refactor reingest scripts

1 files changed, 15 insertions, 66 deletions

diff --git a/sql/dump_reingest_weekly.sql b/sql/dump_reingest_weekly.sql
index e7cd553..65800eb 100644
--- a/sql/dump_reingest_weekly.sql
+++ b/sql/dump_reingest_weekly.sql
@@ -3,11 +3,19 @@ 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.ingest_type = 'pdf'
+        AND ingest_file_result.hit = false
         AND ingest_request.created < NOW() - '8 hour'::INTERVAL
         AND ingest_request.created > NOW() - '8 day'::INTERVAL
-        AND ingest_file_result.hit = false
-        AND ingest_file_result.status like 'spn2-%'
+        AND (ingest_request.ingest_request_source = 'fatcat-changelog'
+             OR ingest_request.ingest_request_source = 'fatcat-ingest')
+        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'
@@ -15,68 +23,9 @@ COPY (
         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_weekly_spn2-error_current.rows.json';
+) TO '/srv/sandcrawler/tasks/reingest_weekly_current.rows.json';
 
-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.ingest_type = 'pdf'
-        AND ingest_file_result.hit = false
-        AND ingest_file_result.status like 'cdx-error'
-        AND ingest_request.created < NOW() - '8 hour'::INTERVAL
-        AND ingest_request.created > NOW() - '8 day'::INTERVAL
-        AND (ingest_request.ingest_request_source = 'fatcat-changelog'
-             OR ingest_request.ingest_request_source = 'fatcat-ingest')
-) TO '/srv/sandcrawler/tasks/reingest_weekly_cdx-error_current.rows.json';
-
-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.ingest_type = 'pdf'
-        AND ingest_file_result.hit = false
-        AND ingest_file_result.status like 'cdx-error'
-        AND ingest_request.created < NOW() - '8 hour'::INTERVAL
-        AND ingest_request.created > NOW() - '8 day'::INTERVAL
-        AND (ingest_request.ingest_request_source != 'fatcat-changelog'
-             AND ingest_request.ingest_request_source != 'fatcat-ingest')
-) TO '/srv/sandcrawler/tasks/reingest_weekly_cdx-error_bulk_current.rows.json';
-
-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.ingest_type = 'pdf'
-        AND ingest_file_result.hit = false
-        AND ingest_file_result.status like 'wayback-error'
-        AND ingest_request.created < NOW() - '8 hour'::INTERVAL
-        AND ingest_request.created > NOW() - '8 day'::INTERVAL
-) TO '/srv/sandcrawler/tasks/reingest_weekly_wayback-error_current.rows.json';
-
-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.ingest_type = 'pdf'
-        AND ingest_file_result.hit = false
-        AND ingest_file_result.status like 'gateway-timeout'
-        AND ingest_request.created < NOW() - '8 hour'::INTERVAL
-        AND ingest_request.created > NOW() - '8 day'::INTERVAL
-        AND (ingest_request.ingest_request_source = 'fatcat-changelog'
-             OR ingest_request.ingest_request_source = 'fatcat-ingest')
-) TO '/srv/sandcrawler/tasks/reingest_weekly_gateway-timeout.rows.json';
-
-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.ingest_type = 'pdf'
-        AND ingest_file_result.hit = false
-        AND ingest_file_result.status like 'petabox-error'
-        AND ingest_request.created < NOW() - '8 hour'::INTERVAL
-        AND ingest_request.created > NOW() - '8 day'::INTERVAL
-        AND (ingest_request.ingest_request_source = 'fatcat-changelog'
-             OR ingest_request.ingest_request_source = 'fatcat-ingest')
-) TO '/srv/sandcrawler/tasks/reingest_weekly_petabox-error_current.rows.json';
+-- bulk re-tries would be:
+--      AND (ingest_request.ingest_request_source != 'fatcat-changelog'
+--           AND ingest_request.ingest_request_source != 'fatcat-ingest')