Multiple loops logic and speed optimization in Python? -

here 2 python functions transfer data 1 file file. both source file , target file have same number of objects different data.

def getblock(rigobj, objname):     rigobj.seek(0)     tag = false     block = ""     line in rigobj:         if line.find("objectalias " + str(objname) + "\n") != -1:             line in rigobj:                 if line.find("beginkeyframe") != -1:                     tag = true                 elif line.lstrip().startswith("0.000 ") , line.rstrip().endswith("default"):                     tag = false                     break                 elif tag:                     block += line     return (block)  def buildscene(sceneobj, rigobj, objlist):     sceneobj.seek(0)     rigobj.seek(0)     newscene = ""     line in sceneobj:         newscene += line         obj in objlist:             if line.find("objectalias " + obj + "\n") != -1:                 tag = true                 line in sceneobj:                     if line.find("beginkeyframe") != -1:                         newscene += line                         newscene += getblock(rigobj, obj)                         tag = false                     elif line.lstrip().startswith("0.000 ") , line.rstrip().endswith("default"):                         newscene += line                         tag = true                         break                     elif tag:                         newscene += line     return (newscene) 

getblock sub-function getting data rigobj;
buildscene main function, has 3 parameters:
first parameter(sceneobj) file want put data into;
second parameter(rigobj) file data from;
third parameter(objlist) list of object's data transfered.

so far, function job, problem bit of slow(sceneobj<10mb, rigobj<2mb, objlist<10 objects), not sure if there logic problem in code, should loop sceneobj first or loop objlist first? affect speed?

update:
both sceneobj , rigobj have similar data this:

lines beginobject                lines                      objectalias xxx           #--> object in transfer list lines                      beginkeyframe 10 12          -9.000 4095 default     #--> transfer begins   lines                   #--> transfer rigobj sceneobj , override lines in sceneobj   -8.000 63 default       #--> same   lines                   #--> same   -7.000 63 default       #--> same   lines                   #--> same   -1.000 63 default       #--> same   lines                   #--> transfer ends   0.000 -1 default           lines                    endkeyframe                endmotion                  lines                      endobject 

the data want transfered , overrided lines bewteen beginkeyframe , 0.000 -1 default of specified objects(by objlist)

most obvious optimization index data getblock function, able seek needed position instead of parsing full file beginning.

like so:

def create_rig_index(rig_obj):     """ function creates dict of offsets specific objectalias          example:               data:                     line offset 100: objectalias xxx                    more lines                    line offset 200: objectalias yyy                    more lines                    line offset 300: objectalias xxx                    more lines                result be:                    xxx: [100, 300]                    yyy: [200]     """     idx = defaultdict( list )     position = 0     line in rig_obj:         strip_line = line.strip()         if strip_line.startswith( "objectalias" ):             obj_name = strip_line.split()[1]             idx[ obj_name ].append( position )          # unfortunately python prevent `tell` calls during iteration.         position += len( bytes( line, 'utf-8' ) )          # if data guaranteed ascii possible use len( line )         # or can write custom line generator on top of read function.     return idx;  def getblock(rigobj, rigidx, objname):     """ same getblock, uses precalculated offsets"""     block = ""     idx in rigidx[ objname ]:         rigobj.seek( idx )         tag = false         line in rigobj:             if line.find("beginkeyframe") != -1:                 tag = true             elif line.lstrip().startswith("0.000 ") , line.rstrip().endswith("default"):                 break             elif tag:                 block += line      return (block) 

in buildscene method should create rig_index before running loop, , use index in getblock function.