Fake Programmers

I compare every peak to which is the closest to my 'default toes', if the paw is rotated at a certain angle one toe is the closest to 2 default toes, so basically it overwrites the previous toe it was connected to

but I wrote that code in january or something, so I'm going to give it another shot, surely I should be able to do it one up

technically you're right as long as I can find the right toes, it doesn't matter if its interpolated or not (to a degree)

DMA57361

@IvoFlipse I guess as long as you don't use the interpolated locations in any calculations the visual location isn't as vital

Ivo Flipse

I'm now try to tag on the center of pressure

which strangely enough is upside down :\

I have no idea how to flip it around without flipping around the rest as well :\

DMA57361

how's it being added to the rest? presumably you've got to flip it before then

@IvoFlipse "paws"? ;) and it does a nice job rendering feet too, I see :)

Ivo Flipse

@DMA57361 I flipped the average contact before sending it to the zone calculating, so that one is covered

yeah, though I'm considering adding a step at 1% as a difference between black and blue, because this image should have some REALLY low pressure in the middle

and its still a tad too blue

but that's because they used to use a logarithmic scale, them cheaters :P

DMA57361

9:05 PM

lol

Ivo Flipse

this is what it looks like in the original software

so that path is pretty much correct

but upside down

DMA57361

yeah

Ivo Flipse

which I thought meant, just display it backwards

but it doesn't seem to make a difference

which probably means I'm not really turning it around :P

Ivo Flipse

9:26 PM

strangely enough, I tried flipping it in every direction

self.axes.plot(copx[::-1], copy, color='w', linewidth=3)
        self.axes.plot(copx, copy[::-1], color='r', linewidth=3)
        self.axes.plot(copx[::-1], copy[::-1], color='b', linewidth=3)
        self.axes.plot(copx, copy, color='g', linewidth=3)

yet only these two are displayed

and that's because the blue and white ones give exactly the same result, which makes sense :\

turning around x and y means its projected sideways

DMA57361

@IvoFlipse so the green and red lines in this image are meant to be reflections of one another? but they don't have the same shape

Ivo Flipse

well I guess if they are direct opposites of each other they are indeed the same

but I just don't get WHY on earth I can't flip it over...

DMA57361

have you examined the contents of copx and copy for any abnormalities?

Ivo Flipse

[ 27.18 25.99895806 25.49275103 25.10337962 24.60003155
24.16227309 24.02869261 24.1056284 24.31926957 24.69225599
25.22815432 25.84745174 26.49427813 27.13406004 27.69877527
28.1248567 28.559772 28.98424173 29.36837908 29.74601267
30.10392086 30.5022812 30.90429792 31.27817115 31.68755597
32.17207495 32.6402097 33.06565381 33.46964794 33.85773938
34.32696542 34.82708222 35.34663548 35.92 36.48951162
37.04386196 37.59465864 38.14534405 38.66371637 39.0891375

130 is at the bottom (because the contacts are upside down technically speaking)

I'm not sure its really possible by just flipping them around :\

basically I need the inverted image, transposed so it starts at the right position

pfff. I wonder where this all started that everythings upside down

I'm very tempted to rewrite this in the future, to make sure the data goes into the database in such a way that its orientation is aligned to how its going to be displayed

or I need a way to convert measurements to some global coordinate system, so I don't have to juggle around with flipping things over all the time :\

DMA57361

9:44 PM

hmmm, yeah, that'd be the "future proof" solution I guess

Ivo Flipse

9:54 PM

@DMA57361 lol well that's what I need

only without the foot being flipped as well ;P

DMA57361

lol

sort of half way there

Ivo Flipse

this is what I get by inverting one of the arrays

something's fishy

this is pretty correct, but the problem is that I had to flip my contact to get this result

which makes you wonder, WHY CAN'T I DO THIS WITH JUST THE WHITE LINE FFS

DMA57361

indeed, most strange

Ivo Flipse

because the data is stored in the database, I can't go back to my calculations and just invert the input either

that's the downside :\

DMA57361

well, I'm going to do that sleep thing I keep hearing good things about... I'll leave you to it

Ivo Flipse

10:08 PM

what's that ;P

I'll keep trying ;)

see you later

Ivo Flipse

10:53 PM

KronoS

11:08 PM

hey @IvoFlipse how ya doin?

