Skip to content

赵月辉 / fastgpt-migrated

Go to a project
Commit 95325346 by Archer Committed by GitHub
Options

perf: vector format (#5516) 

* perf: vector format

* feat: embedding batch size
parent a92917c0
Showing with 433 additions and 35 deletions
document/content/docs/upgrading/4-12/4122.mdx
...@@ -6,9 +6,12 @@ description: 'FastGPT V4.12.2 更新说明' ...@@ -6,9 +6,12 @@ description: 'FastGPT V4.12.2 更新说明'
## 🚀 新增内容 ## 🚀 新增内容
1. 向量模型并发请求设置，不统一设置成 10，避免部分向量模型不支持并发，默认均为 1，可在模型配置中设置。
## ⚙️ 优化 ## ⚙️ 优化
1. 增加工作流**独立分支**异常检测。
2. 向量模型超过 1536 维度进行截断时，强制进行归一化。其他维度是否归一化，完全由配置决定，减少自动判断的计算量。
## 🐛 修复 ## 🐛 修复
...@@ -17,6 +20,7 @@ description: 'FastGPT V4.12.2 更新说明' ...@@ -17,6 +20,7 @@ description: 'FastGPT V4.12.2 更新说明'
3. 移动端，分享链接，异常加载了登录态对话页的导航。 3. 移动端，分享链接，异常加载了登录态对话页的导航。
4. 用户同步可能出现写冲突问题。 4. 用户同步可能出现写冲突问题。
5. 无法完全关闭系统套餐，会存在空对象默认值，导致鉴权异常。 5. 无法完全关闭系统套餐，会存在空对象默认值，导致鉴权异常。
6. 工作流，添加团队应用，搜索无效。
## 🔨 工具更新 ## 🔨 工具更新
......
document/data/doc-last-modified.json
...@@ -97,13 +97,14 @@ ...@@ -97,13 +97,14 @@
"document/content/docs/protocol/terms.en.mdx": "2025-08-03T22:37:45+08:00", "document/content/docs/protocol/terms.en.mdx": "2025-08-03T22:37:45+08:00",
"document/content/docs/protocol/terms.mdx": "2025-08-03T22:37:45+08:00", "document/content/docs/protocol/terms.mdx": "2025-08-03T22:37:45+08:00",
"document/content/docs/toc.en.mdx": "2025-08-04T13:42:36+08:00", "document/content/docs/toc.en.mdx": "2025-08-04T13:42:36+08:00",
"document/content/docs/toc.mdx": "2025-08-13T14:29:13+08:00", "document/content/docs/toc.mdx": "2025-08-20T21:58:13+08:00",
"document/content/docs/upgrading/4-10/4100.mdx": "2025-08-02T19:38:37+08:00", "document/content/docs/upgrading/4-10/4100.mdx": "2025-08-02T19:38:37+08:00",
"document/content/docs/upgrading/4-10/4101.mdx": "2025-08-02T19:38:37+08:00", "document/content/docs/upgrading/4-10/4101.mdx": "2025-08-02T19:38:37+08:00",
"document/content/docs/upgrading/4-11/4110.mdx": "2025-08-05T23:20:39+08:00", "document/content/docs/upgrading/4-11/4110.mdx": "2025-08-05T23:20:39+08:00",
"document/content/docs/upgrading/4-11/4111.mdx": "2025-08-07T22:49:09+08:00", "document/content/docs/upgrading/4-11/4111.mdx": "2025-08-07T22:49:09+08:00",
"document/content/docs/upgrading/4-12/4120.mdx": "2025-08-12T22:45:19+08:00", "document/content/docs/upgrading/4-12/4120.mdx": "2025-08-12T22:45:19+08:00",
"document/content/docs/upgrading/4-12/4121.mdx": "2025-08-15T22:53:06+08:00", "document/content/docs/upgrading/4-12/4121.mdx": "2025-08-15T22:53:06+08:00",
"document/content/docs/upgrading/4-12/4122.mdx": "2025-08-22T09:38:44+08:00",
"document/content/docs/upgrading/4-8/40.mdx": "2025-08-02T19:38:37+08:00", "document/content/docs/upgrading/4-8/40.mdx": "2025-08-02T19:38:37+08:00",
"document/content/docs/upgrading/4-8/41.mdx": "2025-08-02T19:38:37+08:00", "document/content/docs/upgrading/4-8/41.mdx": "2025-08-02T19:38:37+08:00",
"document/content/docs/upgrading/4-8/42.mdx": "2025-08-02T19:38:37+08:00", "document/content/docs/upgrading/4-8/42.mdx": "2025-08-02T19:38:37+08:00",
......
