feat(batch): add batch upload for resume analysis against multiple jobs

[prathamdby]

PR Agent Description

PR Type

Enhancement

Description

• Add batch resume analysis: upload one resume PDF and compare against 2–10 job descriptions in parallel
• New NDJSON streaming API with semaphore-based concurrency (3 parallel AI calls per batch)
• Extracted shared PDF conversion and AI analysis logic into reusable lib module
• Real-time progress UI with live score cards and batch results summary page

File Walkthrough

Enhancement (10 files)

Batch analyze API route

app/api/batch-analyze/route.ts

• New POST endpoint accepting resume PDF + job entries
• Streams NDJSON progress events with semaphore-limited concurrency
• Creates BatchAnalysis record and individual Resume entries

Batch upload form with live progress

app/components/BatchUploadForm.tsx

• Step 1: resume PDF upload with validation
• Step 2: add 2–10 job entries with title, company, description
• Real-time streaming progress with per-job score cards

Batch results page with score summary

app/(dashboard)/app/batch/[id]/page.tsx

• Summary stats: total jobs, completed, failed, avg score, best match
• Grid of ResumeCard components linking to individual analyses
• Handles completed, partial, and failed batch statuses

New BatchAnalysis database model

prisma/schema.prisma

• BatchAnalysis model with status tracking (processing/completed/partial/failed)
• Resume gains optional batchId relation with cascade set-null

Extracted shared resume analysis utilities

lib/resume-analysis.ts

• convertPdfToMarkdown reused for batch and single analysis
• analyzeResumeWithAI wraps AI request with system prompt

Client-side streaming API function

lib/api.ts

• batchAnalyze() posts FormData and reads NDJSON stream
• Parses progressive events via callback pattern

Batch input validation schemas

lib/schemas.ts

• BatchJobEntrySchema: 2–10 jobs with validation
• BatchAnalyzeInputSchema wraps job array

New batch analysis TypeScript types

types/index.d.ts

• BatchAnalysis, BatchJobEntry, BatchStreamEvent types
• Stream event union covers init/progress/complete phases

Rate limit for batch endpoint

lib/rate-limit.ts

• 1 request per 120-second window
• Prevents excessive batch submissions

Navigation links for batch feature

app/components/AppSidebar.tsx

• Sidebar: new Batch Analyze nav item with Layers icon
• Navbar: new Batch tab pill linking to batch-upload

[zeus-review]

PR Agent Review

Note

This PR adds a batch resume analysis feature allowing users to upload one resume and analyze it against 2-10 job descriptions in parallel. It introduces a new NDJSON streaming API endpoint (/api/batch-analyze) with custom concurrency control (Semaphore class limiting to 3 parallel AI calls), a new BatchAnalysis database model to track batch status, two new pages (upload form with real-time progress tracking and results dashboard with score comparisons), refactored shared PDF/AI logic into lib/resume-analysis.ts, rate limiting configuration (1 request per 120 seconds), and UI integration in sidebar and navbar. The implementation spans multiple layers: API routes with streaming responses, database schema changes with new relationships, complex client components managing streaming state, and server components with auth/ownership validation.

Effort	Substantial · 4/5
Findings	No issues on this pass.
Relevant tests	no
Security	No security vulnerabilities were identified by the security specialist. The batch results page includes proper ownership validation (`batch.userId !== session.user.id`) preventing cross-user access. Rate limiting is configured appropriately for the batch endpoint. File size and type validation are in place.
CI	✅ All CI is passing
Follow-ups	Correctness specialist timed out during review - manually verify Semaphore concurrency control, NDJSON stream parsing error handling, and batch status transition logic before merge
Follow-ups	Repository has no test suite (AGENTS.md explicitly states 'No additional test suites exist') - this PR introduces complex concurrent logic that would benefit from test coverage, particularly for error paths and edge cases
Follow-ups	Verify streaming response behavior under error conditions: malformed NDJSON data, client disconnects mid-stream, and AI API timeouts to ensure graceful degradation and prevent resource waste

Coverage partial: correctness specialist failed.

_{5ddbde0 ⋅ general ⋅ 17m 49s ⋅ qwen3.7-plus}

[prathamdby]

/ask What's the purpose of this component?

[zeus-review]

This page component (app/(dashboard)/app/batch-upload/page.tsx) is a batch resume analysis page that lets a user upload a single resume and compare it against up to 10 different job descriptions in parallel.

It renders two things:

1. A header with the copy "One resume, multiple jobs. Analyzed in parallel." and a short description explaining the feature.
2. <BatchUploadForm />, which is the real workhorse. That form handles:
   • Uploading one resume PDF.
   • Adding 2 to 10 job entries (title, company, description), each with an option to import details from a URL or a PDF.
   • Validating the form (each job needs a title and a description of at least 50 characters).
   • Calling batchAnalyze with streaming progress events, then rendering a live progress UI showing each job's status (analyzing, success with score, or error).
   • Once complete, offering a "View all results" button that navigates to /app/batch/{batchId}, plus per-job "View analysis" links to /app/resume/{resumeId}.

So the page itself is just a thin layout wrapper; the bulk of the logic lives in app/components/BatchUploadForm.tsx.