Get simple reports from Peter Ward files, reading LDR/05.

peter_ward_report.py

#!/usr/bin/env python
#-*- coding: utf-8 -*-
"""
Get simple reports from Peter Ward files, reading LDR/05.
Peter Ward files come from a paid subscription as of 201808, processed by ke, jeb, pmg.
Run like this: python peter_ward_report.py -f 'unname18.01'
It produces a csv file like this ...

ldr05,status,lccn
c,changed,n  00004614 
c,changed,n  00005716 
c,changed,n  00007215
...

and a log with counts, like this ...

deletes: 172
changed: 4039
increase: 0
new: 5947
prepub:0

NOTE: the Gary Strawn Authority Loader produces log files with counts as well. 

from 20180808
pmg
"""
import argparse
import csv
import glob
import pymarc
import sys

def leader_reader(filename):
	# write column heading in csv file
	with open("../../Documents/working/authorities/"+filename+"_test_report_out.csv","w+") as outfile:
		outfile.write("ldr05,status,lccn\n")
	
	# read in MARC file (they lack file extensions)
	for m in glob.glob(r'../../Documents/working/authorities/'+filename):
		marc = str(m)
	
	reader = pymarc.MARCReader(file(marc))
	
	# counters for count log
	dcount = 0
	ccount = 0
	acount = 0
	ncount = 0
	pcount = 0
	
	try:
		for record in reader:
			ldr05 = record.leader[5:6]
			lccn = record['001'].value()
			ldr = ''
			if ldr05 == 'd':
				ldr = 'deleted'
				dcount += 1
			elif ldr05 == 'c':
				ldr = 'changed'
				ccount += 1
			elif ldr05 == 'a':
				ldr = 'increase in encoding level'
				acount += 1
			elif ldr05 == 'n':
				ldr = 'new'
				ncount += 1
			elif ldr05 == 'p':
				ldr = 'increase in encoding levevl from prepublication' 
				pcount += 1
			else:
				ldr = 'wth'
				
			with open("../../Documents/working/authorities/"+filename+"_test_report_out.csv","ab") as outfile:
				outfile.write(ldr05+','+ldr+','+lccn+"\n")
	except Exception as e:
			exc_type, exc_obj, exc_tb = sys.exc_info()
			fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1]
			print(exc_type, fname, exc_tb.tb_lineno)

	# write to count log
	with open("../../Documents/working/authorities/"+filename+"_test_report_counts.txt",'wb') as clog:
		clog.write('deletes: %s\nchanged: %s\nincrease: %s\nnew: %s\nprepub:%s\n' % (dcount,ccount,acount,ncount,pcount))


if __name__ == "__main__":
	parser = argparse.ArgumentParser(description='Generate Peter Ward file.')
	parser.add_argument('-f','--filename',type=str,dest="filename",help="The filename of the report as downloaded from the Peter Ward service e.g. unsub18.30'",required=True)
	args = vars(parser.parse_args())

	filename = args['filename']

	leader_reader(filename)