Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 07:14
    rahul799 commented #10
  • 05:58
    robkooper commented #10
  • Jan 23 20:28
    para2x commented #2522
  • Jan 23 20:25
    araiho commented #2522
  • Jan 23 18:16
    para2x commented #2522
  • Jan 23 18:08
    para2x commented #2433
  • Jan 23 16:14
    infotroph labeled #2338
  • Jan 23 16:14
    infotroph unlabeled #2338
  • Jan 23 16:14
    infotroph labeled #2521
  • Jan 23 15:06
    araiho synchronize #2522
  • Jan 23 15:03
    araiho edited #2522
  • Jan 23 15:02
    araiho edited #2522
  • Jan 23 15:00
    araiho review_requested #2522
  • Jan 23 15:00
    araiho labeled #2522
  • Jan 23 15:00
    araiho labeled #2522
  • Jan 23 15:00
    araiho labeled #2522
  • Jan 23 15:00
    araiho assigned #2522
  • Jan 23 15:00
    araiho opened #2522
  • Jan 22 19:50
    KristinaRiemer closed #2296
  • Jan 22 19:50
    KristinaRiemer commented #2296
Sameroom
@sameroom-bot
[David LeBauer, PEcAn] This is similar behavior to what dplyr does with database queries ... just pulls a few rows into memory.
Sameroom
@sameroom-bot
[Chris Black, PEcAn] @dietze @dlebauer yeah, the preview is printed because I specifically requested it with glimpse() — I do endorse benchmarks, but I’ll also note that it’s written by Michael Sumner, who works in high-dimensional geospatial data full time and I trust him to get it right if anyone can.
Sameroom
@sameroom-bot
[Michael Dietze, PEcAn] I was just testing the last release VM before the PEcAn demo at the ED meeting this afternoon. Shiny is still not right -- let's me select the workflow and run ID's (NOTE: THERE ARE STILL TYPOS IN THE GUI) and then crashes when I click load data. Have others encountered this issue? Is there a bug fix I can apply before cloning the AWS VM out for everyone to play with? FYI the specific test run is ED2 at Harvard for 2004 using default met. @tonygard @kooper @ashiklom @emcowdery
Sameroom
@sameroom-bot
[Chris Black, PEcAn] @dietze This is with the 1.5.2 VM? I get the same behavior and have been hoping somebody else will fix it before I have to look any closer.
Sameroom
@sameroom-bot
[Tony Gardella, PEcAn] @dietze Not aware of a bug fix and I don't think I could troubleshoot by the afternoon. I'm definitely going to add this as a discussion topic for tomorrow. I think we need a few people to take a look at shiny.
[Michael Dietze, PEcAn] Yes, this is the 1.5.2 VM
Sameroom
@sameroom-bot
[Tony Gardella, PEcAn] @Channel PEcAn weekly telecon tomorrow (1/17) at 11:00 am eastern. Please add your weekly update to the google document titled "Meeting 1/17/18" in the pecan google drive. Jitsi link will be posted 10 min before call.
Sameroom
@sameroom-bot
[David LeBauer, PEcAn] @tonygard Sorry - I won’t be able to make the meeting
Tess McCabe
@mccabete
@mdietze are @istfer and I and you going to be on the PEcAn call?
Sameroom
@sameroom-bot
[Alexey Shiklomanov, PEcAn] @black11 Adding PEcAn.settings as a dependency for PEcAn.DB creates a circular dependency, because PEcAn.settings also depends on PEcAn.DB.
[Chris Black, PEcAn] @ashiklom D’oh! Right. Let me see where it’s used.
[Michael Dietze, PEcAn] Folks at the ED meeting will miss today's PEcAn meeting. @tonygard I've been looking at the PR log this AM -- could you get updates on today's call about which PR that have 'changes requested' are ready for review and which require nudging the author to make fixes. Would really like to clean up our PR backlog between today and next week's call
[Alexey Shiklomanov, PEcAn] Looks like the only offending line is in runModule.get.trait.data. Could we just move that whole function to PEcAn.settings? Or, alternatively, it might make sense to just move all of the runModule functions to PEcAn.all?
Sameroom
@sameroom-bot
[Michael Dietze, PEcAn] Conceptually I don't think it makes sense to migrate any runModule to settings. 'all' serves a different aggregation purpose and contains no code. Should there be a new package in base for these?
[Michael Dietze, PEcAn] Also, what settings function is runModule.get.trait.data actually calling?
[Chris Black, PEcAn] Looks like it calls PEcAn.settings::is.Settings and PEcAn.settings::is.MultiSettings
Sameroom
@sameroom-bot
[Chris Black, PEcAn] I’d been talking about taking a bunch of the PEcAn.utils functions that are more “actually do the work” than “provide a tool to use when doing the work” and moving to a new package called something like PEcAn.workflow. Would that be an appropriate home for runModule.get.trait.data?
[Chris Black, PEcAn] My list so far of the utils functions I was thinking about moving: do_conversions, get.results, run.write.configs, ensemble, get.parameter.samples
[Michael Dietze, PEcAn] @black11 this is where threads is failing us --- @ashiklom and I were just discussing that very thing
[Chris Black, PEcAn] Bah. Sorry ’bout that.
[Michael Dietze, PEcAn] no worries -- ironically we independently agreed on the package name of 'workflow'
[Chris Black, PEcAn] I might have gotten that from one of you last fall
Sameroom
@sameroom-bot
[Alexey Shiklomanov, PEcAn] I strongly endorse the idea of moving "actually do the work" functions to PEcAn.workflow.
[Alexey Shiklomanov, PEcAn] As an aside, I think we should use Slack threads much more aggressively -- they're the main reason I really wanted to move to Slack in the first place, as they tie together relevant conversations without polluting the main chat.
[Shawn Paul Serbin, PEcAn] @ashiklom good point. One major pain with Gitter is that if I want to follow a "thread" but I am not watching closely, it can get buried far up and in bits throughout the main chat window....but if we can keep to threads, much easier to keep tabs / lurk on converations
Sameroom
@sameroom-bot
[Alexey Shiklomanov, PEcAn] @emcowdery https://github.com/stan-dev/shinystan
Sameroom
@sameroom-bot
[Chris Black, PEcAn] @ashiklom does the “upstream problems” in your commit message for #1816 mean you already tried having it install from https://git.bioconductor.org/packages/rhdf5? I tried briefly and failed, but assumed I was just misunderstanding the remotes: syntax.
[Chris Black, PEcAn] …Heh, just saw your comments in #1806. Woo, juggling communication channels!
Sameroom
@sameroom-bot
[Ann Raiho, PEcAn] @dietze is the allometry vignette the right place to start to figure out how to run dbh's through the allometry module?
Sameroom
@sameroom-bot
[Michael Dietze, PEcAn] @ann.raiho yep!
[Ann Raiho, PEcAn] I ran allom.stats = AllomAve(pfts,ngibbs=500,components=c(3)) and got a length = 6 list but everything but all the elements are NULL except for 6
@ann.raiho yep!
Sameroom
@sameroom-bot
[David LeBauer, PEcAn] @dietze Are there plans for the DART module?
Sameroom
@sameroom-bot
[Michael Dietze, PEcAn] @dlebauer There is no active development going on in DART at the moment, but Toni Viskari still has one DART-based paper that he owes us and one published paper that references that module (so I'd be hesitant to deprecate it until a major release)
Sameroom
@sameroom-bot

