fix: reference file with 'docx' type can not open #844 (#1635)

Chat Use CVmodel (#1607)

fix: after logging out and entering the knowledge base page again, the data before still exists #1306 (#1597)

feat: change all file names to lowercase #1574 (#1575)

fix: remove duplicate MessageItem #1289 (#1566)

feat: add llm Select to KeywordExtractForm #918 (#1492)

feat: save the parameters of the generate operator to the form field … (#1390)

feat: deleting a node does not require a confirmation box to pop up #918 (#1380)

feat: add description text to operators and extract the useFetchModelId to logicHooks.ts and drag the operator to the canvas and initialize the form data #918 (#1379)

feat: add DynamicParameters #918 (#1346)

feat: modify the style of the operator #918 (#1335)

feat: build categorize list from object #918 (#1276)

feat: add DynamicCategorize #918 (#1273)

feat: add LLMSelect (#1270)

Let json files support naive parsing methods #1245 (#1247)

feat: duplicate node #918 (#1136)

feat: add delete menu to graph node #918 (#1133)

feat: add OperateDropdown and send debug message #918 (#1095)

feat: add FlowChatBox #918 (#1086)

feat: fetch flow (#1068)

feat: add custom edge (#1061)

feat: add rerank models to the project #724 #162 (#966)

feat: RAPTOR is not displayed when the parsing method is picture. (#897)

Implements RAPTOR for better chunking #882 (#883)

fix: thumbnails are too large in the chat box #818 (#819)

feat: Support for conversational streaming (#809)

fix: remove Top K in retrieval testing #770 and if the document parsing fails, the error message returned by the backend is displayed (#782)

feat: delete the added model #503 and display an error message when the requested file fails to parse #684 (#708)

fix: fix uploaded file time error #680 (#690)

fix: filter knowledge list by keywords and clear the selected file list after the file is uploaded successfully and add ellipsis pattern to chunk list (#628)

fix: omit long file names (#608)

fix: #567 use modal to upload files in the knowledge base (#601)

feat: #345 even if the backend data returns empty, the skeleton of the chart will be displayed. (#461)

feat: translate EmbedModal #345 (#455)

feat: add overview (#391)

feat: add support for ollama #221 (#260)

change language #245 (#246)

For any type of file, if the parsing method is general, the chunk token number needs to be displayed. #217 (#218)

fix: #209 after saving the knowledge base configuration, jump to the dataset page (#212)

remove showDeleteConfirm function and center the Empty of knowledge list and extract the text of the login page to en.json (#203)

fix: fixed the issue where ChunkMethodModal cannot correctly display … (#189)

feat: add corresponding icons to files (#164)

feat: set the object-fit attribute of the chunk image to contain so that the image is not stretched and if the chunk number of the knowledge base is greater than 0, set the embedding model to disabled (#160)

feat: select the corresponding parsing method according to the file type and after the document is successfully uploaded, use the ChunkMethodModal to select the parsing method. and remove ChunkMethodModal from knowledge-file (#158)

fix: by obtaining the width and height of the pdf and passing it to b… (#151)

feat: add pages to ChunkMethodModal (#143)

add help info (#142)

feat: add image for chunk method (#139)

feat: add loading to ChatContainer and set font family to inter and add tooltip to Form.Item and download documents on the document list page (#136)

fix: disable sending messages if both application and conversation are empty and add loading to all pages (#134)