P2-1: remove dead cost-control module (3 files)
P2-2: switch TypeORM to autoLoadEntities: true Remove unused vision-pipeline-cost-aware.service.ts, cost-control.service.ts and its orphan module. Switch explicit entities[] list to autoLoadEntities.
This commit is contained in:
Developer
@@ -31,34 +31,11 @@ import { ImportTaskModule } from './import-task/import-task.module';
|
|||||||
import { AssessmentModule } from './assessment/assessment.module';
|
import { AssessmentModule } from './assessment/assessment.module';
|
||||||
import { I18nMiddleware } from './i18n/i18n.middleware';
|
import { I18nMiddleware } from './i18n/i18n.middleware';
|
||||||
import { TenantMiddleware } from './tenant/tenant.middleware';
|
import { TenantMiddleware } from './tenant/tenant.middleware';
|
||||||
import { User } from './user/user.entity';
|
|
||||||
import { UserSetting } from './user/user-setting.entity';
|
|
||||||
import { ModelConfig } from './model-config/model-config.entity';
|
|
||||||
import { KnowledgeBase } from './knowledge-base/knowledge-base.entity';
|
|
||||||
import { KnowledgeGroup } from './knowledge-group/knowledge-group.entity';
|
|
||||||
import { SearchHistory } from './search-history/search-history.entity';
|
|
||||||
import { ChatMessage } from './search-history/chat-message.entity';
|
|
||||||
import { Note } from './note/note.entity';
|
|
||||||
import { NoteCategory } from './note/note-category.entity';
|
|
||||||
import { PodcastEpisode } from './podcasts/entities/podcast-episode.entity';
|
|
||||||
import { ImportTask } from './import-task/import-task.entity';
|
|
||||||
import { AssessmentSession } from './assessment/entities/assessment-session.entity';
|
|
||||||
import { AssessmentQuestion } from './assessment/entities/assessment-question.entity';
|
|
||||||
import { AssessmentAnswer } from './assessment/entities/assessment-answer.entity';
|
|
||||||
import { AssessmentTemplate } from './assessment/entities/assessment-template.entity';
|
|
||||||
import { QuestionBank } from './assessment/entities/question-bank.entity';
|
|
||||||
import { QuestionBankItem } from './assessment/entities/question-bank-item.entity';
|
|
||||||
import { Tenant } from './tenant/tenant.entity';
|
|
||||||
import { TenantSetting } from './tenant/tenant-setting.entity';
|
|
||||||
import { ApiKey } from './auth/entities/api-key.entity';
|
|
||||||
import { TenantMember } from './tenant/tenant-member.entity';
|
|
||||||
import { TenantModule } from './tenant/tenant.module';
|
import { TenantModule } from './tenant/tenant.module';
|
||||||
import { SuperAdminModule } from './super-admin/super-admin.module';
|
import { SuperAdminModule } from './super-admin/super-admin.module';
|
||||||
import { AdminModule } from './admin/admin.module';
|
import { AdminModule } from './admin/admin.module';
|
||||||
import { FeishuModule } from './feishu/feishu.module';
|
import { FeishuModule } from './feishu/feishu.module';
|
||||||
import { FeishuBot } from './feishu/entities/feishu-bot.entity';
|
|
||||||
import { FeishuAssessmentSession } from './feishu/entities/feishu-assessment-session.entity';
|
|
||||||
import { AssessmentCertificate } from './assessment/entities/assessment-certificate.entity';