packages/global/core/ai/model.d.ts
...@@ -65,6 +65,7 @@ export type EmbeddingModelItemType = PriceType & ...@@ -65,6 +65,7 @@ export type EmbeddingModelItemType = PriceType &
weight: number; // training weight weight: number; // training weight
hidden?: boolean; // Disallow creation hidden?: boolean; // Disallow creation
normalization?: boolean; // normalization processing normalization?: boolean; // normalization processing
batchSize?: number;
defaultConfig?: Record<string, any>; // post request config defaultConfig?: Record<string, any>; // post request config
dbConfig?: Record<string, any>; // Custom parameters for storage dbConfig?: Record<string, any>; // Custom parameters for storage
queryConfig?: Record<string, any>; // Custom parameters for query queryConfig?: Record<string, any>; // Custom parameters for query
......
packages/global/core/workflow/node/constant.ts
...@@ -110,6 +110,8 @@ export enum FlowNodeTypeEnum { ...@@ -110,6 +110,8 @@ export enum FlowNodeTypeEnum {
systemConfig = 'userGuide', systemConfig = 'userGuide',
pluginConfig = 'pluginConfig', pluginConfig = 'pluginConfig',
globalVariable = 'globalVariable', globalVariable = 'globalVariable',
comment = 'comment',
workflowStart = 'workflowStart', workflowStart = 'workflowStart',
chatNode = 'chatNode', chatNode = 'chatNode',
...@@ -141,7 +143,6 @@ export enum FlowNodeTypeEnum { ...@@ -141,7 +143,6 @@ export enum FlowNodeTypeEnum {
loopStart = 'loopStart', loopStart = 'loopStart',
loopEnd = 'loopEnd', loopEnd = 'loopEnd',
formInput = 'formInput', formInput = 'formInput',
comment = 'comment',
tool = 'tool', tool = 'tool',
toolSet = 'toolSet' toolSet = 'toolSet'
} }
......
packages/service/core/ai/embedding/index.ts
...@@ -23,8 +23,9 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto ...@@ -23,8 +23,9 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto
const formatInput = Array.isArray(input) ? input : [input]; const formatInput = Array.isArray(input) ? input : [input];
// 20 size every request let chunkSize = Number(model.batchSize || 1);
const chunkSize = parseInt(process.env.EMBEDDING_CHUNK_SIZE || '10'); chunkSize = isNaN(chunkSize) ? 1 : chunkSize;
const chunks = []; const chunks = [];
for (let i = 0; i < formatInput.length; i += chunkSize) { for (let i = 0; i < formatInput.length; i += chunkSize) {
chunks.push(formatInput.slice(i, i + chunkSize)); chunks.push(formatInput.slice(i, i + chunkSize));
...@@ -74,14 +75,7 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto ...@@ -74,14 +75,7 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto
const tokens = await Promise.all(chunk.map((item) => countPromptTokens(item))); const tokens = await Promise.all(chunk.map((item) => countPromptTokens(item)));
return tokens.reduce((sum, item) => sum + item, 0); return tokens.reduce((sum, item) => sum + item, 0);
})(), })(),
Promise.all( Promise.all(res.data.map((item) => formatVectors(item.embedding, model.normalization)))
res.data
.map((item) => unityDimensional(item.embedding))
.map((item) => {
if (model.normalization) return normalization(item);
return item;
})
)
]); ]);
return { return {
...@@ -105,28 +99,34 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto ...@@ -105,28 +99,34 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto
} }
} }
function unityDimensional(vector: number[]) { export function formatVectors(vector: number[], normalization = false) {
// normalization processing
function normalizationVector(vector: number[]) {
// Calculate the Euclidean norm (L2 norm)
const norm = Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
if (norm === 0) {
return vector;
}
// Normalize the vector by dividing each component by the norm
return vector.map((val) => val / norm);
}
// 超过上限，截断，并强制归一化
if (vector.length > 1536) { if (vector.length > 1536) {
console.log( console.log(
`The current vector dimension is ${vector.length}, and the vector dimension cannot exceed 1536. The first 1536 dimensions are automatically captured` `The current vector dimension is ${vector.length}, and the vector dimension cannot exceed 1536. The first 1536 dimensions are automatically captured`
); );
return vector.slice(0, 1536); return normalizationVector(vector.slice(0, 1536));
} } else if (vector.length < 1536) {
let resultVector = vector; const vectorLen = vector.length;
const vectorLen = vector.length;
const zeroVector = new Array(1536 - vectorLen).fill(0); const zeroVector = new Array(1536 - vectorLen).fill(0);
return resultVector.concat(zeroVector); vector = vector.concat(zeroVector);
} }
// normalization processing
function normalization(vector: number[]) {
if (vector.some((item) => item > 1)) {
// Calculate the Euclidean norm (L2 norm)
const norm = Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
// Normalize the vector by dividing each component by the norm if (normalization) {
return vector.map((val) => val / norm); return normalizationVector(vector);
} }
return vector; return vector;
......
