Skip to content

feat(neuron-writer): use Meta to allow use of cvode #105

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Sep 8, 2023
Merged

feat(neuron-writer): use Meta to allow use of cvode #105

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
a62e2bf
feat(neuron-writer): use `Meta` to allow use of cvode
sanjayankur31 Aug 23, 2023
98dd808
fix(neuron-writer): set default `dt` even when cvode is in use
sanjayankur31 Aug 23, 2023
5e52dae
feat(neuron-writer): include both cvode and standard integration
sanjayankur31 Aug 24, 2023
fd5f50e
fix(neuron-writer): don't resize recorder vector when using cvode
sanjayankur31 Aug 25, 2023
b676a16
fix(neuron-writer): do not resize recording vector when using cvode
sanjayankur31 Aug 25, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
69 changes: 56 additions & 13 deletions src/main/java/org/neuroml/export/neuron/NeuronWriter.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,7 @@
import org.lemsml.jlems.core.type.Property;
import org.lemsml.jlems.core.type.Requirement;
import org.lemsml.jlems.core.type.Target;
import org.lemsml.jlems.core.type.Meta;
import org.lemsml.jlems.core.type.dynamics.Case;
import org.lemsml.jlems.core.type.dynamics.ConditionalDerivedVariable;
import org.lemsml.jlems.core.type.dynamics.DerivedVariable;
Expand Down Expand Up @@ -329,7 +330,6 @@ public String getMainScript() throws GenerationException, NeuroMLException


Target target = lems.getTarget();

Component simCpt = target.getComponent();

String len = simCpt.getStringValue("length");
Expand All @@ -340,28 +340,58 @@ public String getMainScript() throws GenerationException, NeuroMLException
len = "" + Float.parseFloat(len) * 1000;
}

String dt = simCpt.getStringValue("step");
dt = dt.replaceAll("ms", "").trim();
if(dt.indexOf("s") > 0)
/* cvode usage:
* https://nrn.readthedocs.io/en/latest/hoc/simctrl/cvode.html
* - we do not currently support the local variable time step method
*/
boolean nrn_cvode = false;
String dt = "0.01";
/* defaults from NEURON */
String abs_tol = "None";
String rel_tol = "None";
LemsCollection<Meta> metas = simCpt.metas;
for(Meta m : metas)
{
dt = dt.replaceAll("s", "").trim();
dt = "" + Float.parseFloat(dt) * 1000;
HashMap<String, String> attributes = m.getAttributes();
if (attributes.getOrDefault("for", "").equals("neuron"))
{
if (attributes.getOrDefault("method", "").equals("cvode"))
{
nrn_cvode = true;
abs_tol = attributes.getOrDefault("abs_tolerance", abs_tol);
rel_tol = attributes.getOrDefault("rel_tolerance", rel_tol);
E.info("CVode with abs_tol="+abs_tol+" , rel_tol="+rel_tol+" selected for NEURON simulation");
}
}

}

if (nrn_cvode == false)
{
dt = simCpt.getStringValue("step");
dt = dt.replaceAll("ms", "").trim();
if(dt.indexOf("s") > 0)
{
dt = dt.replaceAll("s", "").trim();
dt = "" + Float.parseFloat(dt) * 1000;
}
}

main.append("class NeuronSimulation():\n\n");
int seed = DLemsWriter.DEFAULT_SEED;
if (simCpt.hasStringValue("seed"))
seed = Integer.parseInt(simCpt.getStringValue("seed"));

main.append(" def __init__(self, tstop, dt, seed="+seed+"):\n\n");
main.append(" def __init__(self, tstop, dt=None, seed="+seed+", abs_tol=None, rel_tol=None):\n\n");


Component targetComp = simCpt.getRefComponents().get("target");

main.append(bIndent+"print(\"\\n Starting simulation in NEURON of %sms generated from NeuroML2 model...\\n\"%tstop)\n\n");
main.append(bIndent+"self.setup_start = time.time()\n");
main.append(bIndent+"self.seed = seed\n");
main.append(bIndent+"self.abs_tol = abs_tol\n");
main.append(bIndent+"self.rel_tol = rel_tol\n");

if (target.reportFile!=null)
{
Expand Down Expand Up @@ -1306,8 +1336,14 @@ else if(cc.getComponentType().isOrExtends(NeuroMLElements.CONTINUOUS_CONNECTION_
main.append(toRec);

