v1.5.1 replaces 1.5.0 as pypi refused it for version naming purpose

This is exactly the same as v1.5.0...
With this new version, you can translate a whole django project at once. Use the new auto_djangopo_lyglot command for that
Also contains a lot of bug fixes + better tests

v1.5.0: translation of a whole django project at once

With this new version, you can translate a whole django project at once. Use the new auto_djangopo_lyglot command for that
Also contains a lot of bug fixes + better tests

What's Changed

• Leolivier/issue12 by @leolivier in #14
• When a po entry has plurals, the translation does not work by @leolivier in #15
• Tests by @leolivier in #16
• Fixed Tests Django and Plurals by @leolivier in #17

Full Changelog: v1.4.1...v1.5.0

v1.4.1: automatic build&push of the docker image to ghcr.io

No change for end users, the only change is that on new version delivery, the new docker image will be pushed to ghcr.io

Full Changelog: v1.4.0...v1.4.1

v1.4.0: Docker images for auto_po_lyglot

Auto_po_lyglot now supports Docker. See the README file for details

What's Changed

• Parameterized models per LLM for UI by @leolivier in #9
• Docker build by @leolivier in #10

Full Changelog: v1.3.3...v1.4.0

v1.3.2: publish content of the README file on PyPi

No change for users: only change is that the description of the package on PyPi will be the full content of the README.md file

Full Changelog: v1.3.1...v1.3.2

v1.3.1: removed sensitive trace

The v1.3.0 version print the Api keys for the commercial LLMs on the console

Full Changelog: v1.3.0...v1.3.1

v1.3.0: Streamlit UI

This release brings a brand new streamlit UI.
To use it, first install the latest version of the package: pip install 'auto-po-lyglot>=1.3.0'
Then create a short python script named auto_po_lyglot_ui.py containing these 2 lines:

from auto_po_lyglot.po_streamlit import streamlit_main
streamlit_main()

And finally run streamlit run auto_po_lyglot_ui.py

What's Changed

• new streamlit UI by @leolivier in #8

Full Changelog: v1.2.1...v1.3.0

v1.2.1: fixes an issue in the packaging deployment

v1.2.0 didn't create a package that can be uploaded on PyPi due to \n present in the package description.

v1.2.0: implements Gemini and Grok Clients

This new release implements a client for Gemini and for Grok.
Note: The Gemini client has been tested, but not the Grok client because the Grok API is not yet available in my country.
If you have problems with Grok, please open a GitHub issue and describe it, I'll try to fix it in a blind way.

What's Changed

• implement a client for Gemini by @leolivier in #6
• Leolivier/issue3 by @leolivier in #7

Full Changelog: v1.1.0...v1.2.0

v1.1.0: a bunch of major enhancements

This new version brings a bunch of major enhancements to the tool. The biggest ones are:

• A package named aut-po-lyglot is now available on PyPi. To install it (in a virtual env preferably), just run pip install auto-po-lyglot
• The tool can now be run directly through the auto_po_lyglot command line. Use auto_po_lyglot -h to see all the parameters and have a look to the README file to see the documentation.
• The prompt has been enhanced a lot. The default prompt is now internal to the python code but can still be overriden using .env. The prompt can be parameterized using different languages and translated sentences (see .env.example)
• Log management
• LLM clients enhancements and tuning
• Full doc in README.md
• new TEMPERATURE parameter in .env (or -t on command line)
• ...

What's Changed

• Pypi-packaging and a lot of other major enhancements by @leolivier in #4

New Contributors

• @leolivier made their first contribution in #4

Full Changelog: v1.0.0...v1.1.0

What still doesn't work:

• a lot of efforts to automate tests using Ollama but as tests are using small Ollama LLM (phi3 or gemma2:2b) and that LLMs always answer different things, it simply does not work :(
• also a lot of effort to manage a cache for the github workflow but still does not work :(