##
##
##  sarchitect designer 2.2 script for 'weighted averaging' of descriptors
##
##  Shaillay Kumar Dogra
##  editor@qsarworld.com
##  10 Oct 2006
##
##  Requires a column ('p-value') that is needed for weighted averaging of descriptors.
##  This 'p-value' column is the 'weights'.
##  Requires the user to provide the set of descriptors that need to be scaled (by p-value).
##  The remaining descriptors are left as is (and not scaled by any value). First value of the set
##  of conformers is taken as the representative value in the 'output' dataset.
##
##
##  Asks for following columns as inputs:
##	-> set of columns that need to be scaled,
##	-> column that has the values with which to scale (p-value),
##	-> conformer-ID column.
##
##
##  Ensure "Attribute Type" is 'categorical' for 'conformer-ID' column 
##  (else  "dataset[mseq_col].getRowIndicesOfCategory(catIdx)" will fail)
##
##
##  Output is a dataset (named "Averaged-Descriptors") 
##  containing unaveraged descriptors (for unselected columns) 
##  and weight-averaged desriptors (for selected columns). 
##
##
##


import script
from script.dataset import *
from script.algorithm import *
from script.project import *
from script.omega import createComponent, showDialog
from javax.swing import *
from math import *
from java.lang import Boolean
from com.strandgenomics.cube.dataset import ColumnFactory
from com.strandgenomics.chem.project import ChemProject




##-------------------------- 
def textarea(text):
	t = JTextArea(text)
	t.setBackground(JLabel().getBackground())
	return t

##------------------------------------------------
## PANEL TO GET DESCRIPTOR-COLUMN SELECTION
def getInputColumns():
	helptext="""
   To weight-average descriptors, select desired columns.
   Ensure selecting numerical columns only!
   Also, don't select the weights (p-value) column.
            """
	helpPanel=createComponent(type="ui",id="help", description="",component=textarea(helptext))
	colPanel = createComponent(type="ColumnSelector",id="columns", description="Select Columns",dataset=script.project.getActiveDataset())
	finalPanel= createComponent(type="group",id="Provide Descriptors for Averaging", description="Set of Descriptors",components=[helpPanel,colPanel])
	result=showDialog(finalPanel)
	return result['columns']

##------------------------------------------
## PANEL TO GET p-value COLUMN
def getPcolumn():
	helptext="""
	Provide the weights (p-value) column:
	   """
	helpPanel=createComponent(type="ui",id="help", description="",component=textarea(helptext))
	colPanel = createComponent(type="column", id="pvalue", description="Select p-value",dataset=script.project.getActiveDataset())
	finalPanel= createComponent(type="group",id="Provide p-value column", description="p-value column?",components=[helpPanel,colPanel])
	result=showDialog(finalPanel)
	return result['pvalue']

##------------------------------------------
## PANEL TO GET mseq COLUMN
def getMseqcolumn():
	helptext="""
	Provide the conformer-ID column:
	   """
	helpPanel=createComponent(type="ui",id="help", description="",component=textarea(helptext))
	colPanel = createComponent(type="column", id="mseq", description="Select Conformer-ID column",dataset=script.project.getActiveDataset())
	finalPanel= createComponent(type="group",id="Provide Conformer-ID column", description="Conformer-ID column?",components=[helpPanel,colPanel])
	result=showDialog(finalPanel)
	return result['mseq']

##------------------------------------------
## GET INDICES OF NOT-SELECTED COLUMNS
def notSelectedColumns(columns):
	selected = [ ]
	for i in range(columns.getSize()):
		selected.append(columns.get(i))
		#print columns.get(i), ":", dataset[columns.get(i)].getName()
	#print selected

	## mseq, p-value get selected in addition to 'columns'; need to add p-value to 'selected' list
	## for mseq,  no need as we need its (unaveraged) entry unlike p-value
	## 'pvalue' already got using "getPcolumn()" before this method gets called

	selected.append(pvalue)   

	total = range(dataset.getColumnCount())
	for i in selected:
		total.remove(i)
	notselected = total
	return notselected
	
##------------------------------------------



## MAIN

dataset=script.project.getActiveDataset()


## GET REQUIRED COLUMNS 

columns=getInputColumns()
#for i in range(columns.getSize()):
#	print columns.get(i), ":", dataset[columns.get(i)].getName()


pvalue=getPcolumn()

mseq_col = getMseqcolumn()



ColsNotSelected =  notSelectedColumns(columns)    ## notSelectedColumns defined above

total_ids = dataset[mseq_col].getCategoryCount()
uniqcol = [ ]


for idx in ColsNotSelected:
	uniqrow = [ ]
	for catIdx in range(total_ids):
		row_set = dataset[mseq_col].getRowIndicesOfCategory(catIdx)
		row_id = min(row_set)
		uniqrow.append(dataset[idx][row_id])
	
	if (dataset[idx].getDatatype()=='string'):   ## 'Id' column in 'unselected columns'
		addCol = ColumnFactory.createStringColumn(dataset[idx].getName(), uniqrow)
		(addCol.getMetaData()).setMark('Identifier')
		uniqcol.append(addCol)
	elif (dataset[idx].getDatatype()=='object'):   ## 'Structure' column in 'unselected columns'
                addCol = ColumnFactory.createObjectColumn(dataset[idx].getName(), uniqrow)
		(addCol.getMetaData()).setMark(ChemProject.RAWSTRUCTURE)
                uniqcol.append(addCol)
	else:
		uniqcol.append(script.dataset.createFloatColumn(dataset[idx].getName(), uniqrow))
	
unscaledset = uniqcol



## SCALE COLUMNS BY p-value
scaledColumns = [ ]
for i in range(columns.getSize()):
	scaledColumns.append(dataset[columns.get(i)] * dataset[pvalue])


total_cols = len(scaledColumns)
averagedset = [ ]

for colIdx in range(total_cols):
	
	averagedcol = [ ]
	for catIdx in range(total_ids):
		row_set = dataset[mseq_col].getRowIndicesOfCategory(catIdx)
	
		avg_rows = 0      
		for i in row_set:
			#print i+1, scaledColumns[colIdx][i]
			avg_rows =  avg_rows + scaledColumns[colIdx][i]   ## for a set of conformers

		#print catIdx, avg_rows
		averagedcol.append(avg_rows)  ## append averaged set of conformers in a column

	averagedset.append(script.dataset.createFloatColumn(scaledColumns[colIdx].getName(), averagedcol))   ## append columns with averaged rows to get dataset

#print averagedset




## DEFINE A NEW CHILD-DATASET
rowIndices = [i for i in range(total_ids)]
colIndices = [ ]
script.project.addSubsetChild(rowIndices, colIndices , name="Averaged-Descriptors", addMarkedColumns=0)
subset=script.project.getActiveDataset()

for i in range(len(unscaledset)):
        addcol = unscaledset[i]
        subset.addColumn(addcol)

for i in range(len(averagedset)):
        addcol = averagedset[i]
        subset.addColumn(addcol)

script.view.Table().show()


## REPORT COMPLETION
parent=script.tool.getTool().getFrame()
mesg = "Done With Script Execution."
JOptionPane.showMessageDialog(parent,mesg,"STATUS!",JOptionPane.INFORMATION_MESSAGE)