feat: implement QuestionBank CRUD with pagination and template query
- Add pagination support to findAll (page, limit query params) - Add findByTemplateId method to service - Add GET /by-template/:templateId endpoint to controller - Service already includes CRUD for QuestionBank and QuestionBankItem
This commit is contained in:
Developer
@@ -0,0 +1,11 @@
|
||||
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 {}
|
||||
@@ -0,0 +1,261 @@
|
||||
/**
|
||||
* 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`;
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,341 @@
|
||||
/**
|
||||
* 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}`);
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,56 @@
|
||||
/**
|
||||
* Vision Pipeline Interface Definitions
|
||||
*/
|
||||
|
||||
import { VisionAnalysisResult } from '../vision/vision.interface';
|
||||
|
||||
export interface PreciseModeOptions {
|
||||
userId: string;
|
||||
tenantId: string;
|
||||
modelId: string;
|
||||
fileId: string;
|
||||
fileName: string;
|
||||
skipQualityCheck?: boolean;
|
||||
maxPages?: number;
|
||||
}
|
||||
|
||||
export interface PipelineResult {
|
||||
success: boolean;
|
||||
fileId: string;
|
||||
fileName: string;
|
||||
totalPages: number;
|
||||
processedPages: number;
|
||||
failedPages: number;
|
||||
results: VisionAnalysisResult[];
|
||||
cost: number;
|
||||
duration: number; // seconds
|
||||
mode: 'precise';
|
||||
}
|
||||
|
||||
export interface ProcessingStatus {
|
||||
status:
|
||||
| 'converting'
|
||||
| 'splitting'
|
||||
| 'checking'
|
||||
| 'analyzing'
|
||||
| 'indexing'
|
||||
| 'completed'
|
||||
| 'failed';
|
||||
progress: number;
|
||||
message: string;
|
||||
cost?: number;
|
||||
}
|
||||
|
||||
export interface FileFormat {
|
||||
extension: string;
|
||||
needsConversion: boolean;
|
||||
supported: boolean;
|
||||
}
|
||||
|
||||
export interface ModeRecommendation {
|
||||
recommendedMode: 'precise' | 'fast';
|
||||
reason: string;
|
||||
estimatedCost?: number;
|
||||
estimatedTime?: number; // seconds
|
||||
warnings?: string[];
|
||||
}
|
||||
@@ -0,0 +1,18 @@
|
||||
import { Module } from '@nestjs/common';
|
||||
import { VisionPipelineService } from './vision-pipeline.service';
|
||||
import { LibreOfficeModule } from '../libreoffice/libreoffice.module';
|
||||
import { Pdf2ImageModule } from '../pdf2image/pdf2image.module';
|
||||
import { VisionModule } from '../vision/vision.module';
|
||||
import { ModelConfigModule } from '../model-config/model-config.module';
|
||||
|
||||
@Module({
|
||||
imports: [
|
||||
LibreOfficeModule,
|
||||
Pdf2ImageModule,
|
||||
VisionModule,
|
||||
ModelConfigModule,
|
||||
],
|
||||
providers: [VisionPipelineService],
|
||||
exports: [VisionPipelineService],
|
||||
})
|
||||
export class VisionPipelineModule {}
|
||||
@@ -0,0 +1,364 @@
|
||||
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 { ModelConfigService } from '../model-config/model-config.service';
|
||||
import {
|
||||
ModeRecommendation,
|
||||
PipelineResult,
|
||||
PreciseModeOptions,
|
||||
ProcessingStatus,
|
||||
} from './vision-pipeline.interface';
|
||||
import {
|
||||
VisionAnalysisResult,
|
||||
VisionModelConfig,
|
||||
} from '../vision/vision.interface';
|
||||
import { I18nService } from '../i18n/i18n.service';
|
||||
|
||||
@Injectable()
|
||||
export class VisionPipelineService {
|
||||
private readonly logger = new Logger(VisionPipelineService.name);
|
||||
|
||||
constructor(
|
||||
private libreOffice: LibreOfficeService,
|
||||
private pdf2Image: Pdf2ImageService,
|
||||
private vision: VisionService,
|
||||
private modelConfigService: ModelConfigService,
|
||||
private configService: ConfigService,
|
||||
private i18nService: I18nService,
|
||||
) {}
|
||||
|
||||
/**
|
||||
* Main processing flow: Precise mode
|
||||
* Returns the processing result, and the caller is responsible for vectorization and indexing.