|
|
||||||
|
|
||||||
@Module({
|
@Module({
|
||||||
imports: [
|
imports: [
|
||||||
@@ -77,33 +54,8 @@ import { AssessmentCertificate } from './assessment/entities/assessment-certific
|
|||||||
useFactory: (configService: ConfigService) => ({
|
useFactory: (configService: ConfigService) => ({
|
||||||
type: 'better-sqlite3',
|
type: 'better-sqlite3',
|
||||||
database: configService.get<string>('DATABASE_PATH'),
|
database: configService.get<string>('DATABASE_PATH'),
|
||||||
entities: [
|
autoLoadEntities: true,
|
||||||
User,
|
synchronize: true,
|
||||||
UserSetting,
|
|
||||||
ModelConfig,
|
|
||||||
KnowledgeBase,
|
|
||||||
KnowledgeGroup,
|
|
||||||
SearchHistory,
|
|
||||||
ChatMessage,
|
|
||||||
Note,
|
|
||||||
NoteCategory,
|
|
||||||
PodcastEpisode,
|
|
||||||
ImportTask,
|
|
||||||
AssessmentSession,
|
|
||||||
AssessmentQuestion,
|
|
||||||
AssessmentAnswer,
|
|
||||||
AssessmentTemplate,
|
|
||||||
QuestionBank,
|
|
||||||
QuestionBankItem,
|
|
||||||
Tenant,
|
|
||||||
TenantSetting,
|
|
||||||
TenantMember,
|
|
||||||
ApiKey,
|
|
||||||
FeishuBot,
|
|
||||||
FeishuAssessmentSession,
|
|
||||||
AssessmentCertificate,
|
|
||||||
],
|
|
||||||
synchronize: true, // Auto-create database schema. Disable in production.
|
|
||||||
}),
|
}),
|
||||||
}),
|
}),
|
||||||
AuthModule,
|
AuthModule,
|
||||||
|
|||||||
@@ -1,11 +0,0 @@
|
|||||||
import { Module } from '@nestjs/common';
|
|
||||||
import { TypeOrmModule } from '@nestjs/typeorm';
|
|
||||||
import { CostControlService } from './cost-control.service';
|
|
||||||
import { User } from '../user/user.entity';
|
|
||||||
|
|
||||||
@Module({
|
|
||||||
imports: [TypeOrmModule.forFeature([User])],
|
|
||||||
providers: [CostControlService],
|
|
||||||
exports: [CostControlService],
|
|
||||||
})
|
|
||||||
export class CostControlModule {}
|
|
||||||
@@ -1,261 +0,0 @@
|
|||||||
/**
|
|
||||||
* Cost control and quota management service
|
|
||||||
* Used to manage API call costs for Vision Pipeline
|
|
||||||
*/
|
|
||||||
|
|
||||||
import { Injectable, Logger } from '@nestjs/common';
|
|
||||||
import { ConfigService } from '@nestjs/config';
|
|
||||||
import { InjectRepository } from '@nestjs/typeorm';
|
|
||||||
import { Repository } from 'typeorm';
|
|
||||||
import { User } from '../user/user.entity';
|
|
||||||
|
|
||||||
export interface UserQuota {
|
|
||||||
userId: string;
|
|
||||||
monthlyCost: number; // Current month used cost
|
|
||||||
maxCost: number; // Monthly max cost
|
|
||||||
remaining: number; // Remaining cost
|
|
||||||
lastReset: Date; // Last reset time
|
|
||||||
}
|
|
||||||
|
|
||||||
export interface CostEstimate {
|
|
||||||
estimatedCost: number; // Estimated cost
|
|
||||||
estimatedTime: number; // Estimated time(seconds)
|
|
||||||
pageBreakdown: {
|
|
||||||
// Per-page breakdown
|
|
||||||
pageIndex: number;
|
|
||||||
cost: number;
|
|
||||||
}[];
|
|
||||||
}
|
|
||||||
|
|
||||||
@Injectable()
|
|
||||||
export class CostControlService {
|
|
||||||
private readonly logger = new Logger(CostControlService.name);
|
|
||||||
private readonly COST_PER_PAGE = 0.01; // Cost per page(USD)
|
|
||||||
