Revert "Feat/parent child retrieval" (#12095)
This commit is contained in:
-LAN-
GitHub
parent
9231fdbf4c
commit
db2aa83a7c
216 changed files with 3116 additions and 9066 deletions
|
|
@ -42,11 +42,6 @@ const translation = {
|
|||
zoomOut: 'Zoom Out',
|
||||
zoomIn: 'Zoom In',
|
||||
openInNewTab: 'Open in new tab',
|
||||
saveAndRegenerate: 'Save & Regenerate Child Chunks',
|
||||
close: 'Close',
|
||||
view: 'View',
|
||||
viewMore: 'VIEW MORE',
|
||||
regenerate: 'Regenerate',
|
||||
},
|
||||
errorMsg: {
|
||||
fieldRequired: '{{field}} is required',
|
||||
|
|
@ -480,10 +475,9 @@ const translation = {
|
|||
documents: 'Documents',
|
||||
hitTesting: 'Retrieval Testing',
|
||||
settings: 'Settings',
|
||||
emptyTip: 'This Knowledge has not been integrated within any application. Please refer to the document for guidance.',
|
||||
emptyTip: 'The Knowledge has not been associated, please go to the application or plug-in to complete the association.',
|
||||
viewDoc: 'View documentation',
|
||||
relatedApp: 'linked apps',
|
||||
noRelatedApp: 'No linked apps',
|
||||
},
|
||||
voiceInput: {
|
||||
speaking: 'Speak now...',
|
||||
|
|
|
|||
|
|
@ -4,9 +4,9 @@ const translation = {
|
|||
creation: 'Create Knowledge',
|
||||
update: 'Add data',
|
||||
},
|
||||
one: 'Data Source',
|
||||
two: 'Document Processing',
|
||||
three: 'Execute & Finish',
|
||||
one: 'Choose data source',
|
||||
two: 'Text Preprocessing and Cleaning',
|
||||
three: 'Execute and finish',
|
||||
},
|
||||
error: {
|
||||
unavailable: 'This Knowledge is not available',
|
||||
|
|
@ -47,7 +47,6 @@ const translation = {
|
|||
notionSyncTitle: 'Notion is not connected',
|
||||
notionSyncTip: 'To sync with Notion, connection to Notion must be established first.',
|
||||
connect: 'Go to connect',
|
||||
cancel: 'Cancel',
|
||||
button: 'Next',
|
||||
emptyDatasetCreation: 'I want to create an empty Knowledge',
|
||||
modal: {
|
||||
|
|
@ -95,49 +94,36 @@ const translation = {
|
|||
},
|
||||
},
|
||||
stepTwo: {
|
||||
segmentation: 'Chunk Settings',
|
||||
segmentation: 'Chunk settings',
|
||||
auto: 'Automatic',
|
||||
autoDescription: 'Automatically set chunk and preprocessing rules. Unfamiliar users are recommended to select this.',
|
||||
custom: 'Custom',
|
||||
customDescription: 'Customize chunks rules, chunks length, and preprocessing rules, etc.',
|
||||
general: 'General',
|
||||
generalTip: 'General text chunking mode, the chunks retrieved and recalled are the same.',
|
||||
parentChild: 'Parent-child',
|
||||
parentChildTip: 'When using the parent-child mode, the child-chunk is used for retrieval and the parent-chunk is used for recall as context.',
|
||||
parentChunkForContext: 'Parent-chunk for Context',
|
||||
childChunkForRetrieval: 'Child-chunk for Retrieval',
|
||||
paragraph: 'Paragraph',
|
||||
paragraphTip: 'This mode splits the text in to paragraphs based on delimiters and the maximum chunk length, using the split text as the parent chunk for retrieval.',
|
||||
fullDoc: 'Full Doc',
|
||||
fullDocTip: 'The entire document is used as the parent chunk and retrieved directly. Please note that for performance reasons, text exceeding 10000 tokens will be automatically truncated.',
|
||||
separator: 'Delimiter',
|
||||
separatorTip: 'A delimiter is the character used to separate text. \\n\\n and \\n are commonly used delimiters for separating paragraphs and lines. Combined with commas (\\n\\n,\\n), paragraphs will be segmented by lines when exceeding the maximum chunk length. You can also use special delimiters defined by yourself (e.g. ***).',
|
||||