|
||||
*/
|
||||
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: ${options.fileName} (User: ${options.userId})`,
|
||||
);
|
||||
|
||||
try {
|
||||
// Step 1: Unification of formats
|
||||
this.logger.log('📄 Step 1/4: Unification of formats');
|
||||
this.updateStatus('converting', 10, 'Converting document format...');
|
||||
try {
|
||||
pdfPath = await this.convertToPDF(filePath);
|
||||
this.logger.log(`✅ Format conversion completed: ${pdfPath}`);
|
||||
} catch (convertError) {
|
||||
this.logger.error(
|
||||
`❌ Format conversion failed: ${convertError.message}`,
|
||||
);
|
||||
throw convertError;
|
||||
}
|
||||
|
||||
// Step 2: Conversion from PDF to images
|
||||
this.logger.log('🖼️ Step 2/4: Conversion from PDF to images');
|
||||
this.updateStatus('splitting', 30, 'Converting PDF to images...');
|
||||
let conversionResult;
|
||||
try {
|
||||
conversionResult = await this.pdf2Image.convertToImages(pdfPath, {
|
||||
density: 300,
|
||||
quality: 85,
|
||||
format: 'jpeg',
|
||||
});
|
||||
} catch (imageError) {
|
||||
this.logger.error(
|
||||
`❌ PDF to image conversion failed: ${imageError.message}`,
|
||||
);
|
||||
throw imageError;
|
||||
}
|
||||
|
||||
if (conversionResult.images.length === 0) {
|
||||
throw new Error(
|
||||
this.i18nService.getMessage('pdfToImageConversionFailed'),
|
||||
);
|
||||
}
|
||||
|
||||
this.logger.log(
|
||||
`✅ PDF to image conversion completed: Total ${conversionResult.totalPages} pages, ${conversionResult.images.length} images generated`,
|
||||
);
|
||||
|
||||
// Limit the number of pages to process
|
||||
imagesToProcess = options.maxPages
|
||||
? conversionResult.images.slice(0, options.maxPages)
|
||||
: conversionResult.images;
|
||||
|
||||
this.logger.log(
|
||||
`📊 Processing ${imagesToProcess.length} pages (${options.maxPages ? 'limited' : 'all'})`,
|
||||
);
|
||||
|
||||
// Step 3: Get Vision model configuration
|
||||
this.logger.log('🤖 Step 3/4: Preparation of Vision model');
|
||||
const modelConfig = await this.getVisionModelConfig(
|
||||
options.modelId,
|
||||
);
|
||||
this.logger.log(
|
||||
`✅ Vision model configuration completed: ${modelConfig.modelId}`,
|
||||
);
|
||||
|
||||
// Step 4: VL model analysis
|
||||
this.logger.log('🔍 Step 4/4: Vision model analysis');
|
||||
this.updateStatus(
|
||||
'analyzing',
|
||||
50,
|
||||
'Analyzing pages using Vision model...',
|
||||
);
|
||||
|
||||
// Display processing progress of each page
|
||||
this.logger.log(
|
||||
`Starting analysis of ${imagesToProcess.length} page contents...`,
|
||||
);
|
||||
const batchResult = await this.vision.batchAnalyze(
|
||||
imagesToProcess.map((img) => img.path),
|
||||
modelConfig,
|
||||
{
|
||||
startIndex: 1,
|
||||
skipQualityCheck: options.skipQualityCheck,
|
||||
onProgress: (current: number, total: number, pageResult?: any) => {
|
||||
const progress = Math.round((current / total) * 100);
|
||||
this.logger.log(
|
||||
`📄 Processing progress: ${current}/${total} (${progress}%) ${pageResult ? `- Page ${pageResult.pageIndex} completed` : ''}`,
|
||||
);
|
||||
},
|
||||
},
|
||||
);
|
||||
|
||||
totalCost = batchResult.estimatedCost;
|
||||
processedPages = batchResult.successCount;
|
||||
failedPages = batchResult.failedCount;
|
||||
results.push(...batchResult.results);
|
||||
|
||||
this.logger.log(
|
||||
`✅ Vision analysis completed: Success ${processedPages} pages, Fail ${failedPages} pages, Cost $${totalCost.toFixed(2)}`,
|
||||
);
|
||||
|
||||
// Step 5: Cleanup of temporary files (images)
|
||||
this.logger.log('🧹 Cleaning up temporary files...');
|
||||
this.updateStatus(
|
||||
'completed',
|
||||
100,
|
||||
'Processing completed. Cleaning up temporary files...',
|
||||
);
|
||||
await this.pdf2Image.cleanupImages(imagesToProcess);
|
||||
|
||||
// If converted to PDF, clean up the converted file
|
||||
if (pdfPath !== filePath) {
|
||||
try {
|
||||
await fs.unlink(pdfPath);
|
||||
this.logger.log('🗑️ Cleaned up converted PDF file');
|
||||
} catch (error) {
|
||||
this.logger.warn(
|
||||
`⚠️ Failed to clean up converted PDF: ${error.message}`,
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
const duration = (Date.now() - startTime) / 1000;
|
||||
|
||||
this.logger.log(
|
||||
`🎉 Precise mode processing completed! ` +
|
||||
`📊 Statistics: ${processedPages}/${imagesToProcess.length} pages success, ` +
|
||||
