joe/everytab
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

download cc-index first with aws cli instead of streaming it

Browse source
This commit is contained in:
Joe Lothan 2026-05-20 08:14:22 -04:00
parent 564919c5cc
commit 9308b5e039
1 changed files with 23 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

33
pipeline/01_cc_index/query.sh
View file

@ -2,8 +2,8 @@
set -euo pipefail set -euo pipefail
# Query Common Crawl columnar index and populate the hosts table. # Query Common Crawl columnar index and populate the hosts table.
# Uses DuckDB with S3 (credential chain) to read parquet files directly, # Downloads CC-Index parquet files locally first (avoids S3 rate limits),
# and the postgres extension to write results into RDS. # then uses DuckDB to query locally and write results into RDS.
usage() { usage() {
cat <<'EOF' cat <<'EOF'
@ -64,15 +64,10 @@ if [ "$LIMIT" -gt 0 ] 2>/dev/null; then
LIMIT_CLAUSE="LIMIT ${LIMIT}" LIMIT_CLAUSE="LIMIT ${LIMIT}"
fi fi
S3_PATH="s3://commoncrawl/cc-index/table/cc-main/warc/crawl=${CRAWL}/subset=warc/*.parquet" S3_PATH="s3://commoncrawl/cc-index/table/cc-main/warc/crawl=${CRAWL}/subset=warc/"
LOCAL_INDEX="/tmp/cc-index"
QUERY=" QUERY="
INSTALL aws;
LOAD aws;
INSTALL httpfs;
LOAD httpfs;
CREATE SECRET (TYPE S3, PROVIDER CREDENTIAL_CHAIN);
INSTALL postgres; INSTALL postgres;
LOAD postgres; LOAD postgres;
ATTACH '${DB_URL}' AS pg (TYPE POSTGRES); ATTACH '${DB_URL}' AS pg (TYPE POSTGRES);
@ -85,7 +80,7 @@ SELECT
first(warc_filename ORDER BY CASE WHEN url_protocol = 'https' THEN 0 ELSE 1 END) AS warc_filename, first(warc_filename ORDER BY CASE WHEN url_protocol = 'https' THEN 0 ELSE 1 END) AS warc_filename,
first(warc_record_offset ORDER BY CASE WHEN url_protocol = 'https' THEN 0 ELSE 1 END) AS warc_record_offset, first(warc_record_offset ORDER BY CASE WHEN url_protocol = 'https' THEN 0 ELSE 1 END) AS warc_record_offset,
first(warc_record_length ORDER BY CASE WHEN url_protocol = 'https' THEN 0 ELSE 1 END) AS warc_record_length first(warc_record_length ORDER BY CASE WHEN url_protocol = 'https' THEN 0 ELSE 1 END) AS warc_record_length
FROM read_parquet('${S3_PATH}') FROM read_parquet('${LOCAL_INDEX}/*.parquet')
WHERE url_path = '/' WHERE url_path = '/'
AND content_mime_type = 'text/html' AND content_mime_type = 'text/html'
AND fetch_status = 200 AND fetch_status = 200
@ -113,6 +108,18 @@ echo "=== CC-Index Query ==="
echo "Crawl: $CRAWL" echo "Crawl: $CRAWL"
echo "S3 path: $S3_PATH" echo "S3 path: $S3_PATH"
echo "Limit: ${LIMIT} (0 = no limit)" echo "Limit: ${LIMIT} (0 = no limit)"
echo ""
# Download CC-Index parquet files locally (avoids S3 rate limits during query)
if [ -d "$LOCAL_INDEX" ] && [ "$(ls -A "$LOCAL_INDEX"/*.parquet 2>/dev/null)" ]; then
echo "Using cached CC-Index at $LOCAL_INDEX"
else
echo "Downloading CC-Index parquet files..."
mkdir -p "$LOCAL_INDEX"
aws s3 sync "$S3_PATH" "$LOCAL_INDEX/" --quiet
echo "Downloaded $(ls "$LOCAL_INDEX"/*.parquet | wc -l) parquet files"
fi
echo "" echo ""
echo "Starting query..." echo "Starting query..."
echo "" echo ""
@ -154,3 +161,9 @@ SELECT json_build_object(
); );
" > "$STATS_FILE" " > "$STATS_FILE"
echo "Stats written to $STATS_FILE" echo "Stats written to $STATS_FILE"
# Clean up local CC-Index files
echo ""
echo "Cleaning up $LOCAL_INDEX..."
rm -rf "$LOCAL_INDEX"
echo "Done"