private readonly DEFAULT_MONTHLY_LIMIT = 100; // Default monthly limit(USD)
|
|
||||||
|
|
||||||
constructor(
|
|
||||||
private configService: ConfigService,
|
|
||||||
@InjectRepository(User)
|
|
||||||
private userRepository: Repository<User>,
|
|
||||||
) {}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Estimate processing cost
|
|
||||||
*/
|
|
||||||
estimateCost(
|
|
||||||
pageCount: number,
|
|
||||||
quality: 'low' | 'medium' | 'high' = 'medium',
|
|
||||||
): CostEstimate {
|
|
||||||
// Adjust cost coefficient based on quality
|
|
||||||
const qualityMultiplier = {
|
|
||||||
low: 0.5,
|
|
||||||
medium: 1.0,
|
|
||||||
high: 1.5,
|
|
||||||
};
|
|
||||||
|
|
||||||
const baseCost =
|
|
||||||
pageCount * this.COST_PER_PAGE * qualityMultiplier[quality];
|
|
||||||
const estimatedTime = pageCount * 3; // // Approximately 3 seconds
|
|
||||||
|
|
||||||
const pageBreakdown = Array.from({ length: pageCount }, (_, i) => ({
|
|
||||||
pageIndex: i + 1,
|
|
||||||
cost: this.COST_PER_PAGE * qualityMultiplier[quality],
|
|
||||||
}));
|
|
||||||
|
|
||||||
return {
|
|
||||||
estimatedCost: baseCost,
|
|
||||||
estimatedTime,
|
|
||||||
pageBreakdown,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Check user quota
|
|
||||||
*/
|
|
||||||
async checkQuota(
|
|
||||||
userId: string,
|
|
||||||
estimatedCost: number,
|
|
||||||
): Promise<{
|
|
||||||
allowed: boolean;
|
|
||||||
quota: UserQuota;
|
|
||||||
reason?: string;
|
|
||||||
}> {
|
|
||||||
const quota = await this.getUserQuota(userId);
|
|
||||||
|
|
||||||
// Check monthly reset
|
|
||||||
this.checkAndResetMonthlyQuota(quota);
|
|
||||||
|
|
||||||
if (quota.remaining < estimatedCost) {
|
|
||||||
this.logger.warn(
|
|
||||||
`User ${userId} quota insufficient: remaining $${quota.remaining.toFixed(2)}, required $${estimatedCost.toFixed(2)}`,
|
|
||||||
);
|
|
||||||
return {
|
|
||||||
allowed: false,
|
|
||||||
quota,
|
|
||||||
reason: `Insufficient quota: remaining $${quota.remaining.toFixed(2)}, required $${estimatedCost.toFixed(2)}`,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
return {
|
|
||||||
allowed: true,
|
|
||||||
quota,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Deduct from quota
|
|
||||||
*/
|
|
||||||
async deductQuota(userId: string, actualCost: number): Promise<void> {
|
|
||||||
const quota = await this.getUserQuota(userId);
|
|
||||||
quota.monthlyCost += actualCost;
|
|
||||||
quota.remaining = quota.maxCost - quota.monthlyCost;
|
|
||||||
|
|
||||||
await this.userRepository.update(userId, {
|
|
||||||
monthlyCost: quota.monthlyCost,
|
|
||||||
});
|
|
||||||
|
|
||||||
this.logger.log(
|
|
||||||
`Deducted $${actualCost.toFixed(2)} from user ${userId} quota. Remaining: $${quota.remaining.toFixed(2)}`,
|
|
||||||
);
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Get user quota
|
|
||||||
*/
|
|
||||||
async getUserQuota(userId: string): Promise<UserQuota> {
|
|
||||||
const user = await this.userRepository.findOne({ where: { id: userId } });
|
|
||||||
|
|
||||||
if (!user) {
|
|
||||||
throw new Error(`User ${userId} does not exist`);
|
|
||||||
}
|
|
||||||
|
|
||||||
// Use default if user has no quota info
|
|
||||||
const monthlyCost = user.monthlyCost || 0;