packages/web/i18n/en/account_model.json
...@@ -7,6 +7,7 @@ ...@@ -7,6 +7,7 @@
"avg_ttfb": "Average first word duration (seconds)", "avg_ttfb": "Average first word duration (seconds)",
"azure": "Azure", "azure": "Azure",
"base_url": "Base url", "base_url": "Base url",
"batch_size": "Number of concurrent requests",
"channel_name": "Channel", "channel_name": "Channel",
"channel_priority": "Priority", "channel_priority": "Priority",
"channel_priority_tip": "The higher the priority channel, the easier it is to be requested", "channel_priority_tip": "The higher the priority channel, the easier it is to be requested",
......
packages/web/i18n/zh-CN/account_model.json
...@@ -7,6 +7,7 @@ ...@@ -7,6 +7,7 @@
"avg_ttfb": "平均首字时长 (秒)", "avg_ttfb": "平均首字时长 (秒)",
"azure": "微软 Azure", "azure": "微软 Azure",
"base_url": "代理地址", "base_url": "代理地址",
"batch_size": "并发请求数",
"channel_name": "渠道名", "channel_name": "渠道名",
"channel_priority": "优先级", "channel_priority": "优先级",
"channel_priority_tip": "优先级越高的渠道，越容易被请求到", "channel_priority_tip": "优先级越高的渠道，越容易被请求到",
......
packages/web/i18n/zh-Hant/account_model.json
...@@ -7,6 +7,7 @@ ...@@ -7,6 +7,7 @@
"avg_ttfb": "平均首字時長 (秒)", "avg_ttfb": "平均首字時長 (秒)",
"azure": "Azure", "azure": "Azure",
"base_url": "代理地址", "base_url": "代理地址",
"batch_size": "並發請求數",
"channel_name": "管道名稱", "channel_name": "管道名稱",
"channel_priority": "優先順序", "channel_priority": "優先順序",
"channel_priority_tip": "優先順序越高的管道，越容易被請求到", "channel_priority_tip": "優先順序越高的管道，越容易被請求到",
......
projects/app/src/components/Select/AIModelSelector.tsx
...@@ -10,12 +10,14 @@ import MyTooltip from '@fastgpt/web/components/common/MyTooltip'; ...@@ -10,12 +10,14 @@ import MyTooltip from '@fastgpt/web/components/common/MyTooltip';
import { ModelProviderList } from '@fastgpt/global/core/ai/provider'; import { ModelProviderList } from '@fastgpt/global/core/ai/provider';
import MultipleRowSelect from '@fastgpt/web/components/common/MySelect/MultipleRowSelect'; import MultipleRowSelect from '@fastgpt/web/components/common/MySelect/MultipleRowSelect';
import { getModelFromList } from '@fastgpt/global/core/ai/model'; import { getModelFromList } from '@fastgpt/global/core/ai/model';
import type { ResponsiveValue } from '@chakra-ui/system';
type Props = SelectProps & { type Props = SelectProps & {
disableTip?: string; disableTip?: string;
noOfLines?: ResponsiveValue<number>;
}; };
const OneRowSelector = ({ list, onChange, disableTip, ...props }: Props) => { const OneRowSelector = ({ list, onChange, disableTip, noOfLines, ...props }: Props) => {
const { t } = useTranslation(); const { t } = useTranslation();
const { llmModelList, embeddingModelList, ttsModelList, sttModelList, reRankModelList } = const { llmModelList, embeddingModelList, ttsModelList, sttModelList, reRankModelList } =
useSystemStore(); useSystemStore();