[Michael Dietze, PEcAn] Just came across this R package and thought it might be handy in PEcAn for things like user- or project-specific defaults/settings

rappdirs: Application Directories: Determine Where to Save Data, Caches, and Logs
https://cran.r-project.org/web/packages/rappdirs/index.html

Sameroom
@sameroom-bot
[Michael Dietze, PEcAn] ED2 users: Marcos just posted a nice slide explaining the structure of patch- and cohort-level variables in the output & history files. https://github.com/EDmodel/ED2/files/1647126/paco_id.n.pdf
@istfer @tmccabe @ashiklom
istfer
@istfer
just saved it :) :+1:
Sameroom
@sameroom-bot
[Michael Dietze, PEcAn] Just saw that they're having the AMS Biogeosciences / Ag For Met meeting in Boise this year in May. Is anyone thinking of going? @sserbin @ankurdesai https://www.ametsoc.org/ams/index.cfm/meetings-events/ams-meetings/33agforst-12fire-4biogeo/?utm_source=Subscribers&utm_medium=Email&utm_campaign=Newsletter&_zs=8xhhc1&_zl=8LjT4
Ankur Desai
@ankurdesai
@mdietze Yes, I will go to some of AFM/Biogeo, that's where i sent in the Mandifore abstract
Alkesh Srivastava
@alkesh47
hi everyone, I'm a gsoc 2018 aspirant and would love to contribute to the project. Could someone guide me to some good first issues so that I may get the dice rolling..
Shawn P. Serbin
@serbinsh

@/all I am having trouble enabling data ingest on shiny due to package depends. I am trying to get the redland depend installed but hitting this wall after build and on test load:


    rasqal_version_release_get              html
    rasqal_version_string                   html
    rasqal_version_string_get               html
    redland                                 html
    roc_output.mergeNamespace               html
    roc_process.mergeNamespace              html
    serializeToCharacter                    html
    serializeToFile                         html
    setNameSpace                            html
    setQueryResultsLimit                    html
    sub-ExternalReference-method            html
    subset-ExternalReference-method         html
** building package indices
** installing vignettes
** testing if installed package can be loaded
Error: package or namespace load failed for ‘redland’ in dyn.load(file, DLLpath = DLLpath, ...):
 unable to load shared object '/data/home/sserbin/R/library/redland/libs/redland.so':
  /data/home/sserbin/R/library/redland/libs/redland.so: undefined symbol: librdf_serializer_check_name
Error: loading failed
Execution halted
ERROR: loading failed
* removing ‘/data/home/sserbin/R/library/redland’

The downloaded source packages are in/tmp/RtmpvYv88H/downloaded_packages’
Warning message:
In install.packages("redland") :
  installation of package ‘redland’ had non-zero exit status

Anyone else see this before?

