BDTopo and OSM workflow now return the name of the computed tables.

[ebocher]

The method getResults() on the OSM and BDTopo workflow returns the name of the computed tables and stored in the local database.
The input parameter has been improved. So the user can send a configuration file or a map to the workflow.

 def osm_parmeters = [
                "description" :"Example of configuration file to run the grid indicators",
                "geoclimatedb" : [
                        "folder" : "/tmp/geoclimate",
                        "name" : "geoclimate_chain_db;AUTO_SERVER=TRUE",
                        "delete" :false
                ],
                "input" : [
                        "osm" : ["Paimpol"]],
                "output" :[
                        "folder" : ["path":  "/tmp/geoclimate",
                                    "tables": ["grid_indicators", "zones", "rsu_lcz"]]],
                "parameters":
                        ["distance" : 0,
                         "grid_indicators": [
                                 "x_size": 1000,
                                 "y_size": 1000,
                                 "indicators": ["LCZ_PRIMARY"]
                         ]
                        ]
        ]
IProcess process = OSM.WorkflowOSM.workflow()
process.execute(input: osm_parmeters)
process.getResults().output["Paimpol"]

will return a map with a key = Paimpol and value = name of tables

If the local H2GIS is not deleted after the workflow chain, the user can process the tables stored in the database.

[ebocher]

Since last H2 update with some errors

Timeout trying to lock table "SURFACE_ACTIVITE"; SQL statement:

[ebocher]

Dear @katzyn
Since the last update on H2 we have some lock issue.
I'm not able to find where it appears.
The H2 database is build with this path "./target/${RsuIndicatorsTests.simpleName}_db;AUTO_SERVER=TRUE"

[katzyn]

Do you have a complete stack trace?

[ebocher]

Not yet a clear stack trace.
I'm on it

[katzyn]

H2 uses a very small lock timeout by default, usually it needs to be increased for applications with concurrent sessions and transactions.

But timeout may also indicate a some kind of deadlock.

[ebocher]

has the lock timeout changed recently ?

[katzyn]

No.

[katzyn]

Did it work in H2 2.0.x?

[ebocher]

This code was working before 2.1.210 with 2.0.206.
I'm try to find a unit test but it's a huge sql chain ;((

[ebocher]

I modified the tests to create a new database each time because it was impossible to isolate a unit test

[ebocher]

@katzyn
The issue is fixed. It was a bad used of setAutocommit on H2GIS side.

[ebocher]

@katzyn
The issue is fixed. It was a bad used of setAutocommit on H2GIS side.