...@@ -55,7 +57,7 @@ const OneRowSelector = ({ list, onChange, disableTip, ...props }: Props) => { ...@@ -55,7 +57,7 @@ const OneRowSelector = ({ list, onChange, disableTip, ...props }: Props) => {
fallbackSrc={HUGGING_FACE_ICON} fallbackSrc={HUGGING_FACE_ICON}
/> />
<Box noOfLines={1}>{modelData.name}</Box> <Box noOfLines={noOfLines}>{modelData.name}</Box>
</Flex> </Flex>
) )
}; };
...@@ -99,7 +101,14 @@ const OneRowSelector = ({ list, onChange, disableTip, ...props }: Props) => { ...@@ -99,7 +101,14 @@ const OneRowSelector = ({ list, onChange, disableTip, ...props }: Props) => {
); );
}; };
const MultipleRowSelector = ({ list, onChange, disableTip, placeholder, ...props }: Props) => { const MultipleRowSelector = ({
list,
onChange,
disableTip,
placeholder,
noOfLines,
...props
}: Props) => {
const { t } = useTranslation(); const { t } = useTranslation();
const { llmModelList, embeddingModelList, ttsModelList, sttModelList, reRankModelList } = const { llmModelList, embeddingModelList, ttsModelList, sttModelList, reRankModelList } =
useSystemStore(); useSystemStore();
...@@ -189,7 +198,7 @@ const MultipleRowSelector = ({ list, onChange, disableTip, placeholder, ...props ...@@ -189,7 +198,7 @@ const MultipleRowSelector = ({ list, onChange, disableTip, placeholder, ...props
fallbackSrc={HUGGING_FACE_ICON} fallbackSrc={HUGGING_FACE_ICON}
w={avatarSize} w={avatarSize}
/> />
<Box noOfLines={1}>{modelData?.name}</Box> <Box noOfLines={noOfLines}>{modelData?.name}</Box>
</Flex> </Flex>
); );
}, [modelList, props.value, t, avatarSize]); }, [modelList, props.value, t, avatarSize]);
...@@ -222,7 +231,7 @@ const MultipleRowSelector = ({ list, onChange, disableTip, placeholder, ...props ...@@ -222,7 +231,7 @@ const MultipleRowSelector = ({ list, onChange, disableTip, placeholder, ...props
}; };
const AIModelSelector = (props: Props) => { const AIModelSelector = (props: Props) => {
return props.list.length > 100 ? ( return props.list.length > 10 ? (
<MultipleRowSelector {...props} /> <MultipleRowSelector {...props} />
) : ( ) : (
<OneRowSelector {...props} /> <OneRowSelector {...props} />
......
projects/app/src/pageComponents/account/model/AddModelBox.tsx
...@@ -479,6 +479,26 @@ export const ModelEditModal = ({ ...@@ -479,6 +479,26 @@ export const ModelEditModal = ({
<Tr> <Tr>
<Td> <Td>
<HStack spacing={1}> <HStack spacing={1}>
<Box>{t('account_model:batch_size')}</Box>
</HStack>
</Td>
<Td textAlign={'right'}>
<Flex justifyContent={'flex-end'}>
<MyNumberInput
defaultValue={1}
register={register}
name="batchSize"
min={1}
step={1}
isRequired
{...InputStyles}
/>
</Flex>
</Td>
</Tr>
<Tr>
<Td>
<HStack spacing={1}>
<Box>{t('account:model.default_token')}</Box> <Box>{t('account:model.default_token')}</Box>
<QuestionTip label={t('account:model.default_token_tip')} /> <QuestionTip label={t('account:model.default_token_tip')} />
</HStack> </HStack>
......
