InnoZ / MAS_Interface

Apache License 2.0
1 stars 0 forks source link

Test of modelling queue #34

Closed 00Bock closed 6 years ago

00Bock commented 7 years ago

Now that we have the DB structure, can we test the modeling queue? Best with the following Scenarios:

BenSto commented 7 years ago

The Queue is not enabled at the moment as the sidekiq branch is still not merged because it would add another dependency (redis-server). But we can start the queue testing as soon as we are able to simulate a real queue.

00Bock commented 7 years ago

with simulate a real queue you mean something like generate 2 'nigel'-models?

00Bock commented 7 years ago

The benefits of queuing both models soon, is that we a) have something to present, b) we can potentially use to design scenario comparisons with eGAP and 3connect-models and c) can potentially monitor the quality of automation in comparison to Hosse's calibrated models from eGAP and 3connect. The sooner we can document quality levels the better. This last benefit is something which can also derive from the MAS repository. @dhosse how much of an task would that be to run before Florence and should we make an seperate MAS-issue from that? I guess that would also make sense since we need to measure the deltas from traffic volumes between scenarios at counting stations and also deltas from overall modal-split.

wese-da commented 7 years ago

I finally managed to create a dump of the plans of the eGAP and 3connect base scenarios. They got pretty big (~ 145 MB), so I put them on _GISDATA/MODELLINGDATA/MASInterface/seeds.dump.

BenSto commented 7 years ago

Just for Documentation: It would be very usefull when the dump also writes entries in the scenario table (seed: true) and also uses correct identifier for the scenario_id (combination of district_id and year)

BenSto commented 7 years ago

What is missing at the moment: the scenarios table is empty and also the grid. Scenarios table will be filled by MATSim I guess. I think grid creation can be done in the seed procedure easily.

00Bock commented 7 years ago

uses correct identifier for the scenario_id (combination of district_id and year)

I think we still need to add a undefined name part of the scenario table 'scenario_name', which can be changed according to the user's preference. E.g.: district_id='23986' year='2017' scenario_name='autonome_shuttles'

scenario_id= 23986_2017_autonome_shuttles