469 update retrofit dataset

[longshuicy]

add additional function to update the existing inventory dataset with additional retrofit and mapping entry key information. https://github.com/IN-CORE/pyincore/pull/470/files#diff-d1b1cb98d647ccb4ebd9f558b2894472d298769fdd3c2bde8d9d4e00205e4de0R73

Since we use fiona object and looping through fiona is too slow, I used geopandas dataframe to merge and udpate; but then save to shapefile so we can resume using fiona to carry out normal fragility matching and calculation

---

To test:

• Use this notebook to test SLC retrofit strategy
  SLC_retrofit_test.zip
• Use this notebook to tests Galveston
  Galveston_retrofit_test.zip
• RUN tests/pyincore/analyses/buildingdamage/test_buildingdamage_retrofit.py and see results in the retrofit folder

Note:

• Joplin takes 31 seconds;
• Lumberton takes 61 seconds.
• SLC if only retrofit 5000+ building with inner join total runtime: 158 seconds;
• Galveston retrofitting 1500 buildings combined damage takes 115 seconds

[longshuicy]

Made one update to use inner join when joining with retrofit strategies and mapping; this will make sure only the building that has retrofit strategy will get computed again. Hence help with the performance.

[ylyangtw]

Encountered FutureWarning in Galveston notebook:

/pyincore/utils/dataprocessutil.py:281: FutureWarning: The behavior of DataFrame.idxmax with all-NA values, or any-NA and skipna=False, is deprecated. In a future version this will raise ValueError
  max_key = dmg_result[dmg_states].idxmax(axis=1)

Just committed a fix if that's ok. Feel free to revert it if it affects anything.
Other than this, both notebooks and test work.

Some test records from my end:
Joplin Tornado Retrofit execution time: 11.40898 seconds
Lumberton Flood Retrofit execution time: 44.84685 seconds
Galveston notebook runtime: 71 & 20
SLC notebook: 162 & 52

Good job on the hardwork!

[longshuicy]

Encountered FutureWarning in Galveston notebook:

/pyincore/utils/dataprocessutil.py:281: FutureWarning: The behavior of DataFrame.idxmax with all-NA values, or any-NA and skipna=False, is deprecated. In a future version this will raise ValueError
  max_key = dmg_result[dmg_states].idxmax(axis=1)

Just committed a fix if that's ok. Feel free to revert it if it affects anything. Other than this, both notebooks and test work.

Some test records from my end: Joplin Tornado Retrofit execution time: 11.40898 seconds Lumberton Flood Retrofit execution time: 44.84685 seconds Galveston notebook runtime: 71 & 20 SLC notebook: 162 & 52

Good job on the hardwork!

Thanks for the fix and thorough test!