separatorPlaceholder: '\\n\\n for paragraphs; \\n for lines',
|
||||
separatorPlaceholder: '\\n\\n for separating paragraphs; \\n for separating lines',
|
||||
maxLength: 'Maximum chunk length',
|
||||
maxLengthCheck: 'Maximum chunk length should be less than {{limit}}',
|
||||
overlap: 'Chunk overlap',
|
||||
overlapTip: 'Setting the chunk overlap can maintain the semantic relevance between them, enhancing the retrieve effect. It is recommended to set 10%-25% of the maximum chunk size.',
|
||||
overlapCheck: 'chunk overlap should not bigger than maximum chunk length',
|
||||
rules: 'Text Pre-processing Rules',
|
||||
rules: 'Text preprocessing rules',
|
||||
removeExtraSpaces: 'Replace consecutive spaces, newlines and tabs',
|
||||
removeUrlEmails: 'Delete all URLs and email addresses',
|
||||
removeStopwords: 'Remove stopwords such as "a", "an", "the"',
|
||||
preview: 'Preview',
|
||||
previewChunk: 'Preview Chunk',
|
||||
preview: 'Confirm & Preview',
|
||||
reset: 'Reset',
|
||||
indexMode: 'Index Method',
|
||||
indexMode: 'Index mode',
|
||||
qualified: 'High Quality',
|
||||
highQualityTip: 'Once finishing embedding in High Quality mode, reverting to Economical mode is not available.',
|
||||
recommend: 'Recommend',
|
||||
qualifiedTip: 'Calling the embedding model to process documents for more precise retrieval helps LLM generate high-quality answers.',
|
||||
qualifiedTip: 'Call default system embedding interface for processing to provide higher accuracy when users query.',
|
||||
warning: 'Please set up the model provider API key first.',
|
||||
click: 'Go to settings',
|
||||
economical: 'Economical',
|
||||
economicalTip: 'Using 10 keywords per chunk for retrieval, no tokens are consumed at the expense of reduced retrieval accuracy.',
|
||||
economicalTip: 'Use offline vector engines, keyword indexes, etc. to reduce accuracy without spending tokens',
|
||||
QATitle: 'Segmenting in Question & Answer format',
|
||||
QATip: 'Enable this option will consume more tokens',
|
||||
QALanguage: 'Segment using',
|
||||
useQALanguage: 'Chunk using Q&A format in',
|
||||
estimateCost: 'Estimation',
|
||||
estimateSegment: 'Estimated chunks',
|
||||
segmentCount: 'chunks',
|
||||
|
|
@ -167,19 +153,10 @@ const translation = {
|
|||
indexSettingTip: 'To change the index method & embedding model, please go to the ',
|
||||
retrievalSettingTip: 'To change the retrieval setting, please go to the ',
|
||||
datasetSettingLink: 'Knowledge settings.',
|
||||
previewChunkTip: 'Click the \'Preview Chunk\' button on the left to load the preview',
|
||||
previewChunkCount: '{{count}} Estimated chunks',
|
||||
switch: 'Switch',
|
||||
qaSwitchHighQualityTipTitle: 'Q&A Format Requires High-quality Indexing Method',
|
||||
qaSwitchHighQualityTipContent: 'Currently, only high-quality index method supports Q&A format chunking. Would you like to switch to high-quality mode?',
|
||||
notAvailableForParentChild: 'Not available for Parent-child Index',
|
||||
notAvailableForQA: 'Not available for Q&A Index',
|
||||
parentChildDelimiterTip: 'A delimiter is the character used to separate text. \\n\\n is recommended for splitting the original document into large parent chunks. You can also use special delimiters defined by yourself.',
|
||||
parentChildChunkDelimiterTip: 'A delimiter is the character used to separate text. \\n is recommended for splitting parent chunks into small child chunks. You can also use special delimiters defined by yourself.',
|
||||
},
|
||||
stepThree: {
|
||||
creationTitle: '🎉 Knowledge created',
|
||||
creationContent: 'We automatically named the Knowledge, you can modify it at any time.',
