diff options
author | Vincent Ambo <mail@tazj.in> | 2023-01-21T18·17+0300 |
---|---|---|
committer | tazjin <tazjin@tvl.su> | 2023-01-22T16·13+0000 |
commit | ed8dd4acd71161893a06dd25567852b1855ac1ab (patch) | |
tree | 5dd998379e20c6601ef734b2fd15ef99a65ef3c6 /corp/russian/data-import/src/db_setup.rs | |
parent | 2b308c64b94a14592f928a7d2511fc74c8846eb3 (diff) |
feat(corp/data-import): add import of OR 'translations' table r/5732
The original dataset contains translations into different languages, but only the English ones are imported here. Note that translations are for lemmata only. Change-Id: Ifb9c32c25fda44c38ad899efca9d205c520c0fa3 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7895 Reviewed-by: tazjin <tazjin@tvl.su> Tested-by: BuildkiteCI
Diffstat (limited to 'corp/russian/data-import/src/db_setup.rs')
-rw-r--r-- | corp/russian/data-import/src/db_setup.rs | 44 |
1 files changed, 44 insertions, 0 deletions
diff --git a/corp/russian/data-import/src/db_setup.rs b/corp/russian/data-import/src/db_setup.rs index 4644edf09424..21cfb2a4652e 100644 --- a/corp/russian/data-import/src/db_setup.rs +++ b/corp/russian/data-import/src/db_setup.rs @@ -191,6 +191,16 @@ CREATE TABLE or_words_forms ( form_bare TEXT, FOREIGN KEY(word_id) REFERENCES words(id) ) STRICT; + +CREATE TABLE or_translations ( + id INTEGER PRIMARY KEY, + word_id INTEGER NOT NULL, + translation TEXT, + example_ru TEXT, + example_tl TEXT, + info TEXT, + FOREIGN KEY(word_id) REFERENCES words(id) +) STRICT; "#, ) .ensure("setting up OpenRussian table schema failed"); @@ -252,3 +262,37 @@ VALUES (?1, ?2, ?3, ?4, ?5, ?6) info!("inserted {} OpenRussian word forms", count); } + +pub fn insert_or_translations<I: Iterator<Item = or_parser::Translation>>( + conn: &Connection, + translations: I, +) { + let mut stmt = conn + .prepare_cached( + "INSERT INTO or_translations (id, word_id, translation, example_ru, example_tl, info) + VALUES (?1, ?2, ?3, ?4, ?5, ?6)", + ) + .ensure("failed to prepare OR translation statement"); + + let mut count = 0; + + for tl in translations { + if tl.lang != "en" { + continue; + } + + stmt.execute(( + tl.id, + tl.word_id, + tl.tl, + tl.example_ru, + tl.example_tl, + tl.info, + )) + .ensure("failed to insert OR translation"); + + count += 1; + } + + info!("inserted {} OpenRussian translations", count); +} |