From b91a7d1d8932fb75f0ae1f0726409f8c2904e3cd Mon Sep 17 00:00:00 2001 From: Barbara Miller Date: Wed, 28 Jun 2023 11:43:53 -0700 Subject: [PATCH] more updates qa prototyping --- warcprox/dedup.py | 45 ++++++++++++++++++++++----------------------- 1 file changed, 22 insertions(+), 23 deletions(-) diff --git a/warcprox/dedup.py b/warcprox/dedup.py index 93a335b..c49edd4 100644 --- a/warcprox/dedup.py +++ b/warcprox/dedup.py @@ -62,21 +62,21 @@ class DedupableMixin(object): else: return recorded_url.response_recorder.payload_size() > self.min_binary_size -class LimitRevisitsPGMixin(object): +class LimitRevisitsPGMixin(): """ Limit revisits recorded to one per revisit_key """ - def __init__(self, options=warcprox.Options()): + def __init__(self): self.datasource = "postgresql://archiveit@db.qa-archive-it.org/archiveit" # "postgresql://user@db_host/db_name" self.datatable = "crawl_revisits" # postgres table in db_name def limit_revisits(self, recorded_url, hash_plus_url=None, revisit_key=None): + # tracks revisits, returns True when we've seen revisit before, else False if not hash_plus_url: - hash_plus_url = b"".join( - (warcprox.digest_str(recorded_url.payload_digest, - self.options.base32), - recorded_url.url) - ).decode() + digest = warcprox.digest_str(recorded_url.payload_digest, + self.options.base32) + digest = digest[5:] if digest.startswith(b'sha1:') else digest + hash_plus_url = b"".join([digest, recorded_url.url]).decode() if not revisit_key: # use ait-job-id if available if ( @@ -84,11 +84,11 @@ class LimitRevisitsPGMixin(object): and "metadata" in recorded_url.warcprox_meta and "ait-job-id" in recorded_url.warcprox_meta["metadata"] ): - revisit_key = recorded_url.warcprox_meta["metadata"]["ait-job-id"] + revisit_key = str(recorded_url.warcprox_meta["metadata"]["ait-job-id"]) else: - revisit_key = 'all' + revisit_key = '__unspecified__' - query = "SELECT exists(SELECT 1 FROM crawl_revisits WHERE hash_plus_url = %s LIMIT 1);" + query = "SELECT exists(SELECT 1 FROM crawl_revisits WHERE hash_plus_url = %s and crawl_id = %s LIMIT 1);" try: conn = psycopg2.connect(self.datasource) @@ -97,23 +97,22 @@ class LimitRevisitsPGMixin(object): return False cur = conn.cursor() try: - cur.execute(query, (hash_plus_url,)) + cur.execute(query, (hash_plus_url, revisit_key)) except Exception as e: - self.logger.warning("exception querying for %s in %s: %s", hash_plus_url, revisit_key, e) + self.logger.warning("exception querying for %s in %s: %s", digest, revisit_key, e) return False result = cur.fetchone() - if result[0]: - logging.info("result[0]: %s", result[0]) - if result[0] and result[0] == True: - logging.info("skipping revisit for url %s and hash %s", recorded_url.url, hash) + if result and result == (True, ): + logging.info("skipping revisit for url %s and hash %s", recorded_url.url, digest) return True - else: - query = "INSERT INTO crawl_revisits (crawl_id, hash_plus_url) VALUES (%s, %s);" - try: - cur.execute(query, (revisit_key, hash_plus_url)) - except Exception as e: - self.logger.warning("exception inserting %s in %s: %s", hash_plus_url, revisit_key, e) + + query = "INSERT INTO crawl_revisits (crawl_id, hash_plus_url) VALUES (%s, %s);" + try: + cur.execute(query, (revisit_key, hash_plus_url)) + conn.commit() + except Exception as e: + self.logger.warning("exception inserting %s in %s for %s: %s", digest, revisit_key, recorded_url.url, e) return False @@ -516,7 +515,7 @@ class BatchTroughLoader(warcprox.BaseBatchPostfetchProcessor, LimitRevisitsPGMix def __init__(self, trough_dedup_db, options=warcprox.Options()): warcprox.BaseBatchPostfetchProcessor.__init__(self, options) - LimitRevisitsPGMixin.__init__(self, options) + LimitRevisitsPGMixin.__init__(self) self.trough_dedup_db = trough_dedup_db def _startup(self):