v1.10.2-lts

[shaohuzhang1]

New features

• Application (X-Pack): Application access now supports Slack;
• Application: Add variable assignment node; by @liuruibin in Variable assign #2346
• Function library: Support import/export functions; by @liuruibin in feat: Import and Export function lib #2294
• Model settings: Add support for Ollama supplier reordering model;

Enhancements

• Application (X-Pack): After starting the thinking process, optimize the reply process in WeChat for Business, Feishu, DingTalk, and public accounts;
• Application (X-Pack): Support uploading pictures during WeChat for Business conversations;
• Application (X-Pack): Support configuring scope parameters for login authentication OIDC settings;
• Application: Support creating blank applications; by @wangdan-fit2cloud in feat: Create workflow applications that support template selection #2377
• Application: Add support for password box and switch components for user input parameters; by @liuruibin in feat: Supports user input field type password #2307
• Application: User input supports custom titles; by @wangdan-fit2cloud in feat: Supports specified user input title(#2288) #2313
• Application: Add support for password box components for parameters of form collection nodes; by @liuruibin in feat: Supports user input field type password #2307
• Application: Add m4a format to the uploaded audio file type; by @wxg0103 in feat: audio support m4a #2277
• Application: Support uploading file parameters when conducting a conversation by calling the application API Key; by @liuruibin in refactor: Support image_list, document_list while using swagger api #2300
• Application: By calling the application API Key mode, support output of thinking process parameters; by @shaohuzhang1 in feat: The OpenAI interface supports the thought process #2392
• Application: The parameters in the user input table in the basic information node support drag-and-drop adjustment of the order;
• Application: The conditional value in the judgement supports variable parsing; by @shaohuzhang1 in feat: The discriminator supports variables #2311
• Application: In the "Execution Details" dialog box of the multi-way recall node, optimize the segment display content (including segment title, document and knowledge base); by @shaohuzhang1,@wangdan-fit2cloud in refactor: Multi channel recall, adding display of knowledge base titles and other information #2309
• Application: When the user exits the workflow editing, prompt the user to save the data; by @wangdan-fit2cloud in perf: Close editor save content #2312
• Application: The maximum number of reference segments of the associated knowledge base is adjusted to 10,000; by @shaohuzhang1 in fix: Reference segmentation limit #2487
• Application: The operation of modifying the node name in the advanced arrangement is changed to "..." → "Rename"; by @wangdan-fit2cloud in perf: Optimization of modifying node names in workflow #2432
• Application: When a single image generation node generates multiple images, the images are arranged horizontally; by @liuruibin in refactor: Support the generated images horizontal tiling display #2423
• Q&A page: Optimize the Q&A page layout to a left-right layout, with AI answers displayed on the left and user questions displayed on the right; by @wangdan-fit2cloud in fix: Display setting bugs #2258
• Q&A page: Optimize the display of historical conversation records when users open the Q&A page; by @wangdan-fit2cloud in feat: Chat page retains the display of the last conversation #2301
• Q&A page: Optimize voice playback to play only the last content; by @liuruibin in refactor: Cancel last autoplay audio #2431
• Q&A page: Support modifying the session title; by @shaohuzhang1 in feat: The demonstration page supports modifying dialogue summaries #2348
• Knowledge base: Support the continued generation of questions for the segment that failed to generate questions; by @shaohuzhang1 in feat: Generate problem support for generating unfinished paragraphs #2299
• Knowledge base: Support sorting the document list; by @liuruibin in feat: Support sorting feat for knowledge base documents #2284
• Knowledge base: Support using the {title} variable to obtain the segment title in the generated question; by @shaohuzhang1 in feat: Generate problem support for title variable #2310
• Function library: ignore case when querying functions; by @liuruibin in refactor: Search ignores capitalization #2279
• Model settings: ignore case when querying models; by @liuruibin in refactor: Search ignores capitalization #2279
• Team members: ignore case when querying members; by @liuruibin in refactor: Search ignores capitalization #2279
• System: optimize the problem of long time consumption when logging in for the first time. by @shaohuzhang1 in perf: enable gzip. #2414

Bug Fixes

• Application: Fixed the problem that the image in the AI ​​reply content cannot be enlarged; by @shaohuzhang1 in fix: Dialogue image cannot be enlarged #2327
• Application: Fixed the problem that the prompt word window in the advanced arrangement is not saved after editing the content after enlarging the prompt word window and pressing the ESC key to close the window; by @wangdan-fit2cloud in perf: Close editor save content #2312
• Application: Fixed the problem that the answer cannot be ended in some cases when using the vLLM supplier's large language model for dialogue; by @shaohuzhang1 in fix: VLLM supplier recalculates token function #2375
• Application: Fixed the problem that the tokens calculation is inaccurate when using the Kimi supplier's large language model for dialogue; by @shaohuzhang1 in fix: Tokens cannot be obtained from the model dialogue #2326
• Application: Fix the problem that the parameter setting is not updated after the image generation node switches the model;
• Application: Fix the problem that when the Excel table contains data from merged cells, the data is missing when reading; by @liuruibin in fix: Fix excel merge cells header #2265
• Application: Fix the problem that the OpenAI call format does not have the thinking process parameter; by @shaohuzhang1 in feat: The OpenAI interface supports the thought process #2392
• Application: Fix the problem that if there is a non-mandatory parameter in the application child node, if the parameter is not set in the parent application, an error will be reported during the dialogue; by @wxg0103 in fix: Defect of embedding application parameters as empty and reporting errors #2278
• Application: Fix the problem that the Ollama supplier's large language model uses the num_ctx parameter, and an error will be reported during the dialogue. by @itaa in fix: ollama model provider can not set 'num_ctx' etc. parameter #2442 #2444
• Q&A page: fix the problem that the browser cannot play voice in the historical conversation record; by @liuruibin in fix: Fix browser text to speech cannot resume and cancel #2378
• Q&A page: fix the problem that refreshing the page will delete the current conversation when asking the first question after creating a new conversation; by @shaohuzhang1 in fix: The workflow cannot function properly after the conversation flow is disconnected #2428
• Q&A page: fix the problem that line breaks cannot be inserted in the middle of the content in the question box; by @wangdan-fit2cloud in fix: Display setting bugs #2258
• Q&A page: fix the problem that "/" is automatically added after the URL address of the document in some cases, resulting in inaccessibility;
• Q&A page (X-Pack): Fix the problem that the new dialogue on the Q&A page is not displayed after turning off the history in the display settings;
• Function library: fix the problem that the output result during debugging is displayed incorrectly when the function return value is 0; by @wangdan-fit2cloud in fix: Resolve the issue of a function returning 0 as a result (#2389) #2413
• Model Settings: Fix the error when adding Alibaba Cloud Bailian's large language model, if it is a full-modal model submission; by @wxg0103 in fix: Alibaba Cloud full modal docking error #2302
• Model Settings: Fix the error when adding the DeepSeek -R1 model to the Azure OpenAI provider; by @liuruibin in refactor: add model_name parameter to AzureChatModel initialization #2483
• System Settings (X-Pack): Fix the problem that the interface parameters in the Swagger document are not fully displayed;
• System Settings (X-Pack): Fix the problem that the Q&A page does not display the application name after setting the website name in the appearance settings;
• System Settings (X-Pack): Fix the problem that the default logo flashes when the login page is loaded.

---

This discussion was created from the release v1.10.2-lts.