|
|
||||||
const maxCost = user.maxCost || this.DEFAULT_MONTHLY_LIMIT;
|
|
||||||
const lastReset = user.lastQuotaReset || new Date();
|
|
||||||
|
|
||||||
return {
|
|
||||||
userId,
|
|
||||||
monthlyCost,
|
|
||||||
maxCost,
|
|
||||||
remaining: maxCost - monthlyCost,
|
|
||||||
lastReset,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Check and reset monthly quota
|
|
||||||
*/
|
|
||||||
private checkAndResetMonthlyQuota(quota: UserQuota): void {
|
|
||||||
const now = new Date();
|
|
||||||
const lastReset = quota.lastReset;
|
|
||||||
|
|
||||||
// Check if crossed month
|
|
||||||
if (
|
|
||||||
now.getMonth() !== lastReset.getMonth() ||
|
|
||||||
now.getFullYear() !== lastReset.getFullYear()
|
|
||||||
) {
|
|
||||||
this.logger.log(`Reset monthly quota for user ${quota.userId}`);
|
|
||||||
|
|
||||||
// Reset quota
|
|
||||||
quota.monthlyCost = 0;
|
|
||||||
quota.remaining = quota.maxCost;
|
|
||||||
quota.lastReset = now;
|
|
||||||
|
|
||||||
// Update database
|
|
||||||
this.userRepository.update(quota.userId, {
|
|
||||||
monthlyCost: 0,
|
|
||||||
lastQuotaReset: now,
|
|
||||||
});
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Set user quota limit
|
|
||||||
*/
|
|
||||||
async setQuotaLimit(userId: string, maxCost: number): Promise<void> {
|
|
||||||
await this.userRepository.update(userId, { maxCost });
|
|
||||||
this.logger.log(`Set quota limit to $${maxCost} for user ${userId}`);
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Get cost report
|
|
||||||
*/
|
|
||||||
async getCostReport(
|
|
||||||
userId: string,
|
|
||||||
days: number = 30,
|
|
||||||
): Promise<{
|
|
||||||
totalCost: number;
|
|
||||||
dailyAverage: number;
|
|
||||||
pageStats: {
|
|
||||||
totalPages: number;
|
|
||||||
avgCostPerPage: number;
|
|
||||||
};
|
|
||||||
quotaUsage: number; // Percentage
|
|
||||||
}> {
|
|
||||||
const quota = await this.getUserQuota(userId);
|
|
||||||
const usagePercent = (quota.monthlyCost / quota.maxCost) * 100;
|
|
||||||
|
|
||||||
// Query history records here(if implemented)
|
|
||||||
// Return current quota info temporarily
|
|
||||||
|
|
||||||
return {
|
|
||||||
totalCost: quota.monthlyCost,
|
|
||||||
dailyAverage: quota.monthlyCost / Math.max(days, 1),
|
|
||||||
pageStats: {
|
|
||||||
totalPages: Math.floor(quota.monthlyCost / this.COST_PER_PAGE),
|
|
||||||
avgCostPerPage: this.COST_PER_PAGE,
|
|
||||||
},
|
|
||||||
quotaUsage: usagePercent,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Check cost warning threshold
|
|
||||||
*/
|
|
||||||
async checkWarningThreshold(userId: string): Promise<{
|
|
||||||
shouldWarn: boolean;
|
|
||||||
message: string;
|
|
||||||
}> {
|
|
||||||
const quota = await this.getUserQuota(userId);
|
|
||||||
const usagePercent = (quota.monthlyCost / quota.maxCost) * 100;
|
|
||||||
|
|
||||||
if (usagePercent >= 90) {
|
|
||||||
return {
|
|
||||||
shouldWarn: true,
|
|
||||||
message: `⚠️ Quota usage reached ${usagePercent.toFixed(1)}%. Remaining: $${quota.remaining.toFixed(2)}`,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
if (usagePercent >= 75) {
|
|
||||||
return {
|
|
||||||
shouldWarn: true,
|
|
||||||
