Mercurial > hg > cc > cirrus_work
annotate bin/test_warc.py @ 52:e20c64917805
better debugging output
author | Henry S. Thompson <ht@inf.ed.ac.uk> |
---|---|
date | Fri, 07 Jul 2023 17:04:05 +0100 |
parents | 212da3fe3b19 |
children | 9837840f3328 |
rev | line source |
---|---|
42
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
1 import warc,sys |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
2 |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
3 OUT=open(sys.stdout.fileno(),'wb') |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
4 |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
5 if (debug:=(sys.argv[1]=='-d')): |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
6 sys.argv.pop(1) |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
7 |
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
8 def showme(wtype,buf,part): |
45 | 9 # This should exactly reproduce a complete warc file if called |
10 # as per version 1 below | |
42
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
11 if debug: |
52 | 12 OUT.write(b"----start %d-----\n"%part) |
13 OUT.write(buf) | |
14 if buf[-1]!=10: | |
15 OUT.write(b'\r\n') | |
16 if part==7: | |
17 OUT.write(b'\r\n') # to match complete file formatting | |
18 if debug: | |
19 OUT.write(b"----end %d-----\n"%part) | |
20 return OUT | |
42
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
21 |
52 | 22 # 1 |
23 warc.warc(sys.argv[1],showme,[b'response','warcinfo','request','metadata'],parts=int(sys.argv[2]),debug=debug) | |
42
689a0e311cd2
make warc.py a library, separate out testing
Henry S. Thompson <ht@inf.ed.ac.uk>
parents:
diff
changeset
|
24 |
45 | 25 # 2 |
52 | 26 #warc.warc(sys.argv[1],showme,[b'warcinfo'],parts=int(sys.argv[2]),debug=debug) |