safety-screen/Nominatim
1
0
Fork 0
forked from hans/Nominatim
Code Pull Requests Activity

tabs-to-spaces

Browse Source
This commit is contained in:
Marc Tobias Metten
2016-09-04 03:19:48 +02:00
parent cd6dcfa574
commit 832547f192
55 changed files with 8236 additions and 8232 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
utils/import_wikipedia.sh
View File

@@ -11,32 +11,32 @@ echo "CREATE TABLE wikipedia_redirect (language text, from_title text, to_title
for i in "${language[@]}"
do
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-page.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-pagelinks.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-langlinks.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-redirect.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-page.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-pagelinks.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-langlinks.sql.gz
wget http://dumps.wikimedia.org/${i}wiki/latest/${i}wiki-latest-redirect.sql.gz
done
for i in "${language[@]}"
do
gzip -dc ${i}wiki-latest-pagelinks.sql.gz | sed "s/\`pagelinks\`/\`${i}pagelinks\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-page.sql.gz | sed "s/\`page\`/\`${i}page\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-langlinks.sql.gz | sed "s/\`langlinks\`/\`${i}langlinks\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-redirect.sql.gz | sed "s/\`redirect\`/\`${i}redirect\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-pagelinks.sql.gz | sed "s/\`pagelinks\`/\`${i}pagelinks\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-page.sql.gz | sed "s/\`page\`/\`${i}page\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-langlinks.sql.gz | sed "s/\`langlinks\`/\`${i}langlinks\`/g" | $mysql2pgsqlcmd | $psqlcmd
gzip -dc ${i}wiki-latest-redirect.sql.gz | sed "s/\`redirect\`/\`${i}redirect\`/g" | $mysql2pgsqlcmd | $psqlcmd
done
for i in "${language[@]}"
do
echo "create table ${i}pagelinkcount as select pl_title as title,count(*) as count from ${i}pagelinks where pl_namespace = 0 group by pl_title;" | $psqlcmd
echo "insert into linkcounts select '${i}',pl_title,count(*) from ${i}pagelinks where pl_namespace = 0 group by pl_title;" | $psqlcmd
echo "insert into wikipedia_redirect select '${i}',page_title,rd_title from ${i}redirect join ${i}page on (rd_from = page_id) where page_namespace = 0 and rd_namespace = 0;" | $psqlcmd
echo "alter table ${i}pagelinkcount add column othercount integer;" | $psqlcmd
echo "update ${i}pagelinkcount set othercount = 0;" | $psqlcmd
for j in "${language[@]}"
do
echo "update ${i}pagelinkcount set othercount = ${i}pagelinkcount.othercount + x.count from (select page_title as title,count from ${i}langlinks join ${i}page on (ll_from = page_id) join ${j}pagelinkcount on (ll_lang = '${j}' and ll_title = title)) as x where x.title = ${i}pagelinkcount.title;" | $psqlcmd
done
echo "insert into wikipedia_article select '${i}', title, count, othercount, count+othercount from ${i}pagelinkcount;" | $psqlcmd
echo "create table ${i}pagelinkcount as select pl_title as title,count(*) as count from ${i}pagelinks where pl_namespace = 0 group by pl_title;" | $psqlcmd
echo "insert into linkcounts select '${i}',pl_title,count(*) from ${i}pagelinks where pl_namespace = 0 group by pl_title;" | $psqlcmd
echo "insert into wikipedia_redirect select '${i}',page_title,rd_title from ${i}redirect join ${i}page on (rd_from = page_id) where page_namespace = 0 and rd_namespace = 0;" | $psqlcmd
echo "alter table ${i}pagelinkcount add column othercount integer;" | $psqlcmd
echo "update ${i}pagelinkcount set othercount = 0;" | $psqlcmd
for j in "${language[@]}"
do
echo "update ${i}pagelinkcount set othercount = ${i}pagelinkcount.othercount + x.count from (select page_title as title,count from ${i}langlinks join ${i}page on (ll_from = page_id) join ${j}pagelinkcount on (ll_lang = '${j}' and ll_title = title)) as x where x.title = ${i}pagelinkcount.title;" | $psqlcmd
done
echo "insert into wikipedia_article select '${i}', title, count, othercount, count+othercount from ${i}pagelinkcount;" | $psqlcmd
done
echo "update wikipedia_article set importance = log(totalcount)/log((select max(totalcount) from wikipedia_article))" | $psqlcmd