Skip to content

Releases: Cyb3rWard0g/floki

0.12.2 - Enhanced Agentic Workflows, Structured Outputs, and Dynamic Execution

09 Mar 23:05
Compare
Choose a tag to compare

What's Changed

  • Add OpenAI Structured Output via JSON Schema & Enhance Chat Client Flexibility by @Cyb3rWard0g in #27
  • Added SentenceTransformers Notebook to cache model downloaded by @Cyb3rWard0g in #28
  • Enable Dynamic State Format for Agentic Workflows & Improve Workflow Initialization by @Cyb3rWard0g in #29
  • Refactored LLM Orchestrator with Dynamic Workflow State, Dapr Monitor Pattern, and Enhanced Execution Control by @Cyb3rWard0g in #30
  • Enhanced Agentic Framework with Workflow-Based Agents, Improved Task and Workflow registration, Dynamic Topic Routing, and Code Execution Capabilities by @Cyb3rWard0g in #32
  • Optimized Assistant Workflows & Multi-Agent Coordination by @Cyb3rWard0g in #33
  • Fixing Agent Assistant Workflow (ToolCalling) Logic by @Cyb3rWard0g in #34

Full Changelog: 0.11.2...0.12.2

0.11.2 - SentenceTransformer Local Cache Directory

23 Jan 06:36
Compare
Choose a tag to compare

0.10.2 - Minor Fixes on Package Dependencies

22 Jan 12:20
Compare
Choose a tag to compare

What's Changed

  • Added lazy import of arxiv library and logging on structured LLM response by @Cyb3rWard0g in #25
  • Updated project dependencies by @Cyb3rWard0g in #26

Full Changelog: 0.10.1...0.10.2

0.10.1 - ElevenLabs Integration, Neo4j Enhancements, and ArxivFetcher Refactor

14 Jan 17:36
7f44599
Compare
Choose a tag to compare

This release introduces the integration of [ElevenLabs' Text-to-Speech client](https://elevenlabs.io/docs/api-reference/text-to-speech/convert, complete with a demonstration notebook for hands-on exploration. Neo4j enhancements include batch operations, vector indexing, and an example notebook for advanced data management. The ArxivFetcher refactor improves metadata handling and code reusability, streamlining workflows and increasing efficiency.

What's Changed

  • Integrate ElevenLabs Text-to-Speech Client and Add Notebook for Demonstration by @Cyb3rWard0g in #21
  • Enhance Neo4j Integration with Batch Operations, Vector Indexing, and Example Notebook by @Cyb3rWard0g in #22
  • Refactor ArxivFetcher to Support Flexible Metadata Handling and Improve Code Reusability by @Cyb3rWard0g in #23

Full Changelog: 0.9.1...0.10.1

0.9.1 - NVIDIA NIM Chat and Embeddings API Integration

12 Jan 06:54
391dc33
Compare
Choose a tag to compare

This release introduces integration with NVIDIA NIM (Inference Microservice), enabling Floki to leverage NVIDIA's Chat and Embeddings APIs. NIM microservices offer standardized APIs, enabling effortless integration into AI-driven applications and workflows.

What's Changed

  • Integration of NVIDIA LLM Client with Enhanced Structured Response Support by @Cyb3rWard0g in #18
  • Integration of NVIDIA Embedding Client and Embedder by @Cyb3rWard0g in #19
  • Fix Structured Output Parsing for Chat Completion Responses and Prepare for v0.9.1 Release by @Cyb3rWard0g in #20

Full Changelog: 0.8.1...0.9.1

0.8.1 - Embedding Functions, PostgreSQL Integration, and LLM Client Updates

08 Jan 12:15
Compare
Choose a tag to compare

This release introduces several key features, including a new Embedding Functions class, PostgreSQL vector storage integration, and significant updates to Floki's OpenAI and Hugging Face client support. These changes improve functionality, add new features, and enhance the integration of core components for streamlined workflows.

What's Changed

  • Update README.md to indicate admin required for dapr init by @knifeyspoony in #13
  • Default azure_ad_token_provider to None by @knifeyspoony in #14
  • Updated Floki OpenAI and Hugging Face LLM Clients and Provided Notebook Updates by @Cyb3rWard0g in #15
  • Enhanced Embedding Functions and Updated Integrations by @Cyb3rWard0g in #16
  • Initial Integration of PostgreSQL Vector Store with Default SentenceTransformer Embedding Function by @Cyb3rWard0g in #17

Full Changelog: 0.7.1...0.8.1

0.7.1 - Dependency Fixes and Setup Updates

02 Jan 04:43
Compare
Choose a tag to compare

This patch release fixes missing dependencies in the project setup, ensuring all required modules are properly included. Updates were made to requirements.txt, setup.py, and pyproject.toml for seamless installation and improved reliability.

Full Changelog: 0.7.0...0.7.1

0.7.0 - Enhanced Arxiv Module, Workflow Refinements, and Core Component Upgrades

31 Dec 22:44
Compare
Choose a tag to compare

This release introduces the Arxiv Module for seamlessly fetching and processing research papers from Arxiv, coupled with significant enhancements to workflow task execution, agent factories, and storage modules. Key improvements include robust handling of async/sync tasks, lazy initialization of LLMs and memory, and streamlined initialization for Chroma-based vector stores. These updates enhance functionality, maintainability, and developer experience, paving the way for more advanced document processing and AI workflows.

What's Changed

  • Integrate Arxiv Module and Refactor Core Components for Improved Functionality by @Cyb3rWard0g in #12

Full Changelog: 0.6.0...0.7.0

0.6.0 - Document Processing with PDF Readers, Text Splitters, and OpenAI Audio Integration

27 Dec 07:47
Compare
Choose a tag to compare

This release introduces tools for document processing, including PDF readers and text splitters designed to prepare text for Retrieval-Augmented Generation (RAG) workflows. With the integration of the OpenAI Audio Client, Floki now enables seamless conversion of text into speech/audio. As a practical demonstration of these features, a basic Doc2Podcast proof of concept (POC) is included, showcasing how these tools can be combined to transform research documents into podcast-style audio content.

What's Changed

  • Integrate OpenAI Audio Client and Introduce ChatClientBase for Enhanced Separation of Concerns by @Cyb3rWard0g in #9
  • PDF Readers, Text Splitter, and Document Indexing Features by @Cyb3rWard0g in #10
  • Doc2Podcast POC, Enhanced Workflow Validation, and Documentation Updates by @Cyb3rWard0g in #11

Full Changelog: 0.5.0...0.6.0

0.5.0 - Enhanced Message Metadata for Event-Driven Workflows

07 Dec 03:57
12205ea
Compare
Choose a tag to compare

This release focuses on improving message metadata encapsulation in Floki's messaging framework, introducing a unified way to manage and process contextual information within event-driven workflows.

What's Changed

  • Enhance Messaging Framework with EventMessageMetadata Encapsulation by @Cyb3rWard0g in #8

Full Changelog: 0.4.0...0.5.0