I351 list results

[fcooper8472]

New behaviour

smif list

smif list will now append an asterisk to model run names if full results exist

smif available_results <run_name>

smif available_results <run_name> will detail the results available for a single <run_name>: for instance (with results):

model run: energy_water_cp_cr
  - sos model: energy_water
    - sector model: water_supply
      - output: cost
        - decision 0: 2010, 2015, 2020
      - output: energy_demand
        - decision 0: 2010, 2015, 2020
      - output: water
        - decision 0: 2010, 2015, 2020
      - output: reservoir_level
        - decision 0: 2010, 2015, 2020
    - sector model: energy_demand
      - output: cost
        - decision 0: 2010, 2015, 2020
      - output: water_demand
        - decision 0: 2010, 2015, 2020

And, without results:

model run: energy_water_cp_cr
  - sos model: energy_water
    - sector model: water_supply
      - output: cost
        - no results
      - output: energy_demand
        - no results
      - output: water
        - no results
      - output: reservoir_level
        - no results
    - sector model: energy_demand
      - output: cost
        - no results
      - output: water_demand
        - no results

Omitting the <run_name> will print all information for every model run.

Closes #351 and #352

[willu47]

Great, so I've had a play against NISMOD2 with this, and discovered a few things!

1. smif list is now very slow. Each model run takes about 3 second, less if it doesn't have any results. This may well be our config code, or disk io? Not sure. I think the quick fix is to make the results functionality optional on the smif list command, and if necessary look into optimisation at a later date. Not a priority at present.
2. Results read out for smif available_results is really good. I don't think I would ever use it for every single model run as it just fills the screen, but the ability to see this for one model run is handy. I think this meets our requirements for now.

Example for digital comms:

(nismod) vagrant@vagrant:/vagrant$ smif available_results digital_comms_test

model run: digital_comms_test
  - sos model: digital_comms_fixed_only
    - sector model: digital_comms_fixed_network
/vagrant/smif/src/smif/data_layer/file/file_metadata_store.py:159: FionaDeprecationWarning: Use fiona.Env() instead.
  with fiona.drivers():
      - output: adoption_cap
        - no results
      - output: distribution_upgrades
        - no results
      - output: distribution_upgrade_costs_fttp
        - no results
      - output: premises_by_distribution
        - no results
      - output: interventions
        - no results
      - output: distribution_upgrades
        - no results
      - output: distribution_upgrade_costs_fttp
        - no results
      - output: distribution_upgrade_costs_fttdp
        - no results
      - output: premises_upgrade_costs_fttp
        - no results
      - output: premises_upgrade_costs_fttdp
        - no results
      - output: premises_by_distribution
        - no results
      - output: premises_adoption_desirability
        - no results
      - output: premises_connection_statistics
        - no results
      - output: premises_passed_with_fttp
        - no results
      - output: percentage_of_premises_passed_with_fttp
        - no results
      - output: premises_connected_with_fttp
        - no results
      - output: percentage_of_premises_connected_with_fttp
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: premises_passed_with_fttdp
        - no results
      - output: percentage_of_premises_passed_with_fttdp
        - no results
      - output: premises_connected_with_fttdp
        - no results
      - output: percentage_of_premises_connected_with_fttdp
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: percentage_of_premises_with_fttc
        - no results
      - output: lad_premises_with_fttp
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: lad_premises_with_fttdp
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: lad_premises_with_fttc
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: lad_premises_with_adsl
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: coverage
        - no results
      - output: percentage_of_premises_connected_with_fttc
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: percentage_of_premises_connected_with_adsl
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: percentage_of_premises_connected_with_docsis3
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: total_cost
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: total_potential_bcr
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: total_potential_benefit
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: rollout_costs
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020
      - output: rollout_bcr
        - decision 1: 2019
        - decision 2: 2019
        - decision 3: 2020

[codecov]

Codecov Report

Merging #358 into develop will increase coverage by 0.58%.
The diff coverage is 100%.

@@             Coverage Diff             @@
##           develop     #358      +/-   ##
===========================================
+ Coverage    70.59%   71.18%   +0.58%     
===========================================
  Files           59       59              
  Lines         5183     5292     +109     
  Branches       615      647      +32     
===========================================
+ Hits          3659     3767     +108     
+ Misses        1431     1430       -1     
- Partials        93       95       +2

Flag	Coverage Δ
#javascript	71.18% <100%> (+0.58%)	⬆️
#python	78.99% <100%> (+0.5%)	⬆️

Impacted Files	Coverage Δ
src/smif/data_layer/store.py	95.23% <100%> (+0.38%)	⬆️
src/smif/data_layer/file/file_data_store.py	87.61% <0%> (+0.53%)	⬆️
src/smif/decision/decision.py	95.15% <0%> (+1.93%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b303134...40ce640. Read the comment docs.

[fcooper8472]

@willu47 I have made the complete results checking optional: it's now an additional flag smif list <-c/--complete>.

Would be interesting to what is actually slowing up the smif list, but unfortunately I've not been able to get my dev version of smif working with nismod2...

I suspect the slowdown must be in _get_canonical_expected_results rather than _get_canonical_available_results but it only seems to be reading some YAML config - any ideas?

[tomalrussell]

I'm happy to ignore appveyor/branch failures here where they relate to conda-forge or dependency errors, so long as appveyor/pr and travis get ✔️s.

[willu47]

A few optional suggestions for refactoring to help enable #359. Would be good to also get some tests in place against the helper functions.

[fcooper8472]

The new functionality builds the list of available results based on calls to store.canonical_expected_results and store.available_results. This removes some code duplication and should also improve performance.

All methods (including new cli functionality) are now unit tested.

Ready for another review @willu47

[willu47]

That all looks great Fergus. I've tested it against the nismod2 repo and it works fine. Thanks a lot!