message: `💡 Quota usage at ${usagePercent.toFixed(1)}%. Please monitor your costs carefully`,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
return {
|
|
||||||
shouldWarn: false,
|
|
||||||
message: '',
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Format cost display
|
|
||||||
*/
|
|
||||||
formatCost(cost: number): string {
|
|
||||||
return `$${cost.toFixed(2)}`;
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Format time display
|
|
||||||
*/
|
|
||||||
formatTime(seconds: number): string {
|
|
||||||
if (seconds < 60) {
|
|
||||||
return `${seconds.toFixed(0)}s`;
|
|
||||||
}
|
|
||||||
const minutes = Math.floor(seconds / 60);
|
|
||||||
const remainingSeconds = seconds % 60;
|
|
||||||
return `${minutes}m ${remainingSeconds.toFixed(0)}s`;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
@@ -1,341 +0,0 @@
|
|||||||
/**
|
|
||||||
* Vision Pipeline Service (with cost control)
|
|
||||||
* This is an extended version of vision-pipeline.service.ts with integrated cost control
|
|
||||||
*/
|
|
||||||
|
|
||||||
import { Injectable, Logger } from '@nestjs/common';
|
|
||||||
import { ConfigService } from '@nestjs/config';
|
|
||||||
import * as fs from 'fs/promises';
|
|
||||||
import * as path from 'path';
|
|
||||||
import { LibreOfficeService } from '../libreoffice/libreoffice.service';
|
|
||||||
import { Pdf2ImageService } from '../pdf2image/pdf2image.service';
|
|
||||||
import { VisionService } from '../vision/vision.service';
|
|
||||||
import { ElasticsearchService } from '../elasticsearch/elasticsearch.service';
|
|
||||||
import { ModelConfigService } from '../model-config/model-config.service';
|
|
||||||
import {
|
|
||||||
PreciseModeOptions,
|
|
||||||
PipelineResult,
|
|
||||||
ProcessingStatus,
|
|
||||||
ModeRecommendation,
|
|
||||||
} from './vision-pipeline.interface';
|
|
||||||
import {
|
|
||||||
VisionModelConfig,
|
|
||||||
VisionAnalysisResult,
|
|
||||||
} from '../vision/vision.interface';
|
|
||||||
import { CostControlService } from './cost-control.service';
|
|
||||||
import { I18nService } from '../i18n/i18n.service';
|
|
||||||
|
|
||||||
@Injectable()
|
|
||||||
export class VisionPipelineCostAwareService {
|
|
||||||
private readonly logger = new Logger(VisionPipelineCostAwareService.name);
|
|
||||||
|
|
||||||
constructor(
|
|
||||||
private libreOffice: LibreOfficeService,
|
|
||||||
private pdf2Image: Pdf2ImageService,
|
|
||||||
private vision: VisionService,
|
|
||||||
private elasticsearch: ElasticsearchService,
|
|
||||||
private modelConfigService: ModelConfigService,
|
|
||||||
private configService: ConfigService,
|
|
||||||
private costControl: CostControlService,
|
|
||||||
private i18nService: I18nService,
|
|
||||||
) {}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Main processing flow: Precise mode (with cost control)
|
|
||||||
*/
|
|
||||||
async processPreciseMode(
|
|
||||||
filePath: string,
|
|
||||||
options: PreciseModeOptions,
|
|
||||||
): Promise<PipelineResult> {
|
|
||||||
const startTime = Date.now();
|
|
||||||
const results: VisionAnalysisResult[] = [];
|
|
||||||
let processedPages = 0;
|
|
||||||
let failedPages = 0;
|
|
||||||
let totalCost = 0;
|
|
