add backtrans for regex pipeline

in test_regex.py, add ability to do backtrans, first trans to english, perform ner, then translate back to target_lang, and match the ner entity between back transed text and original text.

Start in test_regex.py and read the existing regex pipeline to identify its translation and NER entry points. Trace the current test flow, then verify that it translates to English, performs NER, translates back to target_lang, and matches entities between the back-translated and original text.

Written by the indexing model from the issue text.