projects/app/src/pageComponents/app/detail/WorkflowComponents/Flow/hooks/useDebug.tsx
...@@ -48,6 +48,7 @@ export const useDebug = () => { ...@@ -48,6 +48,7 @@ export const useDebug = () => {
const getNodes = useContextSelector(WorkflowNodeEdgeContext, (v) => v.getNodes); const getNodes = useContextSelector(WorkflowNodeEdgeContext, (v) => v.getNodes);
const edges = useContextSelector(WorkflowNodeEdgeContext, (v) => v.edges); const edges = useContextSelector(WorkflowNodeEdgeContext, (v) => v.edges);
const onUpdateNodeError = useContextSelector(WorkflowContext, (v) => v.onUpdateNodeError); const onUpdateNodeError = useContextSelector(WorkflowContext, (v) => v.onUpdateNodeError);
const onRemoveError = useContextSelector(WorkflowContext, (v) => v.onRemoveError);
const onStartNodeDebug = useContextSelector(WorkflowContext, (v) => v.onStartNodeDebug); const onStartNodeDebug = useContextSelector(WorkflowContext, (v) => v.onStartNodeDebug);
const appDetail = useContextSelector(AppContext, (v) => v.appDetail); const appDetail = useContextSelector(AppContext, (v) => v.appDetail);
...@@ -80,6 +81,7 @@ export const useDebug = () => { ...@@ -80,6 +81,7 @@ export const useDebug = () => {
const checkResults = checkWorkflowNodeAndConnection({ nodes, edges }); const checkResults = checkWorkflowNodeAndConnection({ nodes, edges });
if (!checkResults) { if (!checkResults) {
onRemoveError();
const storeNodes = uiWorkflow2StoreWorkflow({ nodes, edges }); const storeNodes = uiWorkflow2StoreWorkflow({ nodes, edges });
return JSON.stringify(storeNodes); return JSON.stringify(storeNodes);
......
projects/app/src/pageComponents/app/detail/WorkflowComponents/context/index.tsx
...@@ -157,6 +157,7 @@ type WorkflowContextType = { ...@@ -157,6 +157,7 @@ type WorkflowContextType = {
nodeList: FlowNodeItemType[]; nodeList: FlowNodeItemType[];
onUpdateNodeError: (node: string, isError: Boolean) => void; onUpdateNodeError: (node: string, isError: Boolean) => void;
onRemoveError: () => void;
onResetNode: (e: { id: string; node: FlowNodeTemplateType }) => void; onResetNode: (e: { id: string; node: FlowNodeTemplateType }) => void;
onChangeNode: (e: FlowNodeChangeProps) => void; onChangeNode: (e: FlowNodeChangeProps) => void;
getNodeDynamicInputs: (nodeId: string) => FlowNodeInputItemType[]; getNodeDynamicInputs: (nodeId: string) => FlowNodeInputItemType[];
...@@ -401,6 +402,9 @@ export const WorkflowContext = createContext<WorkflowContextType>({ ...@@ -401,6 +402,9 @@ export const WorkflowContext = createContext<WorkflowContextType>({
isSaved?: boolean; isSaved?: boolean;
}): boolean { }): boolean {
throw new Error('Function not implemented.'); throw new Error('Function not implemented.');
},
onRemoveError: function (): void {
throw new Error('Function not implemented.');
} }
}); });
...@@ -473,6 +477,17 @@ const WorkflowContextProvider = ({ ...@@ -473,6 +477,17 @@ const WorkflowContextProvider = ({
}); });
}); });
}); });
const onRemoveError = useMemoizedFn(() => {
setNodes((state) => {
return state.map((item) => {
if (item.data.isError) {
item.data.isError = false;
item.selected = false;
}
return item;
});
});
});
// reset a node data. delete edge and replace it // reset a node data. delete edge and replace it
const onResetNode = useMemoizedFn(({ id, node }: { id: string; node: FlowNodeTemplateType }) => { const onResetNode = useMemoizedFn(({ id, node }: { id: string; node: FlowNodeTemplateType }) => {
...@@ -625,6 +640,7 @@ const WorkflowContextProvider = ({ ...@@ -625,6 +640,7 @@ const WorkflowContextProvider = ({
const checkResults = checkWorkflowNodeAndConnection({ nodes, edges }); const checkResults = checkWorkflowNodeAndConnection({ nodes, edges });
if (!checkResults) { if (!checkResults) {
onRemoveError();
const storeWorkflow = uiWorkflow2StoreWorkflow({ nodes, edges }); const storeWorkflow = uiWorkflow2StoreWorkflow({ nodes, edges });
return storeWorkflow; return storeWorkflow;
...@@ -1025,6 +1041,7 @@ const WorkflowContextProvider = ({ ...@@ -1025,6 +1041,7 @@ const WorkflowContextProvider = ({
// node // node
nodeList, nodeList,
onUpdateNodeError, onUpdateNodeError,
onRemoveError,
onResetNode, onResetNode,
onChangeNode, onChangeNode,
getNodeDynamicInputs, getNodeDynamicInputs,
...@@ -1075,6 +1092,7 @@ const WorkflowContextProvider = ({ ...@@ -1075,6 +1092,7 @@ const WorkflowContextProvider = ({
onChangeNode, onChangeNode,
onDelEdge, onDelEdge,
onNextNodeDebug, onNextNodeDebug,
onRemoveError,
onResetNode, onResetNode,
onStartNodeDebug, onStartNodeDebug,
onStopNodeDebug, onStopNodeDebug,
......
