This project aims at identifying the most common queries on Tatoeba. For this purpose, an annual log of the Tatoeba search engine is analyzed.

• the queries are grouped by language
• the queries that are obviously generated by bots are discarded
• the complex queries that include special characters are discarded, except for those with a simple pattern like ="hello world"
• the queries that can't be found in the language lexicon (which contains a maximum of 100,000 entries) are discarded
• for some languages, the lower case, upper case and title case variants of the queries are also tested after no exact match was found in the lexicon
• for some languages, the queries are spell checked
• the occurrences of the queries are counted and these counts are sorted by descending order

The Tatoeba most common queries are available here as CSV files for 150 languages.

Download the Tatoeba search annual log here.