Ivo Flipse

getting nuts

I just can't get these axes right

KronoS

o.O getting nuts or going nuts?

goes to read transcript to get caught up

Ivo Flipse

don't bother

just look at the pictures ;P

KronoS

lol...

so having issues with the data being upside down, but the images right side up?

That's very odd...

Ivo Flipse

well as you can see in the above image

I found that I can have 2 axes

and make them share the x axes (so far so good)

and I just wanted to invert the y_axis

but well... that's where it goes wrong

even if I define the axis y limits from 0-140 and 140-0

it draws an axis from 100 - 25

KronoS

11:15 PM

why do you need the axis inverted?

Ivo Flipse

because the y coordinates are correct, but generated on the image being upside down

I didn't know that when I calculated (and stored it)

KronoS

Ah...

no way to easily inverted the results?

Ivo Flipse

apparently not, or I just couldn't think of one...

KronoS

well it's all stored in a database right?

and the images are stored as pixels

which are stored in matrices

and there's probably 3 or 4 judging by the fact that there is color

so can't you just create a quick program that will read the database, and invert the matrices?

Ivo Flipse

yes, but to redo all the calculations I need to do: add function to calculate correct value, update the values in the database, create a function that triggers this post-processing and make sure it works for all the other results

KronoS

11:20 PM

meaning that if the matrix is 25x25 that the value in 1-1 becomes the value of 25-25

@IvoFlipse Ew....

Ivo Flipse

ah well, guess I need to ask an SO question

KronoS

He he.. well your rep there is getting better as you com across these questions

well on my project I realize that I need to create a unique id for patientid

Ivo Flipse

@KronoS they are unique

KronoS

something like [observerini][date][num]?

Ivo Flipse

but elaborate

you already have that

you see, rather than trying to create an unparseable index

KronoS

11:27 PM

Ah yes they are... until we have a two machines taking data at once, and then we 'merge' the databases.. there will be dupes

Ivo Flipse

unique = id, observerid, date

no there won't, because one observer can only perform one observation at a time

yes the id's will be the same, but it doesn't have to be unique

KronoS

Oh! That's what that means... I thought that by stating unique = id, observerid, date meant that ID and/or observerid and/or date had to be unique INDIVIDUALLY

not the combination of the three

Ivo Flipse

technically speaking you wouldn't import the id, just the observerid + date

nope :)

when you insert it in the main db

they get a new id, based on that db's autoincrement key

perhaps you should just make observerid + date unique

because the id won't migrate along

KronoS

@IvoFlipse speaking of which i was thinking that I might consider using mySQL for the main db and continue to use sqllite for the app...

Ivo Flipse

@TomWijsman any idea why my axes are screwed?

KronoS

11:30 PM

@IvoFlipse that makes sense

Ivo Flipse

Tom Wijsman

Is the white line right? What's that line supposed to mean?

As for the axes, you should get them to have the same offset and scale sot hey match up, and then reverse if needed and hide them from the sides (if you know how).

Ivo Flipse

well the white line in the previous image is upside down

so I added another axes to the figure and low and behold, its plotted in the right direction

but as you can see from the y axis, they aren't the same

#self.axes.set_ylim(self.ylmin, self.ylmax)
self.axes2.set_ylim(self.ylmax, self.ylmin)

even though I'm clearly setting the limits

Tom Wijsman

Doesn't it have to be set_ylim(self.ylmin, self.ylmax) instead?

And well, debugging without context is hard so I'm probably saying non-sense...

Ivo Flipse

@TomWijsman yes, that would be if I wanted the axis to be the same

but I want it upside down (as it is)