What's Changed

• The conversational retrieval functionality is now operating as expected. It successfully sends the conversation history to the language model, allowing the context from previous interactions to be utilized effectively. by @lvalics in #239
• Added support for Ollama as the Language Model (LLM). Ensure Ollama is specified in the .env configuration and the model is preloaded on the server. by @lvalics in #239

Full Changelog: v0.3.6...0.3.7-beta

What's Changed

• Incorporate 'Ollama' into your example.env configuration and make sure to reflect these changes in your .env file for compatibility. by @lvalics
• We've expanded the logging capabilities within settings.py by deploying logging.debug for more detailed insights, although it remains inactive when the DEBUG mode is off. by @lvalics
• We've refreshed the dependencies listed in requirements.txt to ensure the latest support and functionality. by @lvalics
• Updates to LangChain and OpenAI libraries, now feature the exciting addition of Ollama, currently under development. by @lvalics
• This version has a significant update that is still pending comprehensive testing; therefore, we advise against deploying it in production environments until it is fully vetted. by @lvalics
• Modifications have been made to the history module to append the input message to requests sent to OpenAI. These changes are still being tested. (disabled for the moment). by @lvalics
• Implemented translation fixes to enhance application localization. by @lvalics
• Updated docker-compose.yaml to prefix each container with oc_ for better namespace management. by @lvalics
• Ensured existence of Vector Database (QDrant) prior to web crawling operations to address issues encountered with large websites, ensuring Vector Database creation and availability. by @lvalics

What's Changed

• chat.js file will change based on where the server resides by @codebanesr in #145
• Update views_website_datasource.py by @mrtrantuan in #146
• Update Dockerfile by @mrtrantuan in #147
• Adding prebuilt images guide by @codebanesr in #148
• Bump certifi from 2023.5.7 to 2023.7.22 in /dj_backend_server by @dependabot in #151
• "PR: Add Support for llama-2-7b, Improve Execution Speed, and Clean Up Libraries" by @codebanesr in #150
• Enhancement/llama by @codebanesr in #153
• Adding ability to detect machine architecture + manually running migrations by @codebanesr in #157
• Update views_chatbot_settings.py by @lvalics in #166
• Update example.env.docker by @lvalics in #165
• Update settings.py by @lvalics in #164
• Update settings-theme.html by @lvalics in #167
• docs: add lvalics as a contributor for code, and doc by @allcontributors in #172
• PDF upload, require a table change. by @lvalics in #174
• Ivalics openchat by @codebanesr in #173
• Added a new field files_info and a new table for error_handling by @lvalics in #175
• The listing will check if the file(s) exist(s) and show the download … by @lvalics in #176
• small error. by @lvalics in #177
• TO DO list on dj_backend_server by @lvalics in #178
• Adding a custom command to create default super user + More enhancements by @codebanesr in #180
• Small improvements and cosmetics issues. by @lvalics in #181
• Adding nginx config for chatbot by @codebanesr in #185
• HTTPS working version. by @lvalics in #187
• Added dummy certs by @davidsmithdevops in #188
• CORS fixed and started to add APP_URL to replace in NGINX automatically the domain. by @lvalics in #189
• Some fixes by @lvalics in #191
• Staging by @davidsmithdevops in #192
• moved nginx generation to makefile. removed from yaml and Dockerfile. by @lvalics in #193
• Quick fix from datetime and also deactivated failed status, seems rai… by @lvalics in #194
• Replace URL also in search.js by @lvalics in #195
• Redirect to APP_URL if chatbot was deleted or not exist and delete data PDF by @lvalics in #196
• Correct http://0.0.0.0:8000/ URL by @lvalics in #200

New Contributors

• @mrtrantuan made their first contribution in #146
• @dependabot made their first contribution in #151
• @lvalics made their first contribution in #166
• @davidsmithdevops made their first contribution in #188

Full Changelog: 0.33...v0.3.5-beta

What's Changed

• Update README.md by @gharbat in #140
• Enhancement/chain selector by @codebanesr in #142
• updated readme for users, moved environment files inside to avoid con… by @codebanesr in #143
• prebuilt Docker images for our backend servers by @codebanesr in #144

Full Changelog: 0.3.3...0.33

Openchat v0.3.3-beta (August 11, 2023)

Added 🚀

• 🔥 Completely rewrote codebase in Django for improved performance and maintainability. Includes faster web scraping and PDF parsing.

• ➖ Removed dependency on llm-server microservice.

• ➕ Added support for Azure OpenAI and OpenAI APIs.

• ⚡️ Significantly reduced build times.

• 🛠️ Enhanced developer environment with VSCode debugging scripts.

• 🐇 Integrated Celery task queue and debugging.

• 🔃 Faster hot reloading for development.

• ♻️ Completely compatible with the prior version that uses Pinecone and OpenAI, although currently encounters issues when using with Qdrant vector store.

Known Issues

While testing and refinement is still underway, the product is ready for use. We will continuously improve and enhance the product in subsequent releases. Please contact us if any significant issues are encountered.

Special Notes

• The new Django codebase allows for easier customization and extensibility. See docs for details.

• With the addition of Celery, it's important to properly configure and run Celery workers for async task handling.

• A huge thank you to all our users who have provided feedback and support leading up to this milestone release! We couldn't have done it without you.

Documentation

For complete documentation on the new Django codebase, please see:

• Opechat Django Backend - Provides overview, setup instructions, API references, and customization guides for the Django backend services.

• Qdrant offline vector DB support
• Azure OpenAI support
• General fixes

• Introducing the new chat bubble!

Fixes:
#106
#100
#60
#49

• 🎉 🎉 🎉 🎉 🎉 Introducing OpenChat Search, integrate AI in your normal search button. 🎉 🎉 🎉 🎉 🎉
• Now you can customize the initial prompt via one click, this will open many possioblites for your chatbots, for example, you can ask the chatbot to answer in a certain language or certian style
• We added pre-made tempalates for your inital prompt, in one click, your chatbot can act as a search engine, ghost writer, pair programmer, and more!
• We added the history feature, now you can see all the new conversaions in the history tab.
• We fixed a bug that prevented some users from deleting their bots. So now you can delete your current bots with no problems
• We fixed a bug that prevented users from embedding their chatbots on theri web apps, it should work smoothly!
• We did some desgin enhancments in the settings page to make it cleaner.
• After many requests, we now have simple pricing plans! we think it's unbeatable, in returun, you will have access on moreik 8jm features on the cloud based verison, and a lot more capacity (like a lot), the free package still there!
• Now, we have 5000 bots, and we cralwed more than 30000 pages! thanks for the support!
• Zapier integration is available via invite only, connect your chatbots with thousensd of applications such as Slack, intercome, Zendesk, gmail and many. so you can litrally run openchat every where!

Thanks for the awesome contributors:

Ikko Eltociear Ashimine 🤔 💻

Joshua Sindy 🐛

Erjan Kalybek 📖

WoahAI 🐛 💻

• A lot of refactoring in the LLM server
• Introduce full code base ingesting