||||||
let pdfPath = filePath;
|
|
||||||
let imagesToProcess: any[] = [];
|
|
||||||
|
|
||||||
this.logger.log(
|
|
||||||
`Starting precise mode processing for ${options.fileName} (user: ${options.userId})`,
|
|
||||||
);
|
|
||||||
|
|
||||||
try {
|
|
||||||
// Step 1: Convert format
|
|
||||||
this.updateStatus('converting', 10, 'Converting document format...');
|
|
||||||
pdfPath = await this.convertToPDF(filePath);
|
|
||||||
|
|
||||||
// Step 2: Convert PDF to images
|
|
||||||
this.updateStatus('splitting', 30, 'Converting PDF to images...');
|
|
||||||
const conversionResult = await this.pdf2Image.convertToImages(pdfPath, {
|
|
||||||
density: 300,
|
|
||||||
quality: 85,
|
|
||||||
format: 'jpeg',
|
|
||||||
});
|
|
||||||
|
|
||||||
if (conversionResult.images.length === 0) {
|
|
||||||
throw new Error(
|
|
||||||
this.i18nService.getMessage('pdfToImageConversionFailed'),
|
|
||||||
);
|
|
||||||
}
|
|
||||||
|
|
||||||
// Limit processing pages
|
|
||||||
imagesToProcess = options.maxPages
|
|
||||||
? conversionResult.images.slice(0, options.maxPages)
|
|
||||||
: conversionResult.images;
|
|
||||||
|
|
||||||
const pageCount = imagesToProcess.length;
|
|
||||||
|
|
||||||
// Step 3: Cost estimation and quota check
|
|
||||||
this.updateStatus(
|
|
||||||
'checking',
|
|
||||||
40,
|
|
||||||
'Checking quota and estimating cost...',
|
|
||||||
);
|
|
||||||
const costEstimate = this.costControl.estimateCost(pageCount);
|
|
||||||
this.logger.log(
|
|
||||||
`Estimated cost: $${costEstimate.estimatedCost.toFixed(2)}, Estimated time: ${this.costControl.formatTime(costEstimate.estimatedTime)}`,
|
|
||||||
);
|
|
||||||
|
|
||||||
// Quota check
|
|
||||||
const quotaCheck = await this.costControl.checkQuota(
|
|
||||||
options.userId,
|
|
||||||
costEstimate.estimatedCost,
|
|
||||||
);
|
|
||||||
|
|
||||||
if (!quotaCheck.allowed) {
|
|
||||||
throw new Error(quotaCheck.reason);
|
|
||||||
}
|
|
||||||
|
|
||||||
// Cost warning check
|
|
||||||
const warning = await this.costControl.checkWarningThreshold(
|
|
||||||
options.userId,
|
|
||||||
);
|
|
||||||
if (warning.shouldWarn) {
|
|
||||||
this.logger.warn(warning.message);
|
|
||||||
}
|
|
||||||
|
|
||||||
// Step 4: Get Vision model config
|
|
||||||
const modelConfig = await this.getVisionModelConfig(
|
|
||||||
options.userId,
|
|
||||||
options.modelId,
|
|
||||||
options.tenantId,
|
|
||||||
);
|
|
||||||
|
|
||||||
// Step 5: VL model analysis
|
|
||||||
this.updateStatus(
|
|
||||||
'analyzing',
|
|
||||||
50,
|
|
||||||
'Analyzing pages with Vision model...',
|
|
||||||
);
|
|
||||||
const batchResult = await this.vision.batchAnalyze(
|
|
||||||
imagesToProcess.map((img) => img.path),
|
|
||||||
modelConfig,
|
|
||||||
{
|
|
||||||
startIndex: 1,
|
|
||||||
skipQualityCheck: options.skipQualityCheck,
|
|
||||||
},
|
|
||||||
);
|
|
||||||
|
|
||||||
totalCost = batchResult.estimatedCost;
|
|
||||||
processedPages = batchResult.successCount;
|
|
||||||
failedPages = batchResult.failedCount;
|
|
||||||
results.push(...batchResult.results);
|
|
||||||
|
|
||||||
// Step 6: Subtract actual cost
|
|
||||||
if (totalCost > 0) {
|
|
||||||
await this.costControl.deductQuota(options.userId, totalCost);