|
||||
creationContent: 'We automatically named the Knowledge, you can modify it at any time',
|
||||
label: 'Knowledge name',
|
||||
additionTitle: '🎉 Document uploaded',
|
||||
additionP1: 'The document has been uploaded to the Knowledge',
|
||||
|
|
@ -194,11 +171,6 @@ const translation = {
|
|||
modelButtonConfirm: 'Confirm',
|
||||
modelButtonCancel: 'Cancel',
|
||||
},
|
||||
otherDataSource: {
|
||||
title: 'Connect to other data sources?',
|
||||
description: 'Currently, Dify\'s knowledge base only has limited data sources. Contributing a data source to the Dify knowledge base is a fantastic way to help enhance the platform\'s flexibility and power for all users. Our contribution guide makes it easy to get started. Please click on the link below to learn more.',
|
||||
learnMore: 'Learn more',
|
||||
},
|
||||
}
|
||||
|
||||
export default translation
|
||||
|
|
|
|||
|
|
@ -2,14 +2,12 @@ const translation = {
|
|||
list: {
|
||||
title: 'Documents',
|
||||
desc: 'All files of the Knowledge are shown here, and the entire Knowledge can be linked to Dify citations or indexed via the Chat plugin.',
|
||||
learnMore: 'Learn more',
|
||||
addFile: 'Add file',
|
||||
addPages: 'Add Pages',
|
||||
addUrl: 'Add URL',
|
||||
table: {
|
||||
header: {
|
||||
fileName: 'NAME',
|
||||
chunkingMode: 'CHUNKING MODE',
|
||||
fileName: 'FILE NAME',
|
||||
words: 'WORDS',
|
||||
hitCount: 'RETRIEVAL COUNT',
|
||||
uploadTime: 'UPLOAD TIME',
|
||||
|
|
@ -21,7 +19,7 @@ const translation = {
|
|||
},
|
||||
action: {
|
||||
uploadFile: 'Upload new file',
|
||||
settings: 'Chunking Settings',
|
||||
settings: 'Segment settings',
|
||||
addButton: 'Add chunk',
|
||||
add: 'Add a chunk',
|
||||
batchAdd: 'Batch add',
|
||||
|
|
@ -318,76 +316,36 @@ const translation = {
|
|||
completed: 'Embedding completed',
|
||||
error: 'Embedding error',
|
||||
docName: 'Preprocessing document',
|
||||
mode: 'Chunking Setting',
|
||||
segmentLength: 'Maximum Chunk Length',
|
||||
textCleaning: 'Text Preprocessing Rules',
|
||||
mode: 'Segmentation rule',
|
||||
segmentLength: 'Chunks length',
|
||||
textCleaning: 'Text pre-definition and cleaning',
|
||||
segments: 'Paragraphs',
|
||||
highQuality: 'High-quality mode',
|
||||
economy: 'Economy mode',
|
||||
estimate: 'Estimated consumption',
|
||||
stop: 'Stop processing',
|
||||
pause: 'Pause',
|
||||
resume: 'Resume',
|
||||
resume: 'Resume processing',
|
||||
automatic: 'Automatic',
|
||||
custom: 'Custom',
|
||||
hierarchical: 'Parent-child',
|
||||
previewTip: 'Paragraph preview will be available after embedding is complete',
|
||||
parentMaxTokens: 'Parent',
|
||||
childMaxTokens: 'Child',
|
||||
},
|
||||
segment: {
|
||||
paragraphs: 'Paragraphs',
|
||||
chunks_one: 'CHUNK',
|
||||
chunks_other: 'CHUNKS',
|
||||
parentChunks_one: 'PARENT CHUNK',
|
||||
parentChunks_other: 'PARENT CHUNKS',
|
||||
childChunks_one: 'CHILD CHUNK',
|
||||
childChunks_other: 'CHILD CHUNKS',
|
||||
searchResults_zero: 'RESULT',
|
||||
searchResults_one: 'RESULT',
|
||||
searchResults_other: 'RESULTS',
|
||||
empty: 'No Chunk found',
|
||||
clearFilter: 'Clear filter',
|
||||
chunk: 'Chunk',
|
||||
parentChunk: 'Parent-Chunk',
|
||||
newChunk: 'New Chunk',
|
||||
childChunk: 'Child-Chunk',
|
||||
newChildChunk: 'New Child Chunk',
|
||||
keywords: 'KEYWORDS',
|
||||
addKeyWord: 'Add keyword',
|
||||
keywords: 'Key Words',
|
||||
addKeyWord: 'Add key word',
|
||||
keywordError: 'The maximum length of keyword is 20',
|
||||
characters_one: 'character',
|
||||
characters_other: 'characters',
|
||||