Anthony Gardella
@tonygardella
Hello @alkesh47! We are excited about GSOC 2018, but organizations are not notified of their acceptance until February 12th. That being said, we are always excited to have new developers. As a start I'd recommend going through our tutorials to familiarize yourself with PEcAn and then looking through the project ideas posted on our GSOC page. Once you have an idea of how PEcAn works and areas we are looking for improvement, it would be great to hear about what your interests are so we can assign you an issue to work on. If you have any questions along the way feel free to ask here or on github!
Alkesh Srivastava
@alkesh47
Thanks for the info @tonygardella I'll start getting familiar with the codebase and am excited to start with making contributions to the project
femeunier
@femeunier
Hi all, I am a ED developer and I would like to use PEcAn to run sensitivity analysis of a new PFT that my lab colleagues implemented within the model. I'm new to PEcAn but I'm already familiar with the VM interface and I could successfully run the two demos described online. Basically what we would like to do first is to use our current version of the model to test the impact of some parameters on several model outputs (i.e. starting from basics). I cloned pecan from github and could install the necessary packages but now what should I do? Is there any tutorial or documentation to use PEcAn on a local machine (ubuntu)? Is that even recommended? Any help/advice/link to useful material would be appreciated!
Anthony Gardella
@tonygardella
Hi @femeunier I'm happy to help. I'll write up some steps now and post them soon. I do not recommend installing PEcAn locally because you'll need to install the BETY PostgreSQL database as well and will take time to setup. Working on the VM will be a lot easier. In the meantime if you have time read section 14.1.1 on this page. This page outlines how to add a model that does not exist. Adding a different version of a model type, in this case ED2, will have slightly different steps and will be easier to do. I do have one question however, Do you know if any of the pfts we already have associated with the ED2 model are similar to the pfts you would like to add?
Anthony Gardella
@tonygardella
Hi @femeunier I'll put these in our documentation tomorrow. For now, here are the steps
Adding New Version of a Model
  1. Compile your model on the machine you will be executing PEcAn from. If you are using the VM you can use the secure copy file linux command to move files from your machine to the VM: scp -P 6422 your_files_or_files carya@localhost:/home/location/you/want
    Or simply open Rstudio through a browser - http://localhost:6480/rstudio/ and use the "Upload" button to move files onto the VM.
  2. Open BETY (http://localhost:6480/bety) and navigate to the files section. (username - carya, psswrd- Illinois
  3. Create a file record that points to where your model executable is. Go to "Runs tab" select "Files" select "New File". Set Machine to "pecan.vm", file path to "local/path/to/file/", file name to "executable_filename"
  4. Add to the models table. Go to "Runs" tab select "models" and select "New Model". Set Model name to "Edfemeunier version" (name can be anything but keep it short and ), Revision can be a small descriptor of your version "pftoffemeunire" , Modeltype set to "ED2" and then you can set parent if you know your version model is a change to a certain model we have.
  5. Now your version of the model can be recognized by PEcAn and will use the pfts already associated with your model. If you would like to add species or a prior to an existing model you'll just need to create new species in the species tale and do the same thing for priors. If you would like to create entirely new PFTs for your model you can follow the instructions found in: https://pecanproject.github.io/pecan-documentation/develop/adding-to-pecan.html
  6. For ED2 model specifically however when you will need to do a few extra things:
femeunier
@femeunier
Thank you so much @tonygardella for the information. It helped a lot. I was able to proceed all the steps you described and I added my model. I was just now wondering how to use this new model from the command line in Rstudio. I tried to modify the settings manually (link to the binary) from a workflow I generated using the PEcAn VM but it didn't work. I couldn't find any help in the documentation but perhaps I did not look at the appropriate place ...
Anthony Gardella
@tonygardella
@femeunier The key to running pecan from the command line in Rstudio is understanding the pecan.xml(pecan settings file). You'll find it also is the output folder in any of the runs you've done so far. Within the VM runs are kept in : /home/carya/output/. Once you have a settings file that is set up properly, to run pecan you can either open the workflow.R script in the pecan/web directory and run it line by line with your settings file in your working directory or simply execute that script and pass it the settings file. Either way, it's a matter of having the settings file and running the workflow.R with it available.
When you say you modified the settings manually, did you put in the path to the model binary?
Anthony Gardella
@tonygardella
@/all Hello everyone, we are officially moving our chat communications to slack. Our gitter channels will remain open, but expect delayed responses if you post. To join our slack channels simply click the following link and enter your email: https://publicslack.com/slacks/pecanproject/invites/new
Alkesh Srivastava
@alkesh47
@tonygardella I did some research and read through the codebase. I'm interested in the Bety port from Ruby to Python project. So how should I proceed, would you kindly guide me to the relevant issues or shall I pick the liberty of completing the task mentioned in the GSOC application form?