuff but it at last i heard it wasn't working so she was gonna look at you know more straightforward sort of thing that really just fulfills what we needed to do but if need the if we do need the frequencies | 1,430.56 | speaker1 | 1,407.04 | session_0ed514d2 |
how do you do an inversive file search without the cfdf | 1,432.96 | speaker2 | 1,430.8 | session_0ed514d2 |
i don't know | 1,434.66 | speaker1 | 1,434.2 | session_0ed514d2 |
i think you need to have that would have been like af | 1,439.68 | speaker2 | 1,436.28 | session_0ed514d2 |
stand for the | 1,440.83 | speaker1 | 1,439.84 | session_0ed514d2 |
amount that they occur over documents basically the amount they what does they stand for thef is the document frequency the amount that each word occurs no what is it term frequency is the amount that it occurs in each one of them is the amount that occurs in each document and the other one is the amount it occurs generally so if you so the more it occurs in specific documents compared to it general score the better more informative it is about a so the general score | 1,469.75 | speaker2 | 1,440.84 | session_0ed514d2 |
would that be for the whole of the language or the | 1,472.96 | speaker4 | 1,469.76 | session_0ed514d2 |
call data well you know general frequenies general plus a stop list so you remove stuff that doesn't like and then which is going to have basically the equal score so what would the | 1,486.51 | speaker2 | 1,473.52 | session_0ed514d2 |
if we had these fake topics what would they look like | 1,489.67 | speaker4 | 1,486.52 | session_0ed514d2 |
bunch of keywords a bunch | 1,491.3 | speaker2 | 1,489.68 | session_0ed514d2 |
that's | 1,492.35 | speaker4 | 1,492.12 | session_0ed514d2 |
probably the easy just | 1,493.19 | speaker2 | 1,492.36 | session_0ed514d2 |
a list of like | 1,494.07 | speaker4 | 1,493.2 | session_0ed514d2 |
three words or something yeah keyords three words | 1,496.95 | speaker2 | 1,494.08 | session_0ed514d2 |
what's | 1,497.11 | speaker4 | 1,496.96 | session_0ed514d2 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.