INSTALL

Deployment
==========

To deploy the current HEAD at GitHub, run as yourself with
ssh permissions for Toolserver and Tools:

$ ssh yarrow.toolserver.org sudo -u dbreps bash <<EOF &&
cd ~dbreps/src/database-reports &&
git pull --rebase &&
python setup.py install --user --install-scripts ~dbreps/bin &&
crontab crontab.yarrow
EOF
ssh submit.toolserver.org sudo -u dbreps cronie \~dbreps/src/database-reports/cronietab.submit &&
ssh tools-login.wmflabs.org sudo -iu local-dbreps bash <<EOF &&
cd ~/src/database-reports &&
git pull --rebase &&
python setup.py install --user --install-scripts ~/bin &&
crontab crontab.tools
EOF
echo Successfully deployed dbreps to yarrow, submit and tools-login.

Caution: This does not install the required modules per
README (MySQLdb, oursql, python-Levenshtein and wikitools),
nor does it signal an error if they are not installed.


User databases
==============

Some reports store information about the last run in tables in user
databases.  They can be safely purged and excluded from backup, but
their structures must always exist.

To (re-)create these, run as user dbreps:

$ USERDB=p_dbreps
$ sql -u enwiki_p <<EOF
DROP TABLE IF EXISTS $USERDB.bullshit_reviewed_page_titles;
CREATE TABLE $USERDB.bullshit_reviewed_page_titles
(page_id int(8) unsigned PRIMARY KEY,
 lastreviewed TIMESTAMP NOT NULL);

DROP TABLE IF EXISTS $USERDB.potenshblps4_skipped_pages;
CREATE TABLE $USERDB.potenshblps4_skipped_pages
(page_title VARCHAR(255) COLLATE utf8_bin PRIMARY KEY);
LOAD DATA LOCAL INFILE 'data/enwiki/potenshblps4/skipped_pages.txt'
INTO TABLE $USERDB.potenshblps4_skipped_pages (page_title);
EOF