Hana,
Thank you for your help, the model works very well. I did two things, update
the events_coordinator model and change the type of some variables in
gridcells table, like residential_units which was of type smallint(6).
Now, I have an other question, I would like to know what is the equivalent
of control_totals in this model. Is it the target_vacancy?
Thanks,
Hakim
-----Message d'origine-----
De : users-bounces_at_urbansim.org [mailto:users-bounces_at_urbansim.org] De la
part de Hana Sevcikova
Envoyé : mardi 20 novembre 2007 19:56
À : hakim.ouaras_at_u-cergy.fr
Cc : users_at_urbansim.org
Objet : Re: [UrbanSim-Users] Problem in DPLCM results
Hakim,
I'm not sure what version of urbansim you have, but I suspect from the error
log that it is pre-August 07? I'm mentioning it because end of July 07 we
made substantial internal changes to the events_coordinator model. It didn't
change its behaviour but it extended the type of events that can be
processed. Because of these changes it is very difficult for me now to tell
you what is going wrong in your application.
If you don't want to update opus right now, you could try to update just the
events_coordinator model, since the UI hasn't change. Just replace the
urbansim/models/events_coordinator.py file with the most recent one from the
repository. The unit test probably won't work, but if you replace them by
the unit test from the old events_coordinator.py file, it should work.
(Please make copies of everything you're replacing, in case I'm wrong.)
Let me know if this helps.
Hana
Hakim OUARAS wrote:
> Hana,
>
> It is ok for Eugene, in fact I haven't ran the
> development_event_transition_model.
>
> For Paris database, when I ran also the
> development_event_transition_model
> and events_coordinator I have the error message in attached file
> (error_log).
>
> Note that, in our case we use only one development model (Development
> Project activity Location Choice Model) which replace the industrial
> and commercial models.
>
> Thanks,
> Hakim
>
>
>
> -----Message d'origine-----
> De : users-bounces_at_urbansim.org [mailto:users-bounces_at_urbansim.org] De
> la part de Hana Sevcikova Envoyé : lundi 19 novembre 2007 18:25 À :
> hakim.ouaras_at_u-cergy.fr Cc : users_at_urbansim.org Objet : Re:
> [UrbanSim-Users] Problem in DPLCM results
>
> Hakim,
>
> The gridcell attributes are updated by the events_coordinator model,
> in combination with the development_event_transition_model. Do you run
> these two models? And did you run them in your experiment with the Eugene
data?
>
> FYI, one event is a collection of projects to be applied to one gridcell.
> The development_event_transition_model creates those events from the
> results of the development_project_... models and the
> events_coordinator model 'builds' them, i.e. applies them to gridcells.
>
> In case you do run those two models, could you please send a few log
> files and some indicator results where we could see the inconsistency?
> Also, please try to export the gridcells table into your database
> (with opus_core/tools/do_export_cache_to_sql_database.py, using the '-t'
> option) and check manually if e.g. residential_units haven't changed
> in comparison to the baseyear.
>
> Thanks,
>
> Hana
>
>
> Hakim OUARAS wrote:
>> Hi,
>>
>> We implemented a new development project model, it run without any
>> bug until the last year.
>> In the log file we noticed that there are some projects created, but
>> when we view the results with the indicators tool haven't any change
>> in the values of development variables, like residential_units.
>>
>> We observe the same problem when we ran Eugene database with only
>> development project models (note that we ran the development project
>> models with household (transition and location) models to have some
>> projects created).
>>
>> When we ran all models with Eugene database we haven't this problem,
>> there are evolution in development variable values. In the Paris
>> database we have always the problem even if we run all models.
>>
>> We think that the projects are created but the are note located, but
>> why does not work in Eugene!
>>
>> Paris team
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users_at_urbansim.org
>> http://www.urbansim.org/mailman/listinfo/users
>
> _______________________________________________
> Users mailing list
> Users_at_urbansim.org
> http://www.urbansim.org/mailman/listinfo/users
>
_______________________________________________
Users mailing list
Users_at_urbansim.org
http://www.urbansim.org/mailman/listinfo/users
Received on Wed Nov 21 2007 - 07:37:56 PST
This archive was generated by hypermail 2.2.0 : Wed Nov 21 2007 - 07:37:58 PST