projects/app/src/pageComponents/chat/ChatWindow/HomeChatWindow.tsx
...@@ -248,6 +248,7 @@ const HomeChatWindow = ({ myApps }: Props) => { ...@@ -248,6 +248,7 @@ const HomeChatWindow = ({ myApps }: Props) => {
size="sm" size="sm"
bg={'myGray.50'} bg={'myGray.50'}
rounded="full" rounded="full"
noOfLines={[1, 3]}
list={availableModels} list={availableModels}
value={selectedModel} value={selectedModel}
onChange={async (model) => { onChange={async (model) => {
......
projects/app/src/web/core/workflow/utils.ts
...@@ -381,6 +381,11 @@ export const getNodeAllSource = ({ ...@@ -381,6 +381,11 @@ export const getNodeAllSource = ({
}; };
/* ====== Connection ======= */ /* ====== Connection ======= */
// Connectivity check result type
type ConnectivityIssue = {
nodeId: string;
issue: 'isolated' | 'no_input' | 'unreachable_from_start';
};
export const checkWorkflowNodeAndConnection = ({ export const checkWorkflowNodeAndConnection = ({
nodes, nodes,
edges edges
...@@ -388,7 +393,7 @@ export const checkWorkflowNodeAndConnection = ({ ...@@ -388,7 +393,7 @@ export const checkWorkflowNodeAndConnection = ({
nodes: Node<FlowNodeItemType, string | undefined>[]; nodes: Node<FlowNodeItemType, string | undefined>[];
edges: Edge<any>[]; edges: Edge<any>[];
}): string[] | undefined => { }): string[] | undefined => {
// 1. reference check. Required value // Node check
for (const node of nodes) { for (const node of nodes) {
const data = node.data; const data = node.data;
const inputs = data.inputs; const inputs = data.inputs;
...@@ -453,6 +458,15 @@ export const checkWorkflowNodeAndConnection = ({ ...@@ -453,6 +458,15 @@ export const checkWorkflowNodeAndConnection = ({
return [data.nodeId]; return [data.nodeId];
} }
} }
if (data.flowNodeType === FlowNodeTypeEnum.agent) {
const toolConnections = edges.filter(
(edge) =>
edge.source === data.nodeId && edge.sourceHandle === NodeOutputKeyEnum.selectedTools
);
if (toolConnections.length === 0) {
return [data.nodeId];
}
}
// check node input // check node input
if ( if (
...@@ -506,7 +520,7 @@ export const checkWorkflowNodeAndConnection = ({ ...@@ -506,7 +520,7 @@ export const checkWorkflowNodeAndConnection = ({
return [data.nodeId]; return [data.nodeId];
} }
// filter tools node edge // Check node has invalid edge
const edgeFilted = edges.filter( const edgeFilted = edges.filter(
(edge) => (edge) =>
!( !(
...@@ -514,7 +528,7 @@ export const checkWorkflowNodeAndConnection = ({ ...@@ -514,7 +528,7 @@ export const checkWorkflowNodeAndConnection = ({
edge.sourceHandle === NodeOutputKeyEnum.selectedTools edge.sourceHandle === NodeOutputKeyEnum.selectedTools
) )
); );
// check node has edge // Check node has edge
const hasEdge = edgeFilted.some( const hasEdge = edgeFilted.some(
(edge) => edge.source === data.nodeId || edge.target === data.nodeId (edge) => edge.source === data.nodeId || edge.target === data.nodeId
); );
...@@ -522,6 +536,106 @@ export const checkWorkflowNodeAndConnection = ({ ...@@ -522,6 +536,106 @@ export const checkWorkflowNodeAndConnection = ({
return [data.nodeId]; return [data.nodeId];
} }
} }
// Edge check
/**
* Check graph connectivity and identify connectivity issues
*/
const checkConnectivity = (
nodes: Node<FlowNodeItemType, string | undefined>[],
edges: Edge<any>[]
): string[] => {
// Find start node
const startNode = nodes.find(
(node) =>
node.data.flowNodeType === FlowNodeTypeEnum.workflowStart ||
node.data.flowNodeType === FlowNodeTypeEnum.pluginInput
);
if (!startNode) {
// No start node found - this is a critical issue
return nodes.map((node) => node.data.nodeId);
}
const issues: ConnectivityIssue[] = [];
// Build adjacency lists for both directions
const outgoing = new Map<string, string[]>();
const incoming = new Map<string, string[]>();
nodes.forEach((node) => {
outgoing.set(node.data.nodeId, []);
incoming.set(node.data.nodeId, []);
});
edges.forEach((edge) => {
const outList = outgoing.get(edge.source) || [];
outList.push(edge.target);
outgoing.set(edge.source, outList);
const inList = incoming.get(edge.target) || [];
inList.push(edge.source);
incoming.set(edge.target, inList);
});
// Check reachability from start node（Start node/Loop start 可以到达的地方）
const reachableFromStart = new Set<string>();
const dfsFromStart = (nodeId: string) => {
if (reachableFromStart.has(nodeId)) return;
