Dart workflow example #121
Conversation
…pyDARTdiags with DART, including new example to change error variance
…nstead linking to DART docs, providing direct instructions for writing Python script to half obs error variance, etc.
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #121 +/- ##
==========================================
- Coverage 71.78% 69.75% -2.04%
==========================================
Files 4 4
Lines 996 1025 +29
==========================================
Hits 715 715
- Misses 281 310 +29 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
…obs that is missing from the DART docs, since this experiment's obs seq is all identity obs. Having the user write the code to plot the rank histograms. Fixing formatting.
…nto dart_workflow
|
Formatting issues fixed |
hkershaw-brown
left a comment
hkershaw-brown
left a comment
There was a problem hiding this comment.
Part way though this review.
Major comment is to not repeat the dart documentation and focus on how pyDARTdiags fits into DART, rather than how to build and run DART.
|
I will run the example once the fix for #112 is in. |
Co-authored-by: Helen Kershaw <[email protected]>
Co-authored-by: Helen Kershaw <[email protected]>
Co-authored-by: Helen Kershaw <[email protected]>
Co-authored-by: Helen Kershaw <[email protected]>
…on; adding logic to plot a skip level binning and plot a single RH if nothing for levels is passed in
…nction in examples/03_diagnostics/plot_rank_histogram.py
…the two rank histograms into one script
|
Hi @mjs2369 can you push the the .png files, not finding them when I build the docs. /Users/hkershaw/DART/Projects/Diagnostics/JOSS/pull_121/pyDARTdiags/docs/userguide/dartworkflow.rst:211: WARNING: image file not readable: images/rh_l63.png [image.not_readable] |
hkershaw-brown
left a comment
hkershaw-brown
left a comment
There was a problem hiding this comment.
Hi Marlee,
I worked through the dartworkflow and a couple of commands are incorrect.
The text needs a little revision, as it talks about writing a program but the commands are interactive. Can you work through the example and make sure the commands run.
One question on type < 0
Cheers,
Helen
Co-authored-by: Helen Kershaw <[email protected]>
Hi Helen, fixed the commands. Sorry about that, I had manually edited them to get them to work when running through the examples in an interactive session and forgot to commit the fixes to the guide |
| print( | ||
| "Observation type is for identity observations." | ||
| ) # No filtering by type for identity obs | ||
| if qc0.empty: |
There was a problem hiding this comment.
I think you still need to filter on identity obs here.
Currently only checking for qc.0
If there is a mix of identity obs and regular obs you will plot all the identity obs and reg obs together.
I have a real obs test case, I'll go ahead and add this filter by type < 0
hkershaw-brown
force-pushed
the
dart_workflow
branch
from
a309710 to
0bc6481
Compare
regualar obs have the type,e.g. "RAW_STATE_VARIABLE" identity obs are a negative number
hkershaw-brown
left a comment
hkershaw-brown
left a comment
There was a problem hiding this comment.
Nice work on this Marlee!
This PR adds a new page to the User Guide in the pyDARTdiags documentation that introduces users to the Data Assimilation Research Testbed (DART) workflow and demonstrates how to use pyDARTdiags to enhance their DART data assimilation projects.
This page includes instructions to create an observation sequence file with DART, use pyDARTdiags to perform some basic manipulations on the observation sequence, set up and run data assimilation with DART using the Lorenz 63 model, and finally analyze the results with pyDARTdiags diagnostic plotting and observe how the changes made to the observation sequence affected the data assimilation outcome.
This fixes #117
This PR also fixes #112 by making the levels argument optional in the plot_rank_histogram function.
It adds logic to skip level binning and plot a single RH if nothing for levels is passed into the fucntion.
It also adds logic to work with identity obs as a possible value for obs_type to be passed into the plot_rank_histogram function.