Contingency plan code updates

[lmz]

Changes (via @drewlevitt - https://github.com/BayAreaMetro/bayarea_urbansim/pull/343#issuecomment-1839348474)

• Ignore any CSV whose file name contains "CoStar"
• Create output directories as needed if they don't already exist (previously the model would crash if required directories didn't exist)
• Change initial_summary_year from 2015 to 2020
• Rename development pipeline input filename
• In subsidies.py, add a guard clause so affordable housing preservation is not applied prior to 2025
• In core_summaries.py, use try/except logic to avoid a crash when interim_zone_output_all does not yet exist
• In geographic_summaries.py, comment out a guard clause so we get geographic summaries for all years (useful for validation Tableau)
• In travel_model_summaries.py, comment out a guard clause so we get taz1 summaries for all years

• Add PBA50+ validation Tableau workbook (through contingency plan run v7: housing preservation activated in 2025) and supporting Python preprocessing scripts and notebooks

• To see the specific tasks where the Asana app for GitHub is being used, see below:
  • https://app.asana.com/0/0/1206060729501212

[drewlevitt]

Changes:

• Ignore any CSV whose file name contains "CoStar"
• Create output directories as needed if they don't already exist (previously the model would crash if required directories didn't exist)
• Change initial_summary_year from 2015 to 2020
• Rename development pipeline input filename
• In subsidies.py, add a guard clause so affordable housing preservation is not applied prior to 2025
• In core_summaries.py, use try/except logic to avoid a crash when interim_zone_output_all does not yet exist
• In geographic_summaries.py, comment out a guard clause so we get geographic summaries for all years (useful for validation Tableau)
• In travel_model_summaries.py, comment out a guard clause so we get taz1 summaries for all years
• Add PBA50+ validation Tableau workbook (through contingency plan run v7: housing preservation activated in 2025) and supporting Python preprocessing scripts and notebooks

[lmz]

Copied your notes into description; thank you, @drewlevitt !

[theocharides]

Adding a few quick comments here:

• @akselx added some code here to auto-create output folders, so I think the code in this PR importing pathlib and modifying the output folders can be removed since it's covered
• I asked Nazanin to put in that year < 2025 line as a test but I think we can prevent hardcoding the year with this: https://github.com/BayAreaMetro/bayarea_urbansim/pull/347/commits/97ea86524894194ba5b3fad6214bd5cd68f0ebea#diff-8e32919ed5dab8c6af82aeeb375599e2981a9ac4427a7132bf89d899c7499616R58
• Because I started on the PR previously I think I ran into an error with the necessary change to this and I ended up changing it to just this-- but since I can't remember what the error was (oy!) I can also just address that later if it pops up.
• I think one element of our inputs reconfiguration was also to genericize input files in the code and only swap out input files themselves. So updating the development projects file input in this PR may be our first opportunity to implement that- woohoo! From discussion with @akselx I think we would change the datasources.py input to read in a generic variable called something like development_projects and then in run_setup.yaml we could define that variable with the real file name e.g., development_pipeline_with_basis_buildings_nov2023.

[drewlevitt]

Thanks @theocharides !

• I reproduced your not-hardcoded logic: https://github.com/BayAreaMetro/bayarea_urbansim/pull/343/commits/bec441cf15b3cb2e5f7910928e318c439a088f85
• I think the try/except logic around interim_zone_output should work OK, but let's keep an eye on it
• I will go through and make sure that only one of Lisa's and Aksel's "create any missing directories" approaches is in place. Kinda feels like the most comprehensive way to do this is via merging in the branch, then going through the merged files and editing them via a commit to main afterwards. @lmz does that sound OK?
• Looks like Aksel's recent commit does include a genericized approach to grabbing the HE-to-pipeline data :)