Skip to content

v0.8.3
Compare
Choose a tag to compare
@laipz8200 laipz8200 released this 20 Sep 10:25
· 683 commits to main since this release
0.8.3
7f3282e
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Learn about vigilant mode.

✨ What’s New in v0.8.3? ✨

Hey everyone! We’re thrilled to announce the release of version 0.8.3, and it brings something truly game-changing:

🚀 Major Feature Highlight

  • Qwen 2.5 Series Models: The highlight of this release is the integration of the Qwen 2.5 Series models, seamlessly synced with Aliyun Bailian, brought to you by @soulteary in #8538. This major update significantly expands our model offerings, providing enhanced capabilities and greater flexibility for your projects. Whether you're working with AI-driven workflows or model-intensive tasks, the Qwen 2.5 Series offers unmatched performance and versatility.

🛠️ Additional Features & Improvements

In addition to the powerful Qwen models, we’ve also made several other improvements and fixes to optimize your experience:

  • New Tools:
    • ComfyUI for Stable Diffusion: Enhanced image generation capabilities with the integration of ComfyUI by @QunBB in #8160.
    • Siliconflow Image-Gen Tool: Part of the flux dev suite, this tool, introduced by @hjlarry in #8450, further expands your creative toolkit.
  • Timeout Configuration: Thanks to @kurokobo in #8395, you now have full control over text generation and workflow timeouts via environment variables, giving you more flexibility in managing process execution times.

⚙️ Enhancements and Fixes

We've also worked on various optimizations and bug fixes to improve overall stability:

  • Image Preview Optimization: Enhanced SVG rendering and image previews for a smoother user experience, thanks to @charli117 in #8387.
  • Workflow Improvements: Including better result handling and improved multi-retrieval for datasets.
  • Bug Fixes: Numerous fixes across authorization, token usage, and vector search methods to ensure smoother operations.

This release is all about delivering cutting-edge capabilities with the Qwen 2.5 Series models and refining the platform with additional tools and fixes. Huge thanks to our incredible contributors for making this version happen! 🚀

Upgrade Guide

Docker compose deployments

Warning

The docker-compose.yaml has been refactored. If you've made any changes to the file, make sure to check out the "Upgrade to new docker compose deployment" section above for usage and migration tips.

