/// <summary>
/// Sets the Rneuron
/// </summary>
/// <param name="rneuron"></param>
public void SetRneuronWithInputs(Rneuron rneuron)
{
var rneuronId = rneuron.ID;
// add rneuron to cache
Rneurons.TryAdd(rneuronId, rneuron);
bool isFirstInput = true;
RlmInputValue lastInputValue = null;
int cnt = 0;
IComparer <RlmInputKey> distinctComparer = new RlmInputKeyDistinctComparer();
IComparer <RlmInputKey> linearComparer = new RlmInputKeyLinearComparer();
// build dynamic inputs
foreach (var i in rneuron.Input_Values_Reneurons)
{
RlmInputKey inputKey = new RlmInputKey()
{
Value = i.Value, InputNum = cnt, Type = i.InputType
};
inputKey.DoubleValue = (i.InputType == Enums.RlmInputType.Linear) ? Convert.ToDouble(i.Value) : 0D;
RlmInputValue inputVal = null;
if (!isFirstInput)
{
if (lastInputValue.RelatedInputs == null)
{
lastInputValue.RelatedInputs = new SortedList <RlmInputKey, RlmInputValue>(i.InputType == Enums.RlmInputType.Linear ? linearComparer : distinctComparer);
}
if (!lastInputValue.RelatedInputs.TryGetValue(inputKey, out inputVal))
{
inputVal = new RlmInputValue();
lastInputValue.RelatedInputs.Add(inputKey, inputVal);
}
lastInputValue = inputVal;
}
else
{
if (DynamicInputs == null)
{
DynamicInputs = new SortedList <RlmInputKey, RlmInputValue>(i.InputType == Enums.RlmInputType.Linear ? linearComparer : distinctComparer);
}
isFirstInput = false;
if (!DynamicInputs.TryGetValue(inputKey, out inputVal))
{
inputVal = new RlmInputValue();
DynamicInputs.Add(inputKey, inputVal);
}
lastInputValue = inputVal;
}
cnt++;
}
lastInputValue.RneuronId = rneuronId;
}
/// <summary>
/// Gets existing Rneuron and creates a new one if not existing
/// </summary>
/// <param name="inputs">Inputs with value</param>
/// <param name="rnetworkID">Current NetworkId</param>
/// <returns></returns>
public GetRneuronResult GetRneuronFromInputs(IEnumerable <RlmIOWithValue> inputs, long rnetworkID)
{
GetRneuronResult retVal = new GetRneuronResult();
Rneuron rneuron = null;
// generate key based on input values
long rneuronId = Util.GenerateHashKey(inputs.Select(a => a.Value).ToArray());
// create new rneuron if not exists
if (!Rneurons.TryGetValue(rneuronId, out rneuron))
{
rneuron = new Rneuron()
{
ID = rneuronId, Rnetwork_ID = rnetworkID
};
bool isFirstInput = true;
RlmInputValue lastInputValue = null;
int cnt = 0;
IComparer <RlmInputKey> distinctComparer = new RlmInputKeyDistinctComparer();
IComparer <RlmInputKey> linearComparer = new RlmInputKeyLinearComparer();
foreach (var i in inputs)
{
// create IVR instance
var ivr = new Input_Values_Rneuron()
{
ID = Util.GenerateHashKey(rneuronId, i.ID),
Value = i.Value,
Input_ID = i.ID,
Rneuron_ID = rneuronId,
DotNetType = i.DotNetType,
InputType = i.Type
};
rneuron.Input_Values_Reneurons.Add(ivr);
RlmInputKey inputKey = new RlmInputKey()
{
Value = ivr.Value, InputNum = cnt, Type = i.Type
};
inputKey.DoubleValue = (i.Type == Enums.RlmInputType.Linear) ? Convert.ToDouble(ivr.Value) : 0D;
RlmInputValue inputVal = null;
if (!isFirstInput)
{
if (lastInputValue.RelatedInputs == null)
{
lastInputValue.RelatedInputs = new SortedList <RlmInputKey, RlmInputValue>(i.Type == Enums.RlmInputType.Linear ? linearComparer : distinctComparer);
}
if (!lastInputValue.RelatedInputs.TryGetValue(inputKey, out inputVal))
{
inputVal = new RlmInputValue();
lastInputValue.RelatedInputs.Add(inputKey, inputVal);
}
lastInputValue = inputVal;
}
else
{
if (DynamicInputs == null)
{
DynamicInputs = new SortedList <RlmInputKey, RlmInputValue>(i.Type == Enums.RlmInputType.Linear ? linearComparer : distinctComparer);
}
isFirstInput = false;
if (!DynamicInputs.TryGetValue(inputKey, out inputVal))
{
inputVal = new RlmInputValue();
DynamicInputs.Add(inputKey, inputVal);
}
lastInputValue = inputVal;
}
cnt++;
}
lastInputValue.RneuronId = rneuronId;
Rneurons.TryAdd(rneuronId, rneuron);
//rneuron_queue.Add(retVal);
//if (Rneurons.TryAdd(rneuronId, retVal))
//{
//}
//Rneurons2.Enqueue(retVal);
//rneuron_queue.Add(retVal);
retVal.Rneuron = rneuron;
retVal.ExistsInCache = false;
}
else
{
retVal.Rneuron = rneuron;
retVal.ExistsInCache = true;
}
return(retVal);
}