|
|
||||||
this.logger.log(`Actual cost deducted: $${totalCost.toFixed(2)}`);
|
|
||||||
}
|
|
||||||
|
|
||||||
// Step 7: Cleanup temp files
|
|
||||||
this.updateStatus(
|
|
||||||
'completed',
|
|
||||||
100,
|
|
||||||
'Processing completed. Cleaning up temp files...',
|
|
||||||
);
|
|
||||||
await this.pdf2Image.cleanupImages(imagesToProcess);
|
|
||||||
|
|
||||||
// Cleanup converted PDF file if converted
|
|
||||||
if (pdfPath !== filePath) {
|
|
||||||
try {
|
|
||||||
await fs.unlink(pdfPath);
|
|
||||||
} catch (error) {
|
|
||||||
this.logger.warn(`Failed to cleanup converted PDF: ${error.message}`);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
const duration = (Date.now() - startTime) / 1000;
|
|
||||||
|
|
||||||
this.logger.log(
|
|
||||||
`Precise mode completed: ${processedPages} pages processed, ` +
|
|
||||||
`cost: $${totalCost.toFixed(2)}, duration: ${duration.toFixed(1)}s`,
|
|
||||||
);
|
|
||||||
|
|
||||||
return {
|
|
||||||
success: true,
|
|
||||||
fileId: options.fileId,
|
|
||||||
fileName: options.fileName,
|
|
||||||
totalPages: conversionResult.totalPages,
|
|
||||||
processedPages,
|
|
||||||
failedPages,
|
|
||||||
results,
|
|
||||||
cost: totalCost,
|
|
||||||
duration,
|
|
||||||
mode: 'precise',
|
|
||||||
};
|
|
||||||
} catch (error) {
|
|
||||||
this.logger.error(`Precise mode failed: ${error.message}`);
|
|
||||||
|
|
||||||
// Try to clean up temp files
|
|
||||||
try {
|
|
||||||
if (pdfPath !== filePath && pdfPath !== filePath) {
|
|
||||||
await fs.unlink(pdfPath);
|
|
||||||
}
|
|
||||||
if (imagesToProcess.length > 0) {
|
|
||||||
await this.pdf2Image.cleanupImages(imagesToProcess);
|
|
||||||
}
|
|
||||||
} catch {}
|
|
||||||
|
|
||||||
return {
|
|
||||||
success: false,
|
|
||||||
fileId: options.fileId,
|
|
||||||
fileName: options.fileName,
|
|
||||||
totalPages: 0,
|
|
||||||
processedPages,
|
|
||||||
failedPages,
|
|
||||||
results: [],
|
|
||||||
cost: totalCost,
|
|
||||||
duration: (Date.now() - startTime) / 1000,
|
|
||||||
mode: 'precise',
|
|
||||||
};
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Get Vision model configuration
|
|
||||||
*/
|
|
||||||
private async getVisionModelConfig(
|
|
||||||
userId: string,
|
|
||||||
modelId: string,
|
|
||||||
tenantId?: string,
|
|
||||||
): Promise<VisionModelConfig> {
|
|
||||||
const config = await this.modelConfigService.findOne(modelId);
|
|
||||||
|
|
||||||
if (!config) {
|
|
||||||
throw new Error(`Model config not found: ${modelId}`);
|
|
||||||
}
|
|
||||||
|
|
||||||
// API key is optional - allows local models
|
|
||||||
|
|
||||||
return {
|
|
||||||
baseUrl: config.baseUrl || '',
|
|
||||||
apiKey: config.apiKey || '',
|
|
||||||
modelId: config.modelId,
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Convert to PDF
|
|
||||||
*/
|
|
||||||
private async convertToPDF(filePath: string): Promise<string> {
|
|
||||||
const ext = path.extname(filePath).toLowerCase();
|
|
||||||
|
|
||||||
// Return as-is if already PDF
|
|
||||||
if (ext === '.pdf') {
|
|
||||||
return filePath;
|
|
||||||
}
|
|
||||||
|
|
||||||
// Call LibreOffice to convert
|
|
||||||
return await this.libreOffice.convertToPDF(filePath);