reachableFromStart.add(nodeId);
const neighbors = outgoing.get(nodeId) || [];
neighbors.forEach((neighbor) => dfsFromStart(neighbor));
};
dfsFromStart(startNode.data.nodeId);
nodes.forEach((node) => {
if (node.data.flowNodeType === FlowNodeTypeEnum.loopStart) {
dfsFromStart(node.data.nodeId);
}
});
// Check each node for connectivity issues
for (const node of nodes) {
const nodeId = node.data.nodeId;
const nodeType = node.data.flowNodeType;
// Skip system nodes that don't need connectivity checks
if (
nodeType === FlowNodeTypeEnum.systemConfig ||
nodeType === FlowNodeTypeEnum.pluginConfig ||
nodeType === FlowNodeTypeEnum.comment ||
nodeType === FlowNodeTypeEnum.globalVariable ||
nodeType === FlowNodeTypeEnum.emptyNode
) {
continue;
}
const hasIncoming = (incoming.get(nodeId) || []).length > 0;
const hasOutgoing = (outgoing.get(nodeId) || []).length > 0;
const isStartNode = [
FlowNodeTypeEnum.workflowStart,
FlowNodeTypeEnum.pluginInput,
FlowNodeTypeEnum.loopStart
].includes(nodeType);
// Check if node is reachable from start
if (!isStartNode && !reachableFromStart.has(nodeId)) {
issues.push({
nodeId,
issue: 'unreachable_from_start'
});
break;
}
}
return issues.map((issue) => issue.nodeId);
};
const connectivityIssues = checkConnectivity(nodes, edges);
if (connectivityIssues.length > 0) {
return connectivityIssues;
}
}; };
/* ====== Variables ======= */ /* ====== Variables ======= */
......
test/cases/service/core/ai/embedding.test.ts 0 → 100644
import { formatVectors } from '@fastgpt/service/core/ai/embedding/index';
import { describe, expect, it, vi } from 'vitest';
describe('formatVectors function test', () => {
// Helper function to create a normalized vector (L2 norm = 1)
const createNormalizedVector = (length: number): number[] => {
const vector = Array.from({ length }, (_, i) => (i + 1) / length);
const norm = Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
return vector.map((val) => val / norm);
};
// Helper function to create an unnormalized vector
const createUnnormalizedVector = (length: number): number[] => {
return Array.from({ length }, (_, i) => (i + 1) * 10);
};
// Helper function to calculate L2 norm
const calculateNorm = (vector: number[]): number => {
return Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
};
// Helper function to check if vector is normalized (L2 norm H 1)
const isNormalized = (vector: number[]): boolean => {
const norm = calculateNorm(vector);
return Math.abs(norm - 1) < 1e-10;
};
describe('1536 dimension vectors', () => {
it('should handle normalized 1536-dim vector with normalization=true', () => {
const inputVector = createNormalizedVector(1536);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Since input is already normalized, result should be very similar
expect(result).toEqual(
expect.arrayContaining(inputVector.map((val) => expect.closeTo(val, 10)))
);
});
it('should handle normalized 1536-dim vector with normalization=false', () => {
const inputVector = createNormalizedVector(1536);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
expect(result).toEqual(inputVector);
expect(isNormalized(result)).toBe(true);
});
it('should handle unnormalized 1536-dim vector with normalization=true', () => {
const inputVector = createUnnormalizedVector(1536);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Result should be different from input (normalized)
expect(result).not.toEqual(inputVector);
});
it('should handle unnormalized 1536-dim vector with normalization=false', () => {
const inputVector = createUnnormalizedVector(1536);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
expect(result).toEqual(inputVector);
expect(isNormalized(result)).toBe(false);
});
});
describe('Greater than 1536 dimension vectors', () => {
it('should handle normalized >1536-dim vector with normalization=true', () => {
const inputVector = createNormalizedVector(2048);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated to first 1536 elements and then normalized
expect(result).toEqual(
expect.arrayContaining(inputVector.slice(0, 1536).map((val) => expect.any(Number)))
);
});
it('should handle normalized >1536-dim vector with normalization=false', () => {
const inputVector = createNormalizedVector(2048);
const result = formatVectors(inputVector, true); // Always normalized for >1536 dims
