The OpenQuake 1.2 release is a major change with respect to the OpenQuake 1.0 release. You can find a detailed list of the improvements on this page.
A few new features were added and the performance of all calculators has been improved significatively. In some cases even of several orders of magnitude! Moreover the database name and schema have changed, the underlying dependencies have changed and a lot of bugs were fixed. Most of the changes have been backward compatible though.
You should not expect the numbers of an OpenQuake 1.2 computation to be exactly identical to the numbers of an OpenQuake 1.0 computation, because the algorithms or the underlying implementation may have changed, or because bugs may have been fixed. Moreover, you cannot use the hazard computed with OpenQuake 1.0 to run a risk calculation with OpenQuake 1.2: instead, you must recompute the hazard by using the original files.
OpenQuake 1.2 internally uses a new database called openquake2
and has no access to the
old database openquake
used by OpenQuake 1.0. The upgrade procedure will not
touch the old database at all and you can safely remove it, assuming you are not interested in
the old data.
Whereas the internal changes have been huge, the changes visible to the final users are few. I am listing them here:
-
The
openquake
command has been renamed tooq-engine
, and it has different options. You can see them with$ oq-engine --help
-
In particular the command-line options for export have been reduced. Before we had
--export-hazard-output
,--export-hazard-outputs
,--export-risk-output
,--export-risk-outputs
,--list-hazard-outputs
,--list-risk-outputs
; now we have--export-output
,--export-outputs
,--list-outputs
. The--export-type
option has been removed since it was somewhat redundant with the--exports
option. -
There is now a command-line option
$ oq-engine --make-html-report today
which generates a HTML file with performance information about the computations ran today; you can generate the report even for past days by giving an ISO date in the format YYYY-MM-DD.
-
We found a couples of errors in the XML source model files. Old files are now rejected if they contain duplicated IDs or if they contain invalid geometries.
-
The configuration file
openquake.cfg
has changed. There are some new parameters, some have been dropped and the name of the database has changed. If you upgrade to OpenQuake 1.2 you must manually change the fileopenquake.cfg
. You can see the format of the latest configuration file on GitHub. You may also need to change the PostgreSQL permissions in the file pg_hba.conf so that the database openquake2 can be accessed. -
The acceptable values of the parameter
calculation_mode
in the risk configuration files have changed; in particular whereas before the values wereclassical
,event_based
andscenario
, now you must writeclassical_risk
,event_based_risk
andscenario_risk
instead. The namesclassical
,event_based
andscenario
are now reserved to the hazard calculators. -
In the past, when doing a GMF calculation with correlation the syntax
ground_motion_correlation_params = {"vs30_clustering": false}
was accepted; now you must writeground_motion_correlation_params = {"vs30_clustering": False}
with capitalF
; the same fortrue
, now it must be writtenTrue
;0
and1
are accepted flags, as they were before. -
Now you cannot specify both
intensity_measure_types
andintensity_measure_types_and_levels
; if theintensity_measure_types_and_levels
are known (explicitly or by reading them from the risk models) you must remove the lineintensity_measure_types =
from thejob.ini
file, if it is there. -
The format to export the GMFs from a scenario computation has changed: now it is the same as for event based computations.