samuelacombe / tmxsql

TmxSql enables localization engineers to create a Centralized TM Repository from TMX files that they can query using SQL and Regular Expressions.

TmxSql

TmxSql enables localization engineers to work with translation memories (TM) using SQL. You can create a Centralized TM Repository from TMX files that they can query using SQL. The TMX file format, Translation Memory Exchange, is a standard in XML format used to transfer translation data.

For more information on the TMX file format:
https://en.wikipedia.org/wiki/Translation_Memory_eXchange
https://www.gala-global.org/tmx-14b

Getting Started: Use Git to pull down a clone of TmxSql. Here is a quick tutorial. You will need the following files:

• tmxdb.sqlite: Sqlite Database (an empty one will work - a sample is included)
• sql2tmx.py: Python script to import your tmx files.
• enfr_test.tmx: Sample TMX file
• sqlite-manager: FireFox plug in manage sqlite databases right in your browser

Import your TMX Files: Type the following command to import your files into sqlite:

sql2tmx.py enfr_test.tmx

Connect to the Sqlite Database:

1. Add the Sqlite Manager Plugin to FireFox
2. Connect to the database file 'tmxdb.sqlite'

Query the TM Repostiory:

From there you can use sql to query and update your translation memories. Here is an example:

SELECT * FROM TranslationUnits

Some useful queries are provided in 'tmx2sql_useful_queries.sql'

Note: This is my class project for a computer science course at Portland State University, CS 561: Open Source Software Development taught by Bart Massey. https://github.com/psu-oss-2017/psu-oss-2017.github.io/wiki

Acknowledgements: This project was inspired by Gert Van Asshe from Datamundi. Gert has been doing work on TMX using SQL. This project is a result of conversations with Gert. https://www.linkedin.com/in/gertvanassche

Status

Can be used to effectively manage bilingual TMX files using SQL queries. Not ready yet to import TMX files with 3 or more languages.

To Do:

• Import: Allow import of TMX files with multiple target segments for each translation unit.
• Import: Define interfaces to abstract db layer and decouple Sqlite dependency.
• Import: argument: allow *.tmx on DOS
• Import: add optional argument: --dbname
• Import: add optional argument: --company Specify a company name
• Import: add optional argument: --domain Specify a domain if known
• Import: Try catch for exceptions with db transactions
• Import: Break up code into smaller modules
• Testing: Create python unit tests, test on linux
• SQL: Move Source/Target language columns to column headers in views.
• SQL: Add to property table - any other meta data? Header?
• SQL: Performance: Indexes to improve performance, defrag existing db file
• SQL: More sql views
• SQL: Table or ini file of regexps to clean and dedupe data