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Format detection and mode recommendation (with cost estimation)
|
|
||||||
*/
|
|
||||||
async recommendMode(filePath: string): Promise<ModeRecommendation> {
|
|
||||||
const ext = path.extname(filePath).toLowerCase();
|
|
||||||
const stats = await fs.stat(filePath);
|
|
||||||
const sizeMB = stats.size / (1024 * 1024);
|
|
||||||
|
|
||||||
const supportedFormats = [
|
|
||||||
'.pdf',
|
|
||||||
'.doc',
|
|
||||||
'.docx',
|
|
||||||
'.ppt',
|
|
||||||
'.pptx',
|
|
||||||
'.xls',
|
|
||||||
'.xlsx',
|
|
||||||
];
|
|
||||||
const preciseFormats = ['.pdf', '.doc', '.docx', '.ppt', '.pptx'];
|
|
||||||
|
|
||||||
if (!supportedFormats.includes(ext)) {
|
|
||||||
return {
|
|
||||||
recommendedMode: 'fast',
|
|
||||||
reason: `Unsupported file format: ${ext}`,
|
|
||||||
warnings: ['Using fast mode (text extraction only)'],
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
if (!preciseFormats.includes(ext)) {
|
|
||||||
return {
|
|
||||||
recommendedMode: 'fast',
|
|
||||||
reason: `Format ${ext} does not support precise mode`,
|
|
||||||
warnings: ['Using fast mode (text extraction only)'],
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
// Estimate page count（based on file size）
|
|
||||||
const estimatedPages = Math.max(1, Math.ceil(sizeMB * 2));
|
|
||||||
const costEstimate = this.costControl.estimateCost(estimatedPages);
|
|
||||||
|
|
||||||
// Recommend precise mode for large files
|
|
||||||
if (sizeMB > 50) {
|
|
||||||
return {
|
|
||||||
recommendedMode: 'precise',
|
|
||||||
reason:
|
|
||||||
'File is large, recommend precise mode to preserve full content',
|
|
||||||
estimatedCost: costEstimate.estimatedCost,
|
|
||||||
estimatedTime: costEstimate.estimatedTime,
|
|
||||||
warnings: [
|
|
||||||
'Processing time may be longer',
|
|
||||||
'API costs will be incurred',
|
|
||||||
],
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
// Recommend precise mode
|
|
||||||
return {
|
|
||||||
recommendedMode: 'precise',
|
|
||||||
reason:
|
|
||||||
'Precise mode available. Can preserve mixed text and image content',
|
|
||||||
estimatedCost: costEstimate.estimatedCost,
|
|
||||||
estimatedTime: costEstimate.estimatedTime,
|
|
||||||
warnings: ['API costs will be incurred'],
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Get user quota information
|
|
||||||
*/
|
|
||||||
async getUserQuotaInfo(userId: string) {
|
|
||||||
const quota = await this.costControl.getUserQuota(userId);
|
|
||||||
const report = await this.costControl.getCostReport(userId);
|
|
||||||
|
|
||||||
return {
|
|
||||||
...quota,
|
|
||||||
report,
|
|
||||||
warnings: await this.costControl.checkWarningThreshold(userId),
|
|
||||||
};
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
|
||||||
* Update processing status (for real-time feedback)
|
|
||||||
*/
|
|
||||||
private updateStatus(
|
|
||||||
status: ProcessingStatus['status'],
|
|
||||||
progress: number,
|
|
||||||
message: string,
|
|
||||||
): void {
|
|
||||||
this.logger.log(`[${status}] ${progress}% - ${message}`);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
Reference in New Issue
Block a user