Prototype script converting FALCON first string graph to GFA2 format

fsg_to_gfa2.py

read_length = {}
edges = {}

print "H\tVN:Z:2.0"
with open("sg_edges_list") as f:
    for row in f:
        row = row.strip().split()
        """ 000264577:E 000291380:B 000291380  4496     0 23388 99.53 G """
        v, w, sid2, x2, y2, ovlp, idt, t = row
        ovlp = int(ovlp)
        x2 = int(x2)
        y2 = int(y2)
        if sid2 not in read_length:
            # sg_edges_list was not designed to keep the read length
            # this read length is an estimation (for now)
            l = int(ovlp) + abs( x2 - y2)
            read_length[sid2] = l
            rec = ("S", sid2, str(l), "*")
            print "\t".join(rec)

        edges[ (v, w) ] = ( sid2, x2, y2, ovlp, idt, t )

for v, w in edges:
    #if v > w: continue #dedup

    e = edges[ (v, w) ]
    v = v.split(":")
    w = w.split(":")

    vv = v[0] , "B" if v[1] == "E" else "E"
    ww = w[0] , "B" if w[1] == "E" else "E"
    ee = edges[ (":".join(ww), ":".join(vv)) ]

    tag = "T:Z:"+e[-1]
    if v[1] == "B" and w[1] == "B":
        sid1, sid2 = v[0], w[0]
        orientation = "+"
        b1 = "0"  # another representation
        e1 = "$"+ str(ee[1])
        b2 = str(e[1])
        e2 = "$0"
        print "\t".join( ("E", "*", sid1, orientation, sid2, b1, e1, b2, e2, tag) )
        continue

    if v[1] == "E" and w[1] == "E":
        sid1, sid2 = v[0], w[0]
        orientation = "+"
        #b1 = "$" + str(read_length[sid1] - ee[1])
        b1 = str(ee[1])  # another representation
        e1 = "$0"
        b2 = "0"
        e2 = "$"+str(e[1])
        print "\t".join( ("E", "*", sid1, orientation, sid2, b1, e1, b2, e2, tag) )
        continue

    if v[1] == "E" and w[1] == "B":
        sid1, sid2 = v[0], w[0]
        orientation = "-"
        b1 = str(ee[1])
        e1 = "$0"
        b2 = str(e[1])
        e2 = "$0"
        print "\t".join( ("E", "*", sid1, orientation, sid2, b1, e1, b2, e2, tag) )
        continue

    if v[1] == "B" and w[1] == "E":
        sid1, sid2 = w[0], v[0]
        orientation = "-"
        b1 = "0"
        e1 = "$"+str(e[2]-e[1])
        b2 = "0"
        e2 = "$"+str(ee[2]-ee[1])
        print "\t".join( ("E", "*", sid1, orientation, sid2, b1, e1, b2, e2, tag) )