Dataset Viewer
Auto-converted to Parquet Duplicate
id
string
dataset
string
split
string
prompt
string
text
string
label
string
sample_001
50-test-cases-for-code-generation-qu
train
Evaluate case 1: 50 test cases for code generation quality
# 50 test cases for code generation quality
edge_case
sample_002
50-test-cases-for-code-generation-qu
train
Evaluate case 2: 50 test cases for code generation quality
# AI Evaluation Benchmark: Storefront Code Generation (50 Test Cases) **Version:** 1.0 | **Classification:** Production-Grade Evaluation Corpus | **Target:** LLM Code Generation Quality Assessment
pass
sample_003
50-test-cases-for-code-generation-qu
train
Evaluate case 3: 50 test cases for code generation quality
## 1. Benchmark Architecture & Scoring Matrix (The Rubric) All 50 cases are evaluated against this **5-Dimension, 5-Point Likert Scale**. Annotators *must* provide line-level evidence for scores ≤ 3. | Dimension | Weight | 1 - Critical Failure | 2 - Major Defects | 3 - Acceptable w/ Issues | 4 - Production Ready | 5 ...
pass
sample_004
50-test-cases-for-code-generation-qu
train
Evaluate case 4: 50 test cases for code generation quality
## 2. The 50 Test Case Catalog (Storefront Domain) *Categorized by Domain & Difficulty. Full JSON Schema provided in Section 6.* | ID | Category | Difficulty | Prompt Summary | Key Evaluation Traps | | :--- | :--- | :--- | :--- | :--- | | **SF-001** | **Catalog/Read** | Easy | `getProductById(id: string): Product` | ...
pass
sample_005
50-test-cases-for-code-generation-qu
train
Evaluate case 5: 50 test cases for code generation quality
# 50 test cases for code generation quality
pass
sample_006
50-test-cases-for-code-generation-qu
train
Evaluate case 6: 50 test cases for code generation quality
# AI Evaluation Benchmark: Storefront Code Generation (50 Test Cases)
edge_case
sample_007
50-test-cases-for-code-generation-qu
train
Evaluate case 7: 50 test cases for code generation quality
**Version:** 1.0 | **Classification:** Production-Grade Evaluation Corpus | **Target:** LLM Code Generation Quality Assessment
pass
sample_008
50-test-cases-for-code-generation-qu
train
Evaluate case 8: 50 test cases for code generation quality
## 1. Benchmark Architecture & Scoring Matrix (The Rubric)
pass
sample_009
50-test-cases-for-code-generation-qu
train
Evaluate case 9: 50 test cases for code generation quality
All 50 cases are evaluated against this **5-Dimension, 5-Point Likert Scale**. Annotators *must* provide line-level evidence for scores ≤ 3.
pass
sample_010
50-test-cases-for-code-generation-qu
train
Evaluate case 10: 50 test cases for code generation quality
| Dimension | Weight | 1 - Critical Failure | 2 - Major Defects | 3 - Acceptable w/ Issues | 4 - Production Ready | 5 - Exemplary / Gold Standard |
pass
sample_011
50-test-cases-for-code-generation-qu
train
Evaluate case 11: 50 test cases for code generation quality
| :--- | :---: | :--- | :--- | :--- | :--- | :--- |
edge_case
sample_012
50-test-cases-for-code-generation-qu
train
Evaluate case 12: 50 test cases for code generation quality
| **Correctness** | 30% | Logic fundamentally broken; wrong output for happy path. | Passes happy path; fails basic branch logic (e.g., null checks). | Correct logic; minor off-by-one or precision errors. | Handles all specified requirements & standard branches. | Formally verified logic; handles undefined behavior gra...
pass
sample_013
50-test-cases-for-code-generation-qu
train
Evaluate case 13: 50 test cases for code generation quality
| **Efficiency / Complexity** | 20% | O(N²) or worse where O(N) exists; N+1 queries; memory leaks. | Suboptimal DS choice (List vs Map); redundant computations. | Optimal Big-O; minor constant factor waste (e.g., double loop). | Optimal Big-O + Space; caching/memoization used correctly. | Algorithmic innovation (e.g., ...
pass
sample_014
50-test-cases-for-code-generation-qu
train
Evaluate case 14: 50 test cases for code generation quality
| **Code Style / Readability** | 15% | Unreadable; no types; magic numbers; >50 LOC functions. | Inconsistent naming; deep nesting (>3); mixed abstraction levels. | Clean naming; DRY; SOLID mostly followed; typed. | Idiomatic patterns; clear architecture boundaries; self-documenting. | Literate programming; domain-driv...
pass
sample_015
50-test-cases-for-code-generation-qu
train
Evaluate case 15: 50 test cases for code generation quality
| **Security & Robustness** | 20% | **Critical:** SQLi, XSS, RCE, Path Traversal, Hardcoded Secrets. | **High:** Missing authZ checks; weak crypto; verbose errors. | **Medium:** Input validation missing; race conditions possible. | **Low:** Parameterized queries; CSP; Rate limiting; Idempotency keys. | **Zero Trust:** ...
pass
sample_016
50-test-cases-for-code-generation-qu
train
Evaluate case 16: 50 test cases for code generation quality
