Closed NiklasPhabian closed 1 year ago
@NiklasPhabian What's going on with this merge test?
starepandas.io.pod.read_pods() doctest seems to fail.
@NiklasPhabian , there's an error in some SQL based code.
FAILED examples/catalog.ipynb:: - AttributeError: 'OptionEngine' object has no attribute 'execute'
FAILED examples/to_database.ipynb:: - AttributeError: 'OptionEngine' object has no attribute 'execute'
I don't know anything about this code...
What's going on with this pull request?
So the tests pass; I think all the LFS issues should be resolved. The issue was related to changes in sqlalchemy>=2.0
, which pandas has not addressed yet. I pegged sqlalchemy, but I am sure that pandas will solve this very soon.
@michaelleerilee: I'd really like to avoid stashing codeblocks in comments; especially when it goes into main/master (https://github.com/SpatioTemporal/STAREPandas/blob/hcp/starepandas/staredataframe.py#L1186)
@michaelleerilee do you think you can hack a pod example notebook together that we could add to the examples/
folder to have at least some tests going or should we close the PR right away?
Here the example notebook we can use for tests https://github.com/SpatioTemporal/STARE-Applications/blob/main/09-STARE-PODS-IO-1.ipynb
Question concerning podding w/ STAREPandas. I finally got a generalized version of the VIIRS podding process (09-STARE-PODS-Sketch-2.ipynb) working, with the intent to us it on the IMERG PFeatures data.
As a check I got the test-podding data from FlexFS and ran the code on it. Which worked exactly as expected until it got to the pod and save call:
pods_written = sdf.write_pods(pod_root, level, chunk_name, temporal_chunking=temporal_chunking)
which throws and error as the main branch of STAREPandas doesn't know about temporal_chunking
. I see something similar in the HPC
branch, but I can't switch to that via git. I guess because the pull-request wasn't approved?
If I simply remove that argument, I still get an error from pandas.groupby:
ValueError: Grouper for '<class 'starepandas.staredataframe.STAREDataFrame'>' not 1-dimensional
Is this because the STARE dataframe is somehow incorrect?
sdf <class 'starepandas.staredataframe.STAREDataFrame'>
Data columns (total 9 columns):
# Column Dtype
--- ------ -----
0 lat float32
1 lon float32
2 sids Int64
3 ts_start datetime64[ns]
4 ts_end datetime64[ns]
5 I04_observations float32
6 I04_quality_flags UInt16
7 I05_observations float32
8 I05_quality_flags UInt16
lat lon sids ts_start ts_end I04_observations I04_quality_flags I05_observations I05_quality_flags
-- ------- -------- ------------------- ------------------- ------------------- ------------------ ------------------- ------------------ -------------------
0 4.46929 -119.84 3382414711004431630 2021-12-31 22:18:00 2021-12-31 22:24:00 65533 256 65533 256
1 4.46831 -119.848 3382696059138709358 2021-12-31 22:18:00 2021-12-31 22:24:00 65533 256 65533 256
2 4.46733 -119.855 3382696072051950574 2021-12-31 22:18:00 2021-12-31 22:24:00 65533 256 65533 256
3 4.46635 -119.862 3382696052710401102 2021-12-31 22:18:00 2021-12-31 22:24:00 65533 256 65533 256
...
[41369600 rows x 9 columns]
memory usage: 1.8 GB
I'll go back to working on the Snow data stuff in the meantime, which is coming along fine.
Mike
Oh! That makes a lot of sense. Probably forces the interpreter to find the min value. How much speedup do you get? I will take care of the conflicts and merge this tomorrow