open-insight/LSL2/STPROC/MATH_SERVICES.txt
Infineon\StieberD 7762b129af pre cutover push
2024-09-04 20:33:41 -07:00

123 lines
5.9 KiB
Plaintext

Compile function Math_Services(@Service, @Params)
/***********************************************************************************************************************
Name : Math_Services
Description : Handler program for all math services.
Notes : Application errors should be logged using the Error Services module. There are a few methodological
assumptions built into way errors are managed which are important to understand in order to properly
work with Error Services:
- The term 'top' refers to the originating procedure of a call stack and the term 'bottom' refers to
the last routine (or the current routine) within a call stack. Within the OpenInsight Debugger
this will appear backwards since the originating procedure always appears at the bottom of the
list and the current routine appears at the top of the list. We are using this orientation because
it is common to refer to the process of calling other procedures as 'drilling down'.
- The reason for defining the orientation of the call stack is because Error_Services allows for
multiple error conditions to be appended to an original error. In most cases this will happen when
a procedure at the bottom of the stack generates an error condition and then returns to its
calling procedure. This higher level procedure can optionally add more information relevant to
itself. This continues as the call stack 'bubbles' its way back to the top to where the
originating procedure is waiting.
- Native OpenInsight commands that handle errors (e.g., Set_Status, Set_FSError, Set_EventStatus)
preserve their error state until explicitly cleared. This can hinder the normal execution of code
since subsequent procedures (usually SSPs) will fail if a pre-existing error condition exists.
Our philosophy is that error conditions should automatically be cleared before a new procedure
is executed to avoid this problem. However, the nature of Basic+ does not make this easy to
automate for any given stored procedure. Therefore, if a stored procedure wants to conform to our
philosophy then it should include a call into the 'Clear' service request at the top of the
program. Alternatively this can be done through a common insert (see SERVICE_SETUP for example.)
- Service modules will use the SERVICE_SETUP insert and therefore automatically clear out any
error conditions that were set before.
Parameters :
Service [in] -- Name of the service being requested
Param1-10 [in/out] -- Additional request parameter holders
Response [out] -- Response to be sent back to the Controller (MCP) or requesting procedure
Metadata :
History : (Date, Initials, Notes)
11/12/20 djs Original programmer.
***********************************************************************************************************************/
#pragma precomp SRP_PreCompiler
$insert LOGICAL
$insert SERVICE_SETUP
Declare function SRP_Array, Math_Services
GoToService else
Error_Services('Set', Service : ' is not a valid service request within the ' : ServiceModule : ' services module.')
end
Return Response or ""
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Service Parameter Options
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Options STD_DEV_TYPES = 'SAMPLE','POPULATION'
//-----------------------------------------------------------------------------
// SERVICES
//-----------------------------------------------------------------------------
Service GetStdDev(ValueList, StdDevType=STD_DEV_TYPES)
StdDev = 0
If ValueList NE '' then
If StdDevType EQ '' then StdDevType = 'SAMPLE'
Avg = 0
Total = 0
Deviations = ''
ValueList = SRP_Array('Clean', ValueList, 'Trim', @VM)
ValueCount = DCount(ValueList, @VM)
If ValueCount GT 0 then
For each Value in ValueList using @VM setting vPos
Total += Value
Next Value
Avg = Total / ValueCount
// Calculate Std Dev
For each Value in ValueList using @VM setting vPos
Deviation = Value - Avg
Deviations<0, vPos> = Deviation ** 2
Next Value
DevsSum = Sum(Deviations)
If StdDevType EQ 'SAMPLE' then
// Std Dev for a sample, not population
If (ValueCount - 1) GT 0 then
Variance = DevsSum / (ValueCount - 1)
StdDev = Variance ** 0.5
end
end else
// Std Dev for a population
Variance = DevsSum / ValueCount
StdDev = Variance ** 0.5
end
Response = StdDev
end
end
End Service
Service GetCoefficientOfVariation(ValueList, StdDevType=STD_DEV_TYPES)
ValueList = SRP_Array('Clean', ValueList, 'Trim', @VM)
If ValueList NE '' then
StdDev = Math_Services('GetStdDev', ValueList, StdDevType)
NumValues = DCount(ValueList, @VM)
Total = Sum(ValueList)
Avg = Total / NumValues
CV = StdDev / Avg
Response = CV
end
end service