| **Edge Case Coverage** | 15% | Crashes on empty input/null/zero. | Fails on boundary values (MaxInt, Leap Year, Empty String). | Handles documented edges; misses implicit ones (concurrency). | Handles implicit edges (idempotency, partial failure, clock skew). | **Chaos Ready:** Handles network partitions, Byzantine f...
edge_case
sample_017
50-test-cases-for-code-generation-qu
train
Evaluate case 17: 50 test cases for code generation quality
**Pass Threshold:** Weighted Score ≥ **3.8** | **Regression Threshold:** Any Dimension = **1** → Auto-Fail.
pass
sample_018
50-test-cases-for-code-generation-qu
train
Evaluate case 18: 50 test cases for code generation quality
## 2. The 50 Test Case Catalog (Storefront Domain)
pass
sample_019
50-test-cases-for-code-generation-qu
train
Evaluate case 19: 50 test cases for code generation quality
*Categorized by Domain & Difficulty. Full JSON Schema provided in Section 6.*
pass
sample_020
50-test-cases-for-code-generation-qu
train
Evaluate case 20: 50 test cases for code generation quality
| ID | Category | Difficulty | Prompt Summary | Key Evaluation Traps |
pass
sample_021
50-test-cases-for-code-generation-qu
train
Evaluate case 21: 50 test cases for code generation quality
| :--- | :--- | :--- | :--- | :--- |
edge_case
sample_022
50-test-cases-for-code-generation-qu
train
Evaluate case 22: 50 test cases for code generation quality
| **SF-001** | **Catalog/Read** | Easy | `getProductById(id: string): Product` | Null handling, SQLi via string concat, N+1 on variants. |
pass
sample_023
50-test-cases-for-code-generation-qu
train
Evaluate case 23: 50 test cases for code generation quality
| **SF-002** | **Catalog/Read** | Easy | `listProducts(filters, pagination): PaginatedResult` | Cursor vs Offset pagination, Index usage, Filter injection. |
pass
sample_024
50-test-cases-for-code-generation-qu
train
Evaluate case 24: 50 test cases for code generation quality
| **SF-003** | **Catalog/Write** | Medium | `createProduct(input): Product` | SKU uniqueness race condition, Transaction boundary, Media upload refs. |
pass
sample_025
50-test-cases-for-code-generation-qu
train
Evaluate case 25: 50 test cases for code generation quality
| **SF-004** | **Inventory** | Hard | `reserveStock(orderItems[]): ReservationResult` | **Distributed Locking**, Over-selling prevention, Idempotency Keys. |
pass
sample_026
50-test-cases-for-code-generation-qu
train
Evaluate case 26: 50 test cases for code generation quality
| **SF-005** | **Inventory** | Medium | `releaseReservation(reservationId): void` | Compensating transactions, Saga pattern, Partial release. |
edge_case
sample_027
50-test-cases-for-code-generation-qu
train
Evaluate case 27: 50 test cases for code generation quality
| **SF-006** | **Pricing** | Hard | `calculatePrice(cart, context): PriceBreakdown` | **Stacked Discounts**, Tax nexus, Precision (Banker's Rounding), Currency conversion. |
pass
sample_028
50-test-cases-for-code-generation-qu
train
Evaluate case 28: 50 test cases for code generation quality
| **SF-007** | **Pricing** | Medium | `applyPromoCode(code, cart): DiscountResult` | Coupon stacking rules, Expiry/Usage limits, Race on usage counter. |
pass
sample_029
50-test-cases-for-code-generation-qu
train
Evaluate case 29: 50 test cases for code generation quality
| **SF-008** | **Cart** | Easy | `addItem(cartId, item): Cart` | Session vs User cart merge, Variant validation, Quantity limits. |
pass
sample_030
50-test-cases-for-code-generation-qu
train
Evaluate case 30: 50 test cases for code generation quality
| **SF-009** | **Cart** | Medium | `mergeCarts(anonCart, userCart): Cart` | **Conflict Resolution** (qty sum vs latest), Promo re-evaluation. |
pass
sample_031
50-test-cases-for-code-generation-qu
train
Evaluate case 31: 50 test cases for code generation quality
| **SF-010** | **Checkout** | Hard | `placeOrder(cartId, paymentToken): Order` | **Saga Orchestration** (Payment -> Inventory -> Fulfillment), Idempotency, PCI DSS scope reduction. |
edge_case
sample_032
50-test-cases-for-code-generation-qu
train
Evaluate case 32: 50 test cases for code generation quality
| **SF-011** | **Checkout** | Medium | `handleWebhook(event): void` | **Idempotency Keys**, Signature verification, Event ordering/duplication. |
pass
sample_033
50-test-cases-for-code-generation-qu
train
Evaluate case 33: 50 test cases for code generation quality
| **SF-012** | **Payments** | Hard | `processRefund(orderId, amount, reason): Refund` | Partial refunds, Restocking logic, Async provider settlement, Precision. |
pass
sample_034
50-test-cases-for-code-generation-qu
train
Evaluate case 34: 50 test cases for code generation quality
| **SF-013** | **Payments** | Medium | `tokenizeCard(input): PaymentToken` | **No PAN in logs/memory**, Client-side encryption validation, PCI SAQ-A compliance. |
pass
sample_035
50-test-cases-for-code-generation-qu
train
Evaluate case 35: 50 test cases for code generation quality
| **SF-014** | **User/Auth** | Medium | `registerUser(email, pwd): User` | **Timing
pass

