tsv2json

#!/usr/bin/env python2.7
"""
Convert TSV (with header) into a stream of newline-delimited JSON objects.

EXAMPLE: pretty print
cat events.2015.20170206133646.tab | tsv2json | jq .
"""

import sys
import json as jsonmod
try:
    exec "import ujson as jsonmod"
except ImportError:
    pass

warning_count = 0
warning_max = 20
def warning(s):
  global warning_count
  warning_count += 1
  if warning_count > warning_max: return
  print>>sys.stderr, "WARNING: %s" % s

headerline = sys.stdin.readline()
colnames = headerline.rstrip("\n").split("\t")

for line_num,line in enumerate(sys.stdin):
    parts = line.rstrip("\n").split("\t")
    if len(parts) != len(colnames):
        warning("Mismatched %d columns in line %d, but %s in header" % (len(parts), line_num+2, len(colnames)))
        n = min(len(parts),len(colnames))
        dct = dict(zip(colnames[:n], parts[:n]))
    else:
        dct = dict(zip(colnames,parts))
    print jsonmod.dumps(dct)