Open ashiklom opened 5 years ago
For some addition info see: Issu #2245
Also, minor note, but there are a bunch of style guide violations in this code that makes it hard to read. For future reference, please make sure you enable RStudio's built-in code diagnostics (specifically, the "Provide R style diagnostics) option and try to address most of the issues it highlights (you may also need to install the lintr
package for this to work properly). (Or, if you use a different text editor, enable some kind of asynchronous code linter. Pretty much any editor worth its salt will have a nice package for this).
@para2x bump.....could you take a look at this and let us know if this is an easy change/fix?
@serbinsh The reason I haven't taken any actions is because I haven't run ED yet and it's hard for me to test and see if my solution would work (and spending time on learning on how to run ED
is not what I want to spend time on at the moment). But I can have a fix PR and let you guys test it.
@para2x have you tried running the integration tests? They are intended to be able to be run without having to understand how to set up a specific model. For example, you should be able to run ED2 on the VM with the following code:
./web/workflow.R --settings tests/pecan64.ed.xml
In this case, you will need to set the
<database>
<bety>
...
<write>TRUE</write>
FYI, I just ran into the bug in #2245. Got over that hurdle by changing the config file according to @dlebauer suggestion, though I'm still working through the rest of the bugs.
This issue is stale because it has been open 365 days with no activity.
At issue is this code block: https://github.com/PecanProject/pecan/blob/bb018557ccecfeed4f76d02a154ed9a147ed93a4/modules/uncertainty/R/ensemble.R#L211-L220
Even if we are not writing to BETY, we should still be able to read from it, but we are unable to do so because
con
is set toNULL
. Really, the only circumstance wherecon
should be set toNULL
is if the entiredatabase
settings block is missing.This leads to some insidious downstream bugs, as @serbinsh and I discovered. For instance, none of the input paths required for ED2 (e.g.
css
,pss
,site
,thsum
) are set, because without a database connection, we are unable to determine what the required tags for a model are in this block: https://github.com/PecanProject/pecan/blob/develop/modules/uncertainty/R/ensemble.R#L247-L259