This tool consists on two scripts:

• tmxplore.py, a script to determine the language codes available inside a particular TMX file by looking to an excerpt or even the whole file.
• tmxt.py, to effectively transform a TMX to a tab-separated text file using the language code list provided in the command. TU propierties can also be printed if requested.

Requires python3 and the libraries included in requirements.txt

$ python3 tmxplore.py file.tmx
en es

or

$ cat file.tmx | python3 tmxplore.py
en es

$ python3 tmxt.py --codelist en,fr tm.fr-en.tmx en-fr.txt

Other

$ zcat largefile.tmx.gz | python3 tmxt.py --codelist en,es |gzip > bitext.en-es.gz

Printing TU propierties

$ python3 tmxt.py --codelist en,fr,prop1,prop2 tm.fr-en.tmx en-fr.txt

For propierties inside the TUV, the language code must be prepended

$ python3 tmxt.py --codelist en,fr,en-prop1,fr-prop1,prop2 tm.fr-en.tmx en-fr.txt

In case a propierty is repeated multiple times inside the same TU or TUV, the last instance will be printed. To get all of them, add --multiprops option and all the propierty instances will be printed in the same TSV field but encoded as a JSON array. The property can be easily parsed in pythin like

import json
json.loads(fields[3])