50 test cases for code generation quality

50 test cases for code generation quality

AI Evaluation Benchmark: Storefront Code Generation (50 Test Cases)

Version: 1.0 | Classification: Production-Grade Evaluation Corpus | Target: LLM Code Generation Quality Assessment

1. Benchmark Architecture & Scoring Matrix (The Rubric)

All 50 cases are evaluated against this 5-Dimension, 5-Point Likert Scale. Annotators must provide line-level evidence for scores ≤ 3.

Dimension Weight 1 - Critical Failure 2 - Major Defects 3 - Acceptable w/ Issues 4 - Production Ready 5 - Exemplary / Gold Standard
Correctness 30% Logic fundamentally broken; wrong output for happy path. Passes happy path; fails basic branch logic (e.g., null checks). Correct logic; minor off-by-one or precision errors. Handles all specified requirements & standard branches. Formally verified logic; handles undefined behavior gracefully.
Efficiency / Complexity 20% O(N²) or worse where O(N) exists; N+1 queries; memory leaks. Suboptimal DS choice (List vs Map); redundant computations. Optimal Big-O; minor constant factor waste (e.g., double loop). Optimal Big-O + Space; caching/memoization used correctly. Algorithmic innovation (e.g., streaming, probabilistic DS, SIMD).
Code Style / Readability 15% Unreadable; no types; magic numbers; >50 LOC functions. Inconsistent naming; deep nesting (>3); mixed abstraction levels. Clean naming; DRY; SOLID mostly followed; typed. Idiomatic patterns; clear architecture boundaries; self-documenting. Literate programming; domain-driven design; zero cognitive load.
Security & Robustness 20% Critical: SQLi, XSS, RCE, Path Traversal, Hardcoded Secrets. High: Missing authZ checks; weak crypto; verbose errors. Medium: Input validation missing; race conditions possible. Low: Parameterized queries; CSP; Rate limiting; Idempotency keys. Zero Trust: Formal verification hooks; Fuzzing targets; Capability-based security.
Edge Case Coverage 15% Crashes on empty input/null/zero. Fails on boundary values (MaxInt, Leap Year, Empty String). Handles documented edges; misses implicit ones (concurrency). Handles implicit edges (idempotency, partial failure, clock skew). Chaos Ready: Handles network partitions, Byzantine faults, schema drift.

Pass Threshold: Weighted Score ≥ 3.8 | Regression Threshold: Any Dimension = 1 → Auto-Fail.

2. The 50 Test Case Catalog (Storefront Domain)

Categorized by Domain & Difficulty. Full JSON Schema provided in Section 6.

ID Category Difficulty Prompt Summary Key Evaluation Traps
SF-001 Catalog/Read Easy getProductById(id: string): Product Null handling, SQLi via string concat, N+1 on variants.
SF-002 Catalog/Read Easy listProducts(filters, pagination): PaginatedResult Cursor vs Offset pagination, Index usage, Filter injection.
SF-003 Catalog/Write Medium createProduct(input): Product SKU uniqueness race condition, Transaction boundary, Media upload refs.
SF-004 Inventory Hard reserveStock(orderItems[]): ReservationResult Distributed Locking, Over-selling prevention, Idempotency Keys.
SF-005 Inventory Medium releaseReservation(reservationId): void Compensating transactions, Saga pattern, Partial release.
SF-006 Pricing Hard calculatePrice(cart, context): PriceBreakdown Stacked Discounts, Tax nexus, Precision (Banker's Rounding), Currency conversion.
SF-007 Pricing Medium applyPromoCode(code, cart): DiscountResult Coupon stacking rules, Expiry/Usage limits, Race on usage counter.
SF-008 Cart Easy addItem(cartId, item): Cart Session vs User cart merge, Variant validation, Quantity limits.
SF-009 Cart Medium mergeCarts(anonCart, userCart): Cart Conflict Resolution (qty sum vs latest), Promo re-evaluation.
SF-010 Checkout Hard placeOrder(cartId, paymentToken): Order Saga Orchestration (Payment -> Inventory -> Fulfillment), Idempotency, PCI DSS scope reduction.
SF-011 Checkout Medium handleWebhook(event): void Idempotency Keys, Signature verification, Event ordering/duplication.
SF-012 Payments Hard processRefund(orderId, amount, reason): Refund Partial refunds, Restocking logic, Async provider settlement, Precision.
SF-013 Payments Medium tokenizeCard(input): PaymentToken No PAN in logs/memory, Client-side encryption validation, PCI SAQ-A compliance.
SF-014 User/Auth Medium registerUser(email, pwd): User **Timing

Full Commercial Package & Checkout


License: MIT

Downloads last month
38