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated and normalized regardless of normalization flag
});
it('should handle unnormalized >1536-dim vector with normalization=true', () => {
const inputVector = createUnnormalizedVector(2048);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated to first 1536 elements and then normalized
});
it('should handle unnormalized >1536-dim vector with normalization=false', () => {
const inputVector = createUnnormalizedVector(2048);
const result = formatVectors(inputVector, false); // Always normalized for >1536 dims
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated and normalized regardless of normalization flag
});
it('should log warning for vectors with length > 1536', () => {
const consoleSpy = vi.spyOn(console, 'log').mockImplementation(() => {});
const inputVector = createNormalizedVector(2000);
formatVectors(inputVector, false);
expect(consoleSpy).toHaveBeenCalledWith(
expect.stringContaining(
'The current vector dimension is 2000, and the vector dimension cannot exceed 1536'
)
);
consoleSpy.mockRestore();
});
});
describe('Less than 1536 dimension vectors', () => {
it('should handle normalized <1536-dim vector with normalization=true', () => {
const inputVector = createNormalizedVector(512);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// First 512 elements should match input, rest should be 0
expect(result.slice(0, 512)).toEqual(
expect.arrayContaining(inputVector.map((val) => expect.any(Number)))
);
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
});
it('should handle normalized <1536-dim vector with normalization=false', () => {
const inputVector = createNormalizedVector(512);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
// First 512 elements should match input exactly, rest should be 0
expect(result.slice(0, 512)).toEqual(inputVector);
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
// The result remains normalized because adding zeros doesn't change the L2 norm
expect(isNormalized(result)).toBe(true);
});
it('should handle unnormalized <1536-dim vector with normalization=true', () => {
const inputVector = createUnnormalizedVector(512);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be padded with zeros and then normalized
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
});
it('should handle unnormalized <1536-dim vector with normalization=false', () => {
const inputVector = createUnnormalizedVector(512);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
// First 512 elements should match input exactly, rest should be 0
expect(result.slice(0, 512)).toEqual(inputVector);
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
expect(isNormalized(result)).toBe(false);
});
it('should demonstrate that padding preserves normalization status', () => {
// Create a vector that becomes unnormalized after some scaling
const baseVector = [3, 4]; // norm = 5, not normalized
const result = formatVectors(baseVector, false);
expect(result).toHaveLength(1536);
expect(result[0]).toBe(3);
expect(result[1]).toBe(4);
expect(result.slice(2)).toEqual(new Array(1534).fill(0));
expect(isNormalized(result)).toBe(false);
expect(calculateNorm(result)).toBeCloseTo(5, 10);
});
});
describe('Edge cases', () => {
it('should handle zero vector', () => {
const inputVector = new Array(1536).fill(0);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(result).toEqual(inputVector); // Zero vector remains zero after normalization
});
it('should handle single element vector', () => {
const inputVector = [5.0];
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(result[0]).toBeCloseTo(1.0, 10); // Normalized single element should be 1
expect(result.slice(1)).toEqual(new Array(1535).fill(0));
});
it('should handle exactly 1536 dimension vector', () => {
const inputVector = createNormalizedVector(1536);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
});
it('should handle vector with negative values', () => {
const inputVector = [-1, -2, -3];
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
expect(result[0]).toBeLessThan(0); // Should preserve negative values
expect(result[1]).toBeLessThan(0);
expect(result[2]).toBeLessThan(0);
});
});
});
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or sign in to comment