2284 Commits

Author	SHA1	Message	Date
shaohuzhang1	977019b2bd	fix: Application workflow, after the loop ends, there is a recall node that cannot have a conversation (#4319)	2025-11-10 15:21:22 +08:00
shaohuzhang1	fa1aee6c3d	perf: Memory optimization (#4318)	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	3445e05aea	feat: add MAXKB_SANDBOX_PYTHON_BANNED_HOSTS env to ban host for sandbox in tools code.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	aed98a5906	feat: add MAXKB_SANDBOX_PYTHON_BANNED_HOSTS env to ban host for sandbox in tools code.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	e4b90c6e96	feat: add MAXKB_SANDBOX_PYTHON_BANNED_HOSTS env to ban host for sandbox in tools code.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	1186752027	feat: add MAXKB_SANDBOX_PYTHON_BANNED_HOSTS env to ban host for sandbox in tools code.	2025-11-10 15:21:22 +08:00
CaptainB	f417a5aa39	revert: add Gunicorn preload option and initialize shared resources ... This reverts commit 1af14bdb27.	2025-11-10 15:21:22 +08:00
CaptainB	7a9c67d5df	fix: add empty environment variable dictionary for subprocess execution	2025-11-10 15:21:22 +08:00
CaptainB	63beec4dfb	feat: add Gunicorn preload option and initialize shared resources	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	568cc74ae3	fix: remove env from sandbox.	2025-11-10 15:21:22 +08:00
CaptainB	45f97fceb9	revert: add subprocess handling for local model and conditional post handler execution ... This reverts commit 8e3550eb83.	2025-11-10 15:21:22 +08:00
zhangzhanwei	2814e48be1	feat: Folder move order by name	2025-11-10 15:21:22 +08:00
zhangzhanwei	ba4e0b77f5	feat: Folder move	2025-11-10 15:21:22 +08:00
CaptainB	e497ce271f	feat: add subprocess handling for local model and conditional post handler execution	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	3c8ca222bd	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	70e4f5f467	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	d35cf16031	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	41b60ed53c	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	5fcf122003	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	00acd372db	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	5ea3023bde	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	e84391b4c4	perf: revert preload.	2025-11-10 15:21:22 +08:00
liqiang-fit2cloud	7b0df86578	perf: revert preload.	2025-11-10 15:21:22 +08:00
CaptainB	b36faff714	chore: update document deletion logic to handle paragraph IDs and related mappings ... --bug=1063248 --user=刘瑞斌 【知识库】删除文档时，分段与问题的关联关系没有清除 https://www.tapd.cn/62980211/s/1793500	2025-11-03 16:38:21 +08:00
王旭辉	d636d6ea7a	fix(vector): 修复 source_type 判断类型问题 ... - 将 source_type 转换为字符串后再与枚举值比较 - 避免了类型不匹配导致的判断错误 - 确保了数据分块功能的正常运行	2025-11-03 10:44:21 +08:00
shaohuzhang1	fcd258d05c	fix: The maximum limit for application dialogue has been modified to 10000000 times (#4302)	2025-11-03 10:20:20 +08:00
shaohuzhang1	a0d3031b4f	fix: Unable to export dialogue log data due to error (#4298)	2025-10-31 14:00:06 +08:00
zhangzhanwei	2a15017523	fix: XF stt model	2025-10-31 13:49:24 +08:00
zhangzhanwei	d9bda9679d	fix: Migration	2025-10-30 17:23:50 +08:00
CaptainB	d0cd6d657a	refactor: replace get_embedding_model_default_params with get_model_default_params	2025-10-30 17:12:30 +08:00
CaptainB	1b08643b98	fix: enhance embedding model retrieval with default parameters	2025-10-30 17:05:23 +08:00
CaptainB	ed19db07d1	feat: add function to retrieve default parameters for embedding models ... --bug=1063177 --user=刘瑞斌 【知识库】-知识库使用的模型更换维度参数值并重新向量化后，命中测试、检索报错 https://www.tapd.cn/62980211/s/1792117	2025-10-30 17:00:38 +08:00
wxg0103	2de6bd2018	chore: update file hashes for community	2025-10-30 16:24:21 +08:00
zhangzhanwei	aba41c4a4b	fix: Migrate file	2025-10-30 15:38:41 +08:00
wxg0103	ab434b0a9e	fix: add retry mechanism for file uploads and response handling in image.py	2025-10-30 15:34:41 +08:00
zhangzhanwei	274dfdae5c	fix: Model get in migration file	2025-10-30 14:27:46 +08:00
CaptainB	543f83a822	refactor: update ZhipuChatModel to use BaseChatOpenAI and improve token counting ... --bug=1061305 --user=刘瑞斌 【应用】ai对话启用工具后部分模型（智谱）不统计tokens https://www.tapd.cn/62980211/s/1791683	2025-10-30 11:12:43 +08:00
zhangzhanwei	ef162bdc3e	fix: Folder permission of workspace manager	2025-10-30 10:12:24 +08:00
zhangzhanwei	c86306a67e	fix: Common user folder permission	2025-10-29 18:32:09 +08:00
shaohuzhang1	2298bf6992	fix: After enabling identity authentication in the application, the knowledge base retrieves the reference document list variable, but there is no content retrieved when asking questions on the application dialogue page (#4286)	2025-10-29 15:53:37 +08:00
zhangzhanwei	2cf1ad7ea4	fix: In the variable aggregation strategy that returns the first non-null value of each group, empty strings, empty lists, and empty objects are not treated as null values	2025-10-29 15:34:34 +08:00
CaptainB	a4b7b8dbbb	fix: improve error logging in tools.py to include exception info	2025-10-29 15:08:57 +08:00
shaohuzhang1	28001613f8	fix: When the workflow does not respond, there is an error in the historical chat record data (#4282)	2025-10-29 13:43:02 +08:00
shaohuzhang1	2151c51f78	fix: Variable assignment data type converted (#4279)	2025-10-28 18:32:11 +08:00
wangxuhui	57195f277a	fix(permission): 修复权限常量解析操作符索引错误 (#4275) ... - 修正了权限字符串解析时操作符索引，从原来的第三部分改为第二部分 - 确保权限实例创建时操作符获取正确 - 改善权限常量解析的准确性和稳定性 Co-authored-by: 王旭辉 <2569429256@qq.com>	2025-10-28 18:31:31 +08:00
wxg0103	82224271e8	fix: update video handling to return URLs instead of base64 encoding	2025-10-28 18:21:08 +08:00
zhangzhanwei	79b47bb1b3	feat: Node execute	2025-10-28 18:09:52 +08:00
wxg0103	02311687c8	fix: update video format handling in base_video_understand_node.py	2025-10-28 17:55:54 +08:00
wxg0103	28d403f237	fix: update message handling in image.py to support multiple message types	2025-10-28 16:24:21 +08:00
wxg0103	79fce87e93	fix: add video format detection for uploaded files in image.py	2025-10-28 15:54:56 +08:00