characters: 'characters',
|
||||
hitCount: 'Retrieval count',
|
||||
vectorHash: 'Vector hash: ',
|
||||
questionPlaceholder: 'Add question here',
|
||||
questionPlaceholder: 'add question here',
|
||||
questionEmpty: 'Question can not be empty',
|
||||
answerPlaceholder: 'Add answer here',
|
||||
answerPlaceholder: 'add answer here',
|
||||
answerEmpty: 'Answer can not be empty',
|
||||
contentPlaceholder: 'Add content here',
|
||||
contentPlaceholder: 'add content here',
|
||||
contentEmpty: 'Content can not be empty',
|
||||
newTextSegment: 'New Text Segment',
|
||||
newQaSegment: 'New Q&A Segment',
|
||||
addChunk: 'Add Chunk',
|
||||
addChildChunk: 'Add Child Chunk',
|
||||
addAnother: 'Add another',
|
||||
delete: 'Delete this chunk ?',
|
||||
chunkAdded: '1 chunk added',
|
||||
childChunkAdded: '1 child chunk added',
|
||||
editChunk: 'Edit Chunk',
|
||||
editParentChunk: 'Edit Parent Chunk',
|
||||
editChildChunk: 'Edit Child Chunk',
|
||||
chunkDetail: 'Chunk Detail',
|
||||
regenerationConfirmTitle: 'Do you want to regenerate child chunks?',
|
||||
regenerationConfirmMessage: 'Regenerating child chunks will overwrite the current child chunks, including edited chunks and newly added chunks. The regeneration cannot be undone.',
|
||||
regeneratingTitle: 'Regenerating child chunks',
|
||||
regeneratingMessage: 'This may take a moment, please wait...',
|
||||
regenerationSuccessTitle: 'Regeneration completed',
|
||||
regenerationSuccessMessage: 'You can close this window.',
|
||||
edited: 'EDITED',
|
||||
editedAt: 'Edited at',
|
||||
expandChunks: 'Expand chunks',
|
||||
collapseChunks: 'Collapse chunks',
|
||||
},
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -1,9 +1,9 @@
|
|||
const translation = {
|
||||
title: 'Retrieval Test',
|
||||
title: 'Retrieval Testing',
|
||||
settingTitle: 'Retrieval Setting',
|
||||
desc: 'Test the hitting effect of the Knowledge based on the given query text.',
|
||||
desc: 'Test the hitting effect of the Knowledge based on the given query text',
|
||||
dateTimeFormat: 'MM/DD/YYYY hh:mm A',
|
||||
records: 'Records',
|
||||
recents: 'Recents',
|
||||
table: {
|
||||
header: {
|
||||
source: 'Source',
|
||||
|
|
@ -16,19 +16,15 @@ const translation = {
|
|||
placeholder: 'Please enter a text, a short declarative sentence is recommended.',
|
||||
countWarning: 'Up to 200 characters.',
|
||||
indexWarning: 'High quality Knowledge only.',
|
||||
testing: 'Test',
|
||||
testing: 'Testing',
|
||||
},
|
||||
hit: {
|
||||
title: '{{num}} Retrieved Chunks',
|
||||
title: 'RETRIEVAL PARAGRAPHS',
|
||||
emptyTip: 'Retrieval Testing results will show here',
|
||||
},
|
||||
noRecentTip: 'No recent query results here',
|
||||
viewChart: 'View VECTOR CHART',
|
||||
viewDetail: 'View Detail',
|
||||
chunkDetail: 'Chunk Detail',
|
||||
hitChunks: 'Hit {{num}} child chunks',
|
||||
open: 'Open',
|
||||
keyword: 'Keywords',
|
||||
}
|
||||
|
||||
export default translation
|
||||
|
|
|
|||
|
|
@ -7,8 +7,7 @@ const translation = {
|
|||
nameError: 'Name cannot be empty',
|
||||
desc: 'Knowledge Description',
|
||||
descInfo: 'Please write a clear textual description to outline the content of the Knowledge. This description will be used as a basis for matching when selecting from multiple Knowledge for inference.',
|
||||
descPlaceholder: 'Describe what is in this data set. A detailed description allows AI to access the content of the data set in a timely manner. If empty, LangGenius will use the default hit strategy.',
|
||||
helpText: 'Learn how to write a good dataset description.',
|
||||
descPlaceholder: 'Describe what\'s in this Knowledge (optional)',
