diff options
Diffstat (limited to 'notes/cleanups/wayback_timestamps.md')
| -rw-r--r-- | notes/cleanups/wayback_timestamps.md | 103 |
1 files changed, 100 insertions, 3 deletions
diff --git a/notes/cleanups/wayback_timestamps.md b/notes/cleanups/wayback_timestamps.md index c70ec5b2..81785992 100644 --- a/notes/cleanups/wayback_timestamps.md +++ b/notes/cleanups/wayback_timestamps.md @@ -26,14 +26,53 @@ Filter to files with problem of interest: Wow, this is a lot more than I thought! +There might also be some other short URL patterns, check for those: + + zcat file_export.json.gz \ + | pv -l \ + | rg 'web.archive.org/web/\d{1,11}/' \ + | gzip \ + > files_20211007_veryshortts.json.gz + # skipped, mergine with below + + zcat file_export.json.gz \ + | rg 'web.archive.org/web/None/' \ + | pv -l \ + > /dev/null + # 0.00 0:10:06 [0.00 /s] + # whew, that pattern has been fixed it seems + + zcat file_export.json.gz | rg '/None/' | pv -l > /dev/null + # 2.00 0:10:01 [3.33m/s] + + zcat file_export.json.gz \ + | rg 'web.archive.org/web/\d{13}/' \ + | pv -l \ + > /dev/null + # 0.00 0:10:09 [0.00 /s] + +Yes, 4-digit is a popular pattern as well, need to handle those: + + zcat file_export.json.gz \ + | pv -l \ + | rg 'web.archive.org/web/\d{4,12}/' \ + | gzip \ + > files_20211007_moreshortts.json.gz + # 111M 0:13:22 [ 139k/s] + + zcat files_20211007_moreshortts.json.gz | wc -l + + zcat files_20211007_moreshortts.json.gz | shuf -n10000 > files_20211007_moreshortts.10k_sample.json + # 9,958,854 + ## Fetch Complete URL Want to export JSON like: file_entity [existing file entity] - full_urls[] - <short>: <long> + full_urls[]: list of Dicts[str,str] + <short_url>: <full_url> status: str Status one of: @@ -41,5 +80,63 @@ Status one of: - 'success-self': the file already has a fixed URL internally - 'success-db': lookup URL against sandcrawler-db succeeded, and SHA1 matched - 'success-cdx': CDX API lookup succeeded, and SHA1 matched -- 'fail-hash': found a CDX record, but wrong hash - 'fail-not-found': no matching CDX record found + +Ran over a sample: + + cat files_20211007_shortts.10k_sample.json | ./fetch_full_cdx_ts.py > sample_out.json + + cat sample_out.json | jq .status | sort | uniq -c + 5 "fail-not-found" + 576 "success-api" + 7212 "success-db" + 2207 "success-self" + + head -n1000 | ./fetch_full_cdx_ts.py > sample_out.json + + zcat files_20211007_veryshortts.json.gz | head -n1000 | ./fetch_full_cdx_ts.py | jq .status | sort | uniq -c + 2 "fail-not-found" + 168 "success-api" + 208 "success-db" + 622 "success-self" + +Investigating the "fail-not-found", they look like http/https URL +not-exact-matches. Going to put off handling these for now because it is a +small fraction and more delicate. + +Again with the broader set: + + cat files_20211007_moreshortts.10k_sample.json | ./fetch_full_cdx_ts.py > sample_out.json + + cat sample_out.json | jq .status | sort | uniq -c + 9 "fail-not-found" + 781 "success-api" + 6175 "success-db" + 3035 "success-self" + + +## Cleanup Process + +Other possible cleanups to run at the same time, which would not require +external requests or other context: + +- URL has ://archive.org/ link with rel=repository => rel=archive +- mimetype is bogus => clean mimetype +- bogus file => set some new extra field, like scope=stub or scope=partial (?) + +It looks like the rel swap is already implemented in `generic_file_cleanups()`. +From sampling it seems like the mimetype issue is pretty small, so not going to +bite that off now. The "bogus file" issue requires thought, so also skipping. + +## Commands + +Running with 8x parallelism to not break things; expecting some errors along +the way, may need to add handlers for connection errors etc: + + zcat files_20211007_moreshortts.json.gz \ + | parallel -j8 --linebuffer --round-robin --pipe ./fetch_full_cdx_ts.py \ + | pv -l \ + | gzip \ + > files_20211007_moreshortts.fetched.json.gz + +At 300 records/sec, this should take around 9-10 hours to process. |