ldesousa / centaur.loc Goto Github PK

This entry point should recieve as arguments:

• path to the SWMM file
• name of the desired database schema

Section headers can appear in different forms, e.g. [Polygons] and [POLYGONS]

If the method receives an incomplete set of values it must issue an error.

After the creation of the f_optimal stored procedure, the methods getBestCandidateVolumeArea() and getBestCandidateVolumeAreaNumSubcatchments() are now almost identical - merge them.

Considering this case: the two highest ranking candidates by volume*area serve the same subcatchment(s).

There are at least two options to deal with this:

1. Simply remove from the search space candidates serving subcatchments that are already controlled by previously assigned gates.
2. Use a storage threshold above which a subcatchment is no longer used in the calculation.

. Install postgres + postgis
. Create user and database
. Superuser needed to create Postgis extension

The CreateDBSchema class is creating a temporary file in /tmp, but Windows does not like it. Create the file on the execution folder and then delete it.

This fix must be released as hot-fix: micro-version.

This comes down to running the right contents of the SQL script 0300_spatial_enablement. It might be better to divide this script in two, with the enablement part run by the createDBSchema class and the object creation part run by the ImportSWMM class.

The CreateDBSchema class uses psql to execute the SQL script that creates all the necessary objects and functions to run centaur.loc. However, on Windows users are reporting difficulties with this (possibly psql is not available to the JVM.

There are a few alternative methods to run SQL scripts listed in this thread:
https://stackoverflow.com/questions/2071682/how-to-execute-sql-script-file-in-java

As it stands, the ImportSWMM class fails in case there are no polygons in the .inp file. However, all the information needed is present in the regular tables (area). The create_polygons() function must exit orderly if it does not find polygons.

The theoritical storage calculation must take into account into bifurcations upstream. A node becomes a stopping point in the search algorithm if has more than outlet conduit.

When the ImportSWMM class is run on an existing database, it starts by clearing all existing objects (ClearDB() method). However, at present the command to delete Weir objects is failing, and thus an exception is issued when it tries to delete Link objects

Two sub tasks:
. merge v_candidate_volume with v_candidate
. understand with count(*) is different between v_candidate and candidate

Even though these have no actual effect on the model workings, orifices often appear as conduit end points, i.e. a sub-category of node.

Some contributions fields are turning out with negative values. Investigate why.

If the main() method invokes the two routines in succession:

    FloodedSegments.compute(session);
    commitData(session, tx); 

    OptimalByVolumeArea.compute(session, 5, 1, 60);
    commitData(session, tx);

The candidates view is still empty when the gate sitting algorithm executes. Is the commitData() delayed or failing?

There should be a separate entry point to compute flooded segments and a second one to locate the gates.

Right now it is associated with Links in general.

This comes down to forcing an update to the materialised view v_candidate_upstream

All alphanumeric identifiers must be inserted into the name attribute, with a serial used for the Entity's id. The random number generator must be scrapped.