Page MenuHomePhabricator
Paste P10461

Help script for setting up ORES articletopics test data
ActivePublic

Authored by Tgr on Feb 20 2020, 2:24 AM.
Tags
None
Referenced Files
F31624932: raw.txt
Feb 21 2020, 12:48 AM
F31624920: raw.txt
Feb 21 2020, 12:18 AM
F31623437: raw.txt
Feb 20 2020, 3:21 AM
F31623423: raw.txt
Feb 20 2020, 2:24 AM
Subscribers
Tokens
"Like" token, awarded by kostajh.
#!/bin/bash
set -e
# customize these if you don't use vagrant
FORWARDED_PORT=`cat /vagrant/.settings.yaml | ack '^http_port:\s*(\d+)$' --output '$1' || echo 8080`
ACTION_API=http://dev.wiki.local.wmftest.net:$FORWARDED_PORT/w/api.php
ELASTIC_API=http://localhost:9200
# $1: article title, $2: article text, outputs page ID
api_edit() {
curl -s -d "format=json&action=edit&title=$1&text=$2" --data-urlencode 'token=+\' "$ACTION_API"
}
# $1: page ID, $2: ORES data as JSON string
api_index() {
curl -s -XPOST "$ELASTIC_API/wiki_content/page/$1/_update" -H 'Content-Type: application/json' -d '{"doc": {"ores_articletopics": '"$2"' }}'
}
# $1: article title, $2: article text, $3: ORES data as JSON string
api_execute() {
output=$(api_edit $1 $2)
echo "$output" | jq .
page_id=$(echo "$output" | jq '.edit.pageid')
mwscript extensions/CirrusSearch/maintenance/forceSearchIndex.php --ids $page_id
api_index $page_id "$3" | jq .
}
# test data; page name, page content, ES articletopic field content
api_execute Test_books "books" '["Culture.Media.Books|980"]'
api_execute Test_stem "stem" '["STEM.STEM*|970"]'
api_execute Test_both "books stem" '["Culture.Media.Books|600","STEM.STEM*|550"]'
api_execute Test_none "-" '[]'
# check by visiting Special:ApiSandbox#action=query&format=json&prop=cirrusdoc&titles=Test%20books%7CTest%20stem%7CTest%20both%7CTest%20none