`💰 Cost: $${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 processing failed: ${error.message}`);
|
||||
|
||||
// Attempting cleanup of temporary files
|
||||
try {
|
||||
if (pdfPath !== filePath && pdfPath !== filePath) {
|
||||
await fs.unlink(pdfPath);
|
||||
}
|
||||
if (imagesToProcess.length > 0) {
|
||||
await this.pdf2Image.cleanupImages(imagesToProcess);
|
||||
}
|
||||
this.logger.log('🧹 Cleaned up temporary files');
|
||||
} catch {
|
||||
this.logger.warn('⚠️ Failed to clean up temporary files');
|
||||
}
|
||||
|
||||
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(
|
||||
modelId: string,
|
||||
): Promise<VisionModelConfig> {
|
||||
const config = await this.modelConfigService.findOne(
|
||||
modelId,
|
||||
);
|
||||
|
||||
if (!config) {
|
||||
throw new Error(`Model configuration not found: ${modelId}`);
|
||||
}
|
||||
|
||||
// API key is optional - Allows local models
|
||||
|
||||
return {
|
||||
baseUrl: config.baseUrl || '',
|
||||
apiKey: config.apiKey || '',
|
||||
modelId: config.modelId,
|
||||
};
|
||||
}
|
||||
|
||||
/**
|
||||
* Conversion to PDF
|
||||
*/
|
||||
private async convertToPDF(filePath: string): Promise<string> {
|
||||
const ext = path.extname(filePath).toLowerCase();
|
||||
|
||||
// If already PDF, return as is
|
||||
if (ext === '.pdf') {
|
||||
return filePath;
|
||||
}
|
||||
|
||||
// Call LibreOffice to convert
|
||||
const containerPdfPath = await this.libreOffice.convertToPDF(filePath);
|
||||
|
||||
// The path returned from the LibreOffice container is already correct. All point to the same uploads directory.
|
||||
// Inside container: /uploads/xxx.pdf -> Host machine: ../uploads/xxx.pdf
|
||||
const hostPdfPath = containerPdfPath.startsWith('/uploads/')
|
||||
? path.join('..', containerPdfPath) // ../uploads/xxx.pdf
|
||||
: containerPdfPath;
|
||||
|
||||
this.logger.log(`Path conversion: ${containerPdfPath} -> ${hostPdfPath}`);
|
||||
|
||||
// Check existence of file
|
||||
try {
|
||||
await fs.access(hostPdfPath);
|
||||
return hostPdfPath;
|
||||
} catch (error) {
|
||||
this.logger.error(`PDF file does not exist: ${hostPdfPath}`);
|
||||
throw new Error(`PDF file does not exist: ${hostPdfPath}`);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Index results to Elasticsearch (using knowledge-base embedding service)
|
||||
* Note: This method requires the embedding service, so it should be called within knowledge-base.service.
|
||||
*/
|
||||
private async indexResults(
|
||||
results: any[],
|
||||
options: PreciseModeOptions,
|
||||
): Promise<void> {
|
||||
// This method is currently called from knowledge-base.service
|
||||
// vision-pipeline is only responsible for processing and returning results.
|
||||
this.logger.log(
|
||||
`indexResults called with ${results.length} results - should be handled by knowledge-base service`,
|
||||
);
|
||||
throw new Error(
|
||||
'VisionPipelineService.indexResults should not be called directly. Use knowledge-base service instead.',
|
||||
);
|
||||
}
|
||||
|
||||
/**
|
||||
* Format detection and mode recommendation
|
||||
*/
|
||||
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 ${ext} does not support precise mode`,
|
||||
warnings: ['Using fast mode (text extraction only)'],
|
||||
};
|
||||
}
|
||||
|
||||
// File size check
|
||||
if (sizeMB > 50) {
|
||||
return {
|
||||
recommendedMode: 'precise',
|
||||
reason:
|
||||
'The file is large, so precise mode is recommended to retain full information.',
|
||||
estimatedCost: sizeMB * 0.01, // Rough estimate
|
||||
estimatedTime: sizeMB * 12, // Approx. 12 seconds per 1MB
|
||||
warnings: ['Processing time may be long', 'API costs will occur'],
|
||||
};
|
||||
}
|
||||
|
||||
// Precise mode recommended
|
||||
return {
|
||||
recommendedMode: 'precise',
|
||||
reason:
|
||||
'Precise mode is available. Can retain mixed content of text and images.',
|
||||
estimatedCost: sizeMB * 0.01,
|
||||
estimatedTime: sizeMB * 10,
|
||||
warnings: ['API costs will occur'],
|
||||
};
|
||||
}
|
||||
|
||||
/**
|
||||
* Update processing status (for real-time feedback)
|
||||
*/
|
||||
private updateStatus(
|
||||
status: ProcessingStatus['status'],
|
||||
progress: number,
|
||||
message: string,
|
||||
): void {
|
||||
// You can send WebSocket messages or update the database here.
|
||||
this.logger.log(`[${status}] ${progress}% - ${message}`);
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user