import { fetchSourceAssets } from './assets' import type { SourceAsset } from './assets' export type DocumentStage = 'waiting' | 'downloading' | 'uploading' | 'parsing' | 'indexing' | 'completed' | 'reused' | 'blocked' | 'failed' export const stageLabels: Record = { waiting: '等待处理', downloading: '下载中', uploading: '上传中', parsing: '解析中', indexing: '建立索引', completed: '解析完成', reused: '内容重复,复用已有文档', blocked: '文件超限', failed: '需要处理', } export type PipelineState='waiting'|'active'|'done'|'problem' export interface PipelineStep {label:string;state:PipelineState;detail:string} export function documentPipeline(item:SourceAsset):PipelineStep[]{ let source:PipelineStep={label:'取得文件',state:'waiting',detail:'等待取得文件'} if(item.content_state==='verified')source={label:'取得文件',state:'done',detail:'文件已校验'} else if(item.content_state==='downloading')source={label:'取得文件',state:'active',detail:'下载中'} else if(item.content_state==='over_limit')source={label:'取得文件',state:'problem',detail:'文件超限'} else if(['invalid','failed'].includes(item.content_state))source={label:'取得文件',state:'problem',detail:'文件校验失败'} let upload:PipelineStep={label:'上传测试库',state:'waiting',detail:'等待上传'} if(item.legacy_status==='content_alias')upload={label:'上传测试库',state:'done',detail:'复用已有内容'} else if(item.ingest_state==='uploaded')upload={label:'上传测试库',state:'done',detail:'上传完成'} else if(item.ingest_state==='uploading')upload={label:'上传测试库',state:'active',detail:'上传中'} else if(item.ingest_state==='failed')upload={label:'上传测试库',state:'problem',detail:'上传失败'} let processing:PipelineStep={label:'解析与索引',state:'waiting',detail:'等待解析'} if(item.legacy_status==='content_alias')processing={label:'解析与索引',state:'done',detail:'复用已有解析结果'} else if(item.processing_state==='completed')processing={label:'解析与索引',state:'done',detail:'解析完成'} else if(item.processing_state==='processing')processing={label:'解析与索引',state:'active',detail:'解析中'} else if(item.processing_state==='finalizing')processing={label:'解析与索引',state:'active',detail:'建立索引中'} else if(['failed','stalled'].includes(item.processing_state))processing={label:'解析与索引',state:'problem',detail:item.processing_state==='stalled'?'处理停滞':'解析失败'} let retrieval:PipelineStep={label:'测试检索',state:'waiting',detail:'未证明已启用'} if(item.retrieval_state==='enabled')retrieval={label:'测试检索',state:'done',detail:'测试环境已启用'} else if(['enabling','disabling'].includes(item.retrieval_state))retrieval={label:'测试检索',state:'active',detail:'状态变更中'} else if(item.retrieval_state==='failed')retrieval={label:'测试检索',state:'problem',detail:'检索状态失败'} return [source,upload,processing,retrieval] } export function documentStage(item: SourceAsset): DocumentStage { if (item.source_change === 'unavailable' || item.source_change === 'deleted') return 'failed' if (item.content_state === 'over_limit') return 'blocked' if (['invalid','failed'].includes(item.content_state) || item.ingest_state === 'failed' || ['failed','stalled'].includes(item.processing_state)) return 'failed' if (item.content_state === 'downloading') return 'downloading' if (item.legacy_status === 'content_alias' && !item.has_publication) return 'reused' if (!item.has_publication) return 'waiting' if (item.ingest_state === 'uploading') return 'uploading' switch (item.processing_state) { case 'completed': return 'completed' case 'processing': return 'parsing' case 'finalizing': return 'indexing' default: return 'waiting' } } export function documentProblem(item: SourceAsset): string { if (item.source_change === 'unavailable' || item.source_change === 'deleted') return '来源文件缺失或暂不可访问,请核对来源权限和文件位置。' if (documentStage(item) === 'blocked') return '超过当前导入大小上限,尚未上传;请拆分文件或由管理员调整限制。' if (item.processing_state === 'stalled') return '长时间未收到处理进展,请检查解析服务。' if (documentStage(item) === 'failed') return '处理未完成,请查看错误详情;符合重试条件时可重新处理。' if (documentStage(item) === 'reused') return '文件内容完全相同,未重复上传。' return '' } export function visibilityOrigin(item: SourceAsset): string { switch (item.visibility_origin) { case 'pp_public_tag': return '来自PP公开标记' case 'manual': return '人工指定' case 'configured_scope': return '来源范围设置' case 'file_override': return '文件单独设置' case 'directory_rule': return item.scope_rule_path ? `继承目录设置:${item.scope_rule_path}` : '继承来源范围设置' default: return '设置来源待核对' } } export function connectionKey(item: SourceAsset): string { return item.connection_id || `${item.source_type}:${item.connection_name}` } export function directoryGroups(items: SourceAsset[]) { const groups = new Map() for (const item of items) { const connection = connectionKey(item) const path = item.directory_path || '' // JSON tuple avoids cross-source collisions, even when labels contain separators. const key = JSON.stringify([connection, path]) const group = groups.get(key) || {key, connection, path, count:0} group.count++ groups.set(key, group) } return [...groups.values()].sort((a,b) => a.path.localeCompare(b.path,'zh-CN') || a.connection.localeCompare(b.connection)) } export function operationalDocuments(items:SourceAsset[],includeTestDocuments:boolean) { return includeTestDocuments ? items : items.filter(item=>item.source_type!=='manual') } export interface SourceProgress {connection:string;total:number;completed:number;active:number;attention:number;waiting:number;reused:number} export function sourceProgress(items:SourceAsset[]):SourceProgress[] { const result=new Map() for(const item of items){ const connection=connectionKey(item) const progress=result.get(connection)||{connection,total:0,completed:0,active:0,attention:0,waiting:0,reused:0} const stage=documentStage(item) progress.total++ if(stage==='completed') progress.completed++ if(['downloading','uploading','parsing','indexing'].includes(stage)) progress.active++ if(['blocked','failed'].includes(stage)) progress.attention++ if(stage==='waiting') progress.waiting++ if(stage==='reused') progress.reused++ result.set(connection,progress) } return [...result.values()].sort((a,b)=>a.connection.localeCompare(b.connection)) } export interface DocumentFilter { connection:string; directory:string|null; query:string; visibility:string; stage:string } export function filterDocuments(items: SourceAsset[], filter: DocumentFilter) { const query = filter.query.trim().toLocaleLowerCase() return items.filter(item => (!filter.connection || connectionKey(item) === filter.connection) && (filter.directory === null || (item.directory_path || '') === filter.directory) && (!filter.visibility || item.visibility === filter.visibility) && (!filter.stage || documentStage(item) === filter.stage) && (!query || `${item.title || ''} ${item.directory_path || ''}`.toLocaleLowerCase().includes(query))) } export async function fetchDocumentCatalog(signal?: AbortSignal): Promise { const items: SourceAsset[] = [] const ids = new Set() let expected: number | undefined do { const page = await fetchSourceAssets({state:'',limit:100,offset:items.length},signal) if (expected === undefined) expected = page.summary.total if (!Number.isInteger(expected) || expected < 0 || expected > 10000 || page.summary.total !== expected || page.offset !== items.length) throw new Error('文档清单已变化或过大,请刷新后重试。') for (const item of page.items) { if (!item.id || ids.has(item.id)) throw new Error('文档分页发生变化,请刷新后重试。') ids.add(item.id) items.push(item) } if (!page.items.length && items.length !== expected) throw new Error('文档清单未加载完整,请重试。') } while (items.length < expected) if (items.length !== expected) throw new Error('文档数量核验失败,请重试。') return items }