|
||||
descWrite: 'Learn how to write a good Knowledge description.',
|
||||
permissions: 'Permissions',
|
||||
permissionsOnlyMe: 'Only me',
|
||||
|
|
@ -17,10 +16,9 @@ const translation = {
|
|||
me: '(You)',
|
||||
indexMethod: 'Index Method',
|
||||
indexMethodHighQuality: 'High Quality',
|
||||
indexMethodHighQualityTip: 'Calling the embedding model to process documents for more precise retrieval helps LLM generate high-quality answers.',
|
||||
upgradeHighQualityTip: 'Once upgrading to High Quality mode, reverting to Economical mode is not available',
|
||||
indexMethodHighQualityTip: 'Call Embedding model for processing to provide higher accuracy when users query.',
|
||||
indexMethodEconomy: 'Economical',
|
||||
indexMethodEconomyTip: 'Using 10 keywords per chunk for retrieval, no tokens are consumed at the expense of reduced retrieval accuracy.',
|
||||
indexMethodEconomyTip: 'Use offline vector engines, keyword indexes, etc. to reduce accuracy without spending tokens',
|
||||
embeddingModel: 'Embedding Model',
|
||||
embeddingModelTip: 'Change the embedded model, please go to ',
|
||||
embeddingModelTipLink: 'Settings',
|
||||
|
|
@ -34,7 +32,6 @@ const translation = {
|
|||
externalKnowledgeID: 'External Knowledge ID',
|
||||
retrievalSettings: 'Retrieval Settings',
|
||||
save: 'Save',
|
||||
indexMethodChangeToEconomyDisabledTip: 'Not available for downgrading from HQ to ECO',
|
||||
},
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -1,13 +1,5 @@
|
|||
const translation = {
|
||||
knowledge: 'Knowledge',
|
||||
chunkingMode: {
|
||||
general: 'General',
|
||||
parentChild: 'Parent-child',
|
||||
},
|
||||
parentMode: {
|
||||
paragraph: 'Paragraph',
|
||||
fullDoc: 'Full-doc',
|
||||
},
|
||||
externalTag: 'External',
|
||||
externalAPI: 'External API',
|
||||
externalAPIPanelTitle: 'External Knowledge API',
|
||||
|
|
@ -20,7 +12,6 @@ const translation = {
|
|||
learnHowToWriteGoodKnowledgeDescription: 'Learn how to write a good knowledge description',
|
||||
externalAPIPanelDescription: 'The external knowledge API is used to connect to a knowledge base outside of Dify and retrieve knowledge from that knowledge base.',
|
||||
externalAPIPanelDocumentation: 'Learn how to create an External Knowledge API',
|
||||
localDocs: 'Local Docs',
|
||||
documentCount: ' docs',
|
||||
wordCount: ' k words',
|
||||
appCount: ' linked apps',
|
||||
|
|
@ -124,10 +115,8 @@ const translation = {
|
|||
change: 'Change',
|
||||
changeRetrievalMethod: 'Change retrieval method',
|
||||
},
|
||||
docsFailedNotice: 'documents indexed failed',
|
||||
docsFailedNotice: 'documents failed to be indexed',
|
||||
retry: 'Retry',
|
||||
documentsDisabled: '{{num}} documents disabled - inactive for over 30 days',
|
||||
enable: 'Enable',
|
||||
indexingTechnique: {
|
||||
high_quality: 'HQ',
|
||||
economy: 'ECO',
|
||||
|
|
@ -157,15 +146,6 @@ const translation = {
|
|||
nTo1RetrievalLegacy: 'N-to-1 retrieval will be officially deprecated from September. It is recommended to use the latest Multi-path retrieval to obtain better results. ',
|
||||
nTo1RetrievalLegacyLink: 'Learn more',
|
||||
nTo1RetrievalLegacyLinkText: ' N-to-1 retrieval will be officially deprecated in September.',
|
||||
batchAction: {
|
||||
selected: 'Selected',
|
||||
enable: 'Enable',
|
||||
disable: 'Disable',
|
||||
archive: 'Archive',
|
||||
delete: 'Delete',
|
||||
cancel: 'Cancel',
|
||||
},
|
||||
preprocessDocument: '{{num}} Preprocess Documents',
|
||||
}
|
||||
|
||||
export default translation
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue