Closed joewheaton closed 5 years ago
@banderson1618 can you scope these changes above please? It is relevant to your XML recode as in principle we should have as close as alignment as possible in the nesting and folder structure between:
Also, let me know if anything does not make sense!
Since the only major change that I'm seeing is putting the outputs in an Outputs
folder, and changing one digit numbers to two digit numbers, I think this should be a pretty quick change. I'm still focusing on XML changes, since that seems more pressing to me, but if you want, @joewheaton, I can set that down and work on this.
@banderson1618 that makes sense to focus on XML. This will play into that and it is related... so as long as you charge ahead on XML with this in mind, I think we're good. Then pick this up when it makes sense next (& before we have analysts crank through all the Idaho and TNC runs).
Cheers, Joe
HI Braden, I should have looked at this earlier, but I have not. I'm going to suggest that we change the BRAT project folder structure to be more consistent with GCD folder structure and more intuitive.
Current Project
Right now, as best I can tell you just have an
Inputs
, aTemp
folder and as manyOutput_n
that autonumber. You also have alog
file in root and aproject.rs.xml
riverscapes project file.Your inputs:
This works. I imagine adding an 05_DataCapture events for the Survey123 BRAT_cIS runs as well as Dam Surveys and Dam Complex Surveys (all three of are of type point). So maybe something like:
05_DataCapture
BRAT_cIS_Events
cIS_001
cIS_002
Dam_Events
BD_001
BD_002
Complex_Events
BDC_001
BDC_002
An Output Realization:
The outputs seem logical. It seems like you have two shapefiles out in root of
02_Analyses
and thenWhat I dislike
Output_1
make itOutput_01
so that if we get more than 10 the sort order is logical. This should apply to all realizations of inputs, intermediates and outputs (e.g. instead ofEx_Veg_1
make itEx_Veg_01
)Output_01
,Output_02
,Output_03
, ....Output_0n
should be gathered up in a parent directory ofOutputs
.I'm not sure about
I like how you use two digit, numeric prefixes to keep the sub-folders in order (e.g. 01 02) . However, this could get confusing to sum with things that are 'realizations'. I think its okay as long as 'realizations' (things there are more than one of) are kept as suffixes (i.e. after _00)
Intermediates
I see the logic in having all the intermediates in its own
Output_0n
realization folder:GCD
Here's an example of a GCD project: And this is how that directory structure is accessed by the GCD Project Explorer:
You can see how there are a bunch of libraries under inputs (DEM Surveys, Reference Surfaces, Masks and Profile Routes), which are analogous to your
01_Vegetation
,02_Network
,03_Topography
,04_Anthropogenic
. So those concepts port across nicely, and the idea of having as many different versions of inputs is also nice.Note that in GCD we have
Analyses
instead ofOutputs
. Not sure if it matters. But then we have different types of analyses likeCD
for Change Detection, andIC
for inter comparison. I do imagine we'll add new analyses to BRAT like:We should think about where these should be housed.
Thoughts @wally-mac and @philipbaileynar?