split-essex.py

#!/usr/bin/env python
#=========================================================================
# This is OPEN SOURCE SOFTWARE governed by the MIT License.
# Copyright (C)2026 William H. Majoros <bmajoros@duke.edu>
#=========================================================================
import sys
import ProgramName
from EssexParser import EssexParser

def makeFilename(filestem,n):
    return filestem+str(n+1)+".essex"

#=========================================================================
# main()
#=========================================================================
if(len(sys.argv)!=4):
    exit(ProgramName.get()+" <in.essex> <elements-per-partition> <out-filestem>\n")
(infile,MAX_PER_CHUNK,filestem)=sys.argv[1:]
MAX_PER_CHUNK=int(MAX_PER_CHUNK)

chunkNum=0
fh=open(makeFilename(filestem,chunkNum),"wt")
numInChunk=0
parser=EssexParser(infile)
while(True):
    tree=parser.nextElem()
    if(tree is None): break
    tree.print(fh)
    numInChunk+=1
    if(numInChunk>=MAX_PER_CHUNK):
        fh.close()
        chunkNum+=1
        numInChunk=0
        fh=open(makeFilename(filestem,chunkNum),"wt")
fh.close()
parser.close()