  1. Back up your customized docker-compose YAML file (optional)

    cd docker
cp docker-compose.yaml docker-compose.yaml.$(date +%s).bak

  2. Get the latest code from the main branch

    git checkout main
git pull origin main

  3. Stop the service，Command, please execute in the docker directory

    docker compose down

  4. Back up data

    tar -cvf volumes-$(date +%s).tgz volumes

  5. Upgrade services

    docker compose up -d

Source Code deployments

  1. Stop API server, Worker and Web frontend Server.

  2. Get the latest code from the main branch:

    git checkout main
git pull origin main

  3. Update Python dependencies:

    cd api
poetry install

  4. Then, let's run the migration script:

    poetry shell
flask db upgrade

  5. Finally, run API server, Worker and Web frontend Server again.

What's Changed

  • fix: ark token usage is none by @sinomoe in #8351
  • fix: score_threshold handling in vector search methods by @laipz8200 in #8356
  • chore: fix unnecessary string concatation in single line by @bowenliang123 in #8311
  • fix: o1-series model encounters an error when the generate mode is blocking by @takatost in #8363
  • fix: when edit load balancing config not pass the empty filed value hidden by @iamjoel in #8366
  • fix: add before send to remove langfuse defaultErrorResponse by @ZhouhaoJiang in #8361
  • fix: edit load balancing not pass id by @iamjoel in #8370
  • fix(workflow): both parallel and single branch errors occur in if-else by @takatost in #8378
  • chore: update firecrawl scrape to V1 api by @hjlarry in #8367
  • support hunyuan-turbo by @xiandan-erizo in #8372
  • chore: refurish python code by applying Pylint linter rules by @bowenliang123 in #8322
  • fix (#8322 followup): resolve the violation of pylint rules by @bowenliang123 in #8391
  • chore: improve usage of striping prefix or suffix of string with Ruff 0.6.5 by @bowenliang123 in #8392
  • fix: sandbox issue related httpx and requests by @Yeuoly in #8397
  • fix: o1-mini 65563 -> 65536 by @crazywoola in #8388
  • fix: pyproject.toml typo by @ZuzooVn in #8396
  • fix(docker): aliyun oss path env key by @incca in #8394
  • Fix: Support Bedrock cross region inference #8190 (Update Model name to distinguish between different region groups) by @ybalbert001 in #8402
  • fix: resolve runtime error when self.folder is None by @swingchen01 in #8401
  • fix:logs and rm unused codes in CacheEmbedding by @Howe829 in #8409
  • fix(workflow): the answer node after the iteration node containing the answer was output prematurely by @takatost in #8419
  • fix: when the variable does not exist, an error should be prompted by @HiChen404 in #8413
  • Fix: operation postion of answer in logs by @JzoNgKVO in #8411
  • feat: allow users to specify timeout for text generations and workflows by environment variable by @kurokobo in #8395
  • fix: sys_var startwith 'sys.' not 'sys' #8421 by @yanxiyue in #8422
  • chore: update Jina embedding model by @DresAaron in #8376
  • Fix: #8190 (Update Model nam… by @ybalbert001 in #8426
  • fix: delete the delay for the tooltips inside the add tool panel by @YIXIAO0 in #8436
  • chore(workflow): Optimize the iteration when selecting a variable from a branch in the output variable causes iteration index err by @takatost in #8440
  • Bugfix/fix feishu plugins by @hgnulb in #8443
  • fix: Improve scrolling behavior for Conversation Opener by @ZuzooVn in #8437
  • add WorkflowClient.get_result, increase version number by @twisker in #8435
  • add svg render & Image preview optimization by @charli117 in #8387
  • fix:docker-compose.middleware.yaml start the Weaviate container by default (#8446) by @hwzhuhao in #8447
  • multi-retrival use dataset's top-k by @JohnJyong in #8416
  • fix: internal error when tool authorization by @takatost in #8449
  • [Python SDK] Add KnowledgeBaseClient and the corresponding test cases. by @twisker in #8465
  • fix(RunOnce): change to form submission instead of onKeyDown and onClick by @MH4GF in #8460
  • chore: workflow BRANCH, PARALLEL i18n by @ZuzooVn in #8452
  • feat: add flux dev of siliconflow image-gen tool by @hjlarry in #8450
  • feat: Add base URL settings and secure_ascii options to the Brave search tool by @XiaoLey in #8463
  • chore: update the .gitignore file to include opensearch,pgvector,and myscale by @hwzhuhao in #8470
  • feat: Add ComfyUI tool for Stable Diffusion by @QunBB in #8160
  • chore: Deprecate gpt-3.5-turbo-0613 and gpt-3.5-turbo-16k-0613 models by @laipz8200 in #8500
  • refactor: rename task_type to task for jina embeddings v3 by @bwanglzu in #8488
  • validate user permission before enter app detail page by @douxc in #8527
  • feat: add format util unit and add pre-commit unit check by @LLmoskk in #8427
  • fix: ci issues(missing duckduckgo-search==6.2.11, ruff lint issue) by @takatost in #8543
  • fix: thread_pool submit count in parallel workflow not releasing by @takatost in #8549
  • feat: sync Qwen API with Aliyun Bailian by @soulteary in #8538
  • add storage error log by @JohnJyong in #8556
  • chore: improve delimiter by @iamjoel in #8552
  • feat: add hunyuan-vision by @muzea in #8529
  • fix: fix qwen series model type by @soulteary in #8580
  • Fix: update qwen model and model config by @soulteary in #8584
  • ComfyUI tool use the new internal enumeration class "VariableKey" by @QunBB in #8533
  • chore: fix webpack dependencies order by @hjlarry in #8542
  • Update version to 0.8.3 in packaging and docker-compose files by @laipz8200 in #8590

New Contributors

Full Changelog: 0.8.2...0.8.3

Contributors

ZuzooVn, soulteary, and 32 other contributors
Assets 2
Loading
0 Join discussion