aboutsummaryrefslogtreecommitdiffstats
path: root/sql/dump_reingest_spn.sql
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2021-09-30 16:49:47 -0700
committerBryan Newbold <bnewbold@archive.org>2021-09-30 16:49:47 -0700
commitf6125848f627ae9bfd3a36d807d2349e1c66bfe3 (patch)
tree8419947e17dd27621a5b17098eacaedf1a0227a5 /sql/dump_reingest_spn.sql
parent24edc1c754d35df1020907a9c8941f16eaf71440 (diff)
downloadsandcrawler-f6125848f627ae9bfd3a36d807d2349e1c66bfe3.tar.gz
sandcrawler-f6125848f627ae9bfd3a36d807d2349e1c66bfe3.zip
refactor reingest scripts
Diffstat (limited to 'sql/dump_reingest_spn.sql')
-rw-r--r--sql/dump_reingest_spn.sql25
1 files changed, 25 insertions, 0 deletions
diff --git a/sql/dump_reingest_spn.sql b/sql/dump_reingest_spn.sql
new file mode 100644
index 0000000..a8ed72f
--- /dev/null
+++ b/sql/dump_reingest_spn.sql
@@ -0,0 +1,25 @@
+
+COPY (
+ SELECT row_to_json(ingest_request.*) FROM ingest_request
+ LEFT JOIN ingest_file_result ON ingest_file_result.base_url = ingest_request.base_url
+ WHERE ingest_request.ingest_type = 'pdf'
+ AND ingest_file_result.hit = false
+ AND ingest_request.created < NOW() - '2 hour'::INTERVAL
+ AND ingest_request.created > NOW() - '31 day'::INTERVAL
+ AND ingest_request.ingest_request_source = 'savepapernow-web'
+ AND (
+ ingest_file_result.status like 'spn2-%'
+ -- OR ingest_file_result.status like 'cdx-error'
+ -- OR ingest_file_result.status like 'wayback-error'
+ -- OR ingest_file_result.status like 'wayback-content-error'
+ OR ingest_file_result.status like 'petabox-error'
+ -- OR ingest_file_result.status like 'gateway-timeout'
+ )
+ AND ingest_file_result.status != 'spn2-error:invalid-url-syntax'
+ AND ingest_file_result.status != 'spn2-error:filesize-limit'
+ AND ingest_file_result.status != 'spn2-error:not-found'
+ AND ingest_file_result.status != 'spn2-error:blocked-url'
+ AND ingest_file_result.status != 'spn2-error:too-many-redirects'
+ AND ingest_file_result.status != 'spn2-error:network-authentication-required'
+ AND ingest_file_result.status != 'spn2-error:unknown'
+) TO '/srv/sandcrawler/tasks/reingest_spn.rows.json';