LoadLogPropertyTable.py

#pylint: disable=no-init,invalid-name
import time
import datetime
import numbers
import bisect
import numpy
from mantid.api import * # PythonAlgorithm, AlgorithmFactory, WorkspaceProperty
from mantid.kernel import * # StringArrayProperty
from mantid.simpleapi import * # needed for Load

class LoadLogPropertyTable(PythonAlgorithm):

    def summary(self):
        """ Return summary
        """
        return "Creates a table of Run number against the log values for that run for a range of files.\
         It can use a single log value or a list of log values."

    # same concept as built in "CreateLogPropertyTable" but loads its own workspaces and needn't hold all in memory at once
    # select log values to put in table (list)
    # special cases for:
    # beamlog_(counts, frames, etc): last few points end up in next run's log. Find Maximum.
    # comment (separate function)
    # time series, take average for t>0 (if available)
    def PyInit(self):
        self.declareProperty(FileProperty(name="FirstFile",defaultValue="",action=FileAction.Load,extensions = ["nxs","raw"]),
                             "The first file to load from")
        self.declareProperty(FileProperty(name="LastFile",defaultValue="",action=FileAction.Load,extensions = ["nxs","raw"]),
                             "The Last file to load from, must be in the same directory, all files in between will also be used")
        self.declareProperty(StringArrayProperty("LogNames",direction=Direction.Input),
                             "The comma seperated list of properties to include. \n"+\
                             "The full list will be printed if an invalid value is used.")
        self.declareProperty(WorkspaceProperty("OutputWorkspace","",Direction.Output),"Table of results")

    def category(self):
        return "DataHandling\\Logs;Muon\\DataHandling"

    def getGeneralLogValue(self,ws,name,begin):
        # get log value
        # average time series over run
        # for beamlog, etc return flag=true and value to push into previous run
        if name=="comment":
            return (ws.getComment(),False,0)

        try:
            v=ws.getRun().getProperty(name)
        except:
            possibleLogs = ws.getRun().keys()
            possibleLogs.insert(0,'comment')
            message =  "The log name '" + name + "' was not found, possible choices are: " + str(possibleLogs)
            raise ValueError(message)
        try:
            times2=[]
            if hasattr(v,"unfiltered"):
                v=v.unfiltered()
            for tt in v.times:
                times2.append((datetime.datetime(*(time.strptime(str(tt),"%Y-%m-%dT%H:%M:%S")[0:6]))-begin).total_seconds())
        except StandardError:
            # print "probably not a time series"
            pass
        if name[0:8]=="Beamlog_" and (name.find("Counts")>0 or name.find("Frames")>0):
            i=bisect.bisect_right(times2,2) # allowance for "slow" clearing of DAE
            #print "returning max beam log, list cut 0:",i,":",len(times2)
            return (numpy.amax(v.value[i:]),True,numpy.amax(v.value[:i]))
        if v.__class__.__name__ =="TimeSeriesProperty_dbl" or v.__class__.__name__ =="FloatTimeSeriesProperty":
            i=bisect.bisect_left(times2,0)
            return (numpy.average(v.value[i:]),False,0)
        return (v.value,False,0)

    #pylint: disable=too-many-branches
    def PyExec(self):

        firstFileName=self.getProperty("FirstFile").value
        lastFileName=self.getProperty("LastFile").value

        firstRunNum, firstFileFirstDigit,firstFileLastDigit = self.getRunNumber(firstFileName)
        lastRunNum, lastFileFirstDigit,LastFileLastDigit = self.getRunNumber(lastFileName)

        if firstFileName[:lastFileFirstDigit] != lastFileName[:lastFileFirstDigit]:
            raise Exception("Files from different directories or instruments")
        if firstFileName[firstFileName.rindex('.')] != lastFileName[firstFileName.rindex('.')]:
            raise Exception("Files of different types")
        if firstFileLastDigit-firstFileFirstDigit != LastFileLastDigit-lastFileFirstDigit:
            raise Exception("File numbering error")
        if lastRunNum < firstRunNum:
            raise Exception("Run numbers must increase")

        # table. Rows=runs, columns=logs (col 0 = run number)
        collist=self.getProperty("LogNames").value
        wsOutput=WorkspaceFactory.createTable()
        wsOutput.addColumn("int","RunNumber")

        # loop and load files. Absolute numbers for now.
        for loopRunNum in range(firstRunNum,lastRunNum+1):
            # create a file path for intervening files, based from the 1st filename
            thispath=firstFileName[:firstFileFirstDigit] + \
                     str(loopRunNum).zfill(firstFileLastDigit-firstFileFirstDigit) + \
                     firstFileName[firstFileLastDigit:]

            loadedWs = self.loadMetaData(thispath)
            if loadedWs is None:
                continue

            #check if the ws is a group
            ws = loadedWs
            if ws.id() == 'WorkspaceGroup':
                ws=ws[0]

            begin=datetime.datetime(*(time.strptime(ws.getRun().getProperty("run_start").value,"%Y-%m-%dT%H:%M:%S")[0:6])) # start of day
            vallist=[loopRunNum]
            for col in collist:
                try:
                    (colValue, leftover, lval)=self.getGeneralLogValue(ws, col, begin)
                except ValueError:
                    # this is a failure to find the named log
                    raise
                vallist.append(colValue)
                if loopRunNum==firstRunNum:
                    if isinstance(colValue, numbers.Number):
                        wsOutput.addColumn("double",col)
                    else:
                        wsOutput.addColumn("str",col)
                if leftover and loopRunNum>firstRunNum:
                    if lval>wsOutput.cell(col,loopRunNum-firstRunNum-1):
                        wsOutput.setCell(col,loopRunNum-firstRunNum-1, lval)
            wsOutput.addRow(vallist)

        self.setProperty("OutputWorkspace",wsOutput)

    def loadMetaData(self, thispath):
        loadedWs = None
        try:
            loadAlg = self.createChildAlgorithm('Load')
            #set Filename first
            loadAlg.setProperty('Filename', thispath)
            loadAlg.setProperty('OutputWorkspace', '__CopyLogsTmp')
            try:
                #try to set MetaDataOnly
                loadAlg.setProperty('MetaDataOnly', True)
            except (ValueError,RuntimeError):
                #If that fails set SpectrumMin and SpectrumMax
                loadAlg.setProperty('SpectrumMin', 1)
                loadAlg.setProperty('SpectrumMax', 1)
            loadAlg.execute()

            outWSPropName = 'OutputWorkspace'
            try:
                loadedWs = loadAlg.getProperty(outWSPropName).value
            except RuntimeError:
                raise RuntimeError("No output workspace for " + thispath)
        except (ValueError,RuntimeError):
            return None

        return loadedWs

    def getRunNumber(self, fileName):
        # Find last . and step back until you find a digit
        lastDigitIndex = fileName.rindex('.')
        while not fileName[lastDigitIndex - 1].isdigit():
            lastDigitIndex -= 1

        # Keep going back until you find the start of the number sequence
        firstDigitIndex = lastDigitIndex - 1
        while fileName[firstDigitIndex - 1].isdigit():
            firstDigitIndex -= 1
        runNumber = int(fileName[firstDigitIndex:lastDigitIndex])
        return runNumber, firstDigitIndex, lastDigitIndex


AlgorithmFactory.subscribe(LoadLogPropertyTable())