anthrove/fuzzysearch
4
0
Fork 0
mirror of https://github.com/Syfaro/fuzzysearch.git synced 2024-11-23 23:32:32 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix schema name.

Browse Source
This commit is contained in:
Syfaro 2020-01-21 14:50:47 -06:00
parent 880c5f385e
commit ceeef0b9e9
4 changed files with 20 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
src/bin/import.rs
View File

@ -66,7 +66,7 @@ async fn main() {
}); });
db.execute( db.execute(
"CREATE TABLE IF NOT EXISTS post (id INTEGER PRIMARY KEY, data JSONB)", "CREATE TABLE IF NOT EXISTS e621 (id INTEGER PRIMARY KEY, hash BIGINT, data JSONB, hash_error TEXT)",
&[], &[],
) )
.await .await
@ -84,13 +84,13 @@ async fn main() {
.await .await
.expect("Unable to create function"); .expect("Unable to create function");
db.execute("DROP TRIGGER IF EXISTS call_extract_post_data ON post", &[]) db.execute("DROP TRIGGER IF EXISTS call_extract_post_data ON e621", &[])
.await .await
.expect("Unable to drop trigger"); .expect("Unable to drop trigger");
db.execute("CREATE TRIGGER call_extract_post_data BEFORE INSERT ON post FOR EACH ROW EXECUTE PROCEDURE extract_post_data()", &[]).await.expect("Unable to create trigger"); db.execute("CREATE TRIGGER call_extract_post_data BEFORE INSERT ON e621 FOR EACH ROW EXECUTE PROCEDURE extract_post_data()", &[]).await.expect("Unable to create trigger");
let mut min_id = db let mut min_id = db
.query_one("SELECT MIN(id) FROM post", &[]) .query_one("SELECT MIN(id) FROM e621", &[])
.await .await
.map(|row| row.get("min")) .map(|row| row.get("min"))
.expect("Unable to get min post"); .expect("Unable to get min post");
@ -109,7 +109,7 @@ async fn main() {
min_id = ids.into_iter().min(); min_id = ids.into_iter().min();
db.execute( db.execute(
"INSERT INTO post (data) SELECT json_array_elements($1::json)", "INSERT INTO e621 (data) SELECT json_array_elements($1::json)",
&[&post_data], &[&post_data],
) )
.await .await

6
src/bin/load_hashes.rs
View File

@ -45,7 +45,7 @@ async fn load_next_posts(
id, id,
data->>'file_url' file_url data->>'file_url' file_url
FROM FROM
post e621
WHERE WHERE
hash IS NULL AND hash IS NULL AND
hash_error IS NULL AND hash_error IS NULL AND
@ -109,7 +109,7 @@ async fn main() {
db.get() db.get()
.await .await
.unwrap() .unwrap()
.execute("UPDATE post SET hash = $2 WHERE id = $1", &[&id, &num]) .execute("UPDATE e621 SET hash = $2 WHERE id = $1", &[&id, &num])
.await .await
.expect("Unable to update hash in database"); .expect("Unable to update hash in database");
} }
@ -121,7 +121,7 @@ async fn main() {
.await .await
.unwrap() .unwrap()
.execute( .execute(
"UPDATE post SET hash_error = $2 WHERE id = $1", "UPDATE e621 SET hash_error = $2 WHERE id = $1",
&[&id, &desc], &[&id, &desc],
) )
.await .await

12
src/bin/query_hash.rs
View File

@ -58,7 +58,7 @@ async fn main() {
artists_agg.artists artists, artists_agg.artists artists,
sources_agg.sources sources sources_agg.sources sources
FROM FROM
post, e621,
LATERAL ( LATERAL (
SELECT array_agg(v) artists FROM jsonb_array_elements_text(data->'artist') v SELECT array_agg(v) artists FROM jsonb_array_elements_text(data->'artist') v
) artists_agg, ) artists_agg,
@ -85,9 +85,17 @@ async fn main() {
}); });
for row in rows { for row in rows {
let dist = row.distance.unwrap_or_else(u64::max_value);
if dist > 5 {
println!(
"Skipping https://e621.net/post/show/{}, distance too high: {}",
row.id, dist
);
continue;
}
println!( println!(
"Possible match: [distance of {}] https://e621.net/post/show/{} by {}", "Possible match: [distance of {}] https://e621.net/post/show/{} by {}",
row.distance.unwrap_or_else(u64::max_value), dist,
row.id, row.id,
row.artists row.artists
.map(|artists| artists.join(", ")) .map(|artists| artists.join(", "))

4
src/bin/update.rs
View File

@ -63,7 +63,7 @@ async fn main() {
}); });
let max_id: i32 = db let max_id: i32 = db
.query_one("SELECT max(id) FROM post", &[]) .query_one("SELECT max(id) FROM e621", &[])
.await .await
.map(|row| row.get("max")) .map(|row| row.get("max"))
.expect("Unable to get max post"); .expect("Unable to get max post");
@ -85,7 +85,7 @@ async fn main() {
min_id = ids.into_iter().min(); min_id = ids.into_iter().min();
db.execute( db.execute(
"INSERT INTO post (data) SELECT json_array_elements($1::json) ON CONFLICT DO NOTHING", "INSERT INTO e621 (data) SELECT json_array_elements($1::json) ON CONFLICT DO NOTHING",
&[&post_data], &[&post_data],
) )
.await .await