main.append(bIndent+"h.tstop = tstop\n\n");
main.append(bIndent+"h.dt = dt\n\n");
main.append(bIndent+"h.steps_per_ms = 1/h.dt\n\n");
main.append(bIndent+"if self.abs_tol is not None and self.rel_tol is not None:\n");
main.append(bIndent+" cvode = h.CVode()\n");
main.append(bIndent+" cvode.active(1)\n");
main.append(bIndent+" cvode.atol(self.abs_tol)\n");
main.append(bIndent+" cvode.rtol(self.rel_tol)\n");
main.append(bIndent+"else:\n");
main.append(bIndent+" h.dt = dt\n");
main.append(bIndent+" h.steps_per_ms = 1/h.dt\n\n");

if(!nogui)
{
Expand Down Expand Up @@ -1350,7 +1386,8 @@ else if(cc.getComponentType().isOrExtends(NeuroMLElements.CONTINUOUS_CONNECTION_
columnsPre.get(timeRef).add(bIndent+"h(' objectvar v_" + timeRef + " ')");
columnsPre.get(timeRef).add(bIndent+"h(' { v_" + timeRef + " = new Vector() } ')");
columnsPre.get(timeRef).add(bIndent+"h(' { v_" + timeRef + ".record(&t) } ')");
columnsPre.get(timeRef).add(bIndent+"h.v_" + timeRef + ".resize((h.tstop * h.steps_per_ms) + 1)");
columnsPre.get(timeRef).add(bIndent+"if self.abs_tol is None or self.rel_tol is None:\n");
columnsPre.get(timeRef).add(bIndent+" h.v_" + timeRef + ".resize((h.tstop * h.steps_per_ms) + 1)");

columnsPost0.get(timeRef).add(bIndent+"py_v_" + timeRef + " = [ t/1000 for t in h.v_" + timeRef + ".to_python() ] # Convert to Python list for speed...");

Expand Down Expand Up @@ -1410,7 +1447,8 @@ else if(cc.getComponentType().isOrExtends(NeuroMLElements.CONTINUOUS_CONNECTION_
columnsPre.get(outfileId).add(bIndent+"h(' objectvar v_" + colId + " ')");
columnsPre.get(outfileId).add(bIndent+"h(' { v_" + colId + " = new Vector() } ')");
columnsPre.get(outfileId).add(bIndent+"h(' { v_" + colId + ".record(&" + lqp.getNeuronVariableReference() + ") } ')");
columnsPre.get(outfileId).add(bIndent+"h.v_" + colId + ".resize((h.tstop * h.steps_per_ms) + 1)");
columnsPre.get(outfileId).add(bIndent+"if self.abs_tol is None or self.rel_tol is None:\n");
columnsPre.get(outfileId).add(bIndent+" h.v_" + colId + ".resize((h.tstop * h.steps_per_ms) + 1)");

float conv = NRNUtils.getNeuronUnitFactor(lqp.getDimension().getName());
String factor = (conv == 1) ? "" : " / " + conv;
Expand Down Expand Up @@ -1571,7 +1609,12 @@ else if (eofFormat.equals(EventWriter.FORMAT_ID_TIME))

main.append(bIndent+"self.initialized = True\n");
main.append(bIndent+"sim_start = time.time()\n");
main.append(bIndent+"print(\"Running a simulation of %sms (dt = %sms; seed=%s)\" % (h.tstop, h.dt, self.seed))\n\n");

main.append(bIndent+"if self.abs_tol is not None and self.rel_tol is not None:\n");
main.append(bIndent+" print(\"Running a simulation of %sms (cvode abs_tol = %sms, rel_tol = %sms; seed=%s)\" % (h.tstop, self.abs_tol, self.rel_tol, self.seed))\n");
main.append(bIndent+"else:\n");
main.append(bIndent+" print(\"Running a simulation of %sms (dt = %sms; seed=%s)\" % (h.tstop, h.dt, self.seed))\n\n");

main.append(bIndent+"try:\n");
main.append(bIndent+" h.run()\n");
main.append(bIndent+"except Exception as e:\n");
Expand Down Expand Up @@ -1709,7 +1752,7 @@ else if (eofFormat.equals(EventWriter.FORMAT_ID_TIME))

main.append("if __name__ == '__main__':\n\n");

main.append(" ns = NeuronSimulation(tstop="+len+", dt="+dt+", seed="+seed+")\n\n");
main.append(" ns = NeuronSimulation(tstop="+len+", dt="+dt+", seed="+seed+", abs_tol="+abs_tol+", rel_tol="+rel_tol+")\n\n");

main.append(" ns.run()\n\n");

Expand Down