Áleran Vocals
Piano, entrenamiento con micrófono y juegos de oído para alumnos de canto: notas, escalas, afinación, rango vocal, resistencia e intervalos.
翻译准备中
尚未检测到更新。GitHub 发布版本后会显示在这里。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证TDS Portal Solver is a production-grade, highly secure, browser-based sandbox environment designed for the automated execution of deterministic solver logic. Engineered specifically as a local study companion for the IIT Madras Tools in Data Science (TDS) Course
翻译准备中
取自仓库 README
正在加载完整 README。
查看完整 README ↗TDS Portal is a browser-based study workspace for the IIT Madras Tools in Data Science (TDS) curriculum. It combines deterministic client-side solvers, interactive guides and quizzes, diagnostics, and integrations with separately hosted services where a question requires an API. Solver results and external-service availability depend on their respective inputs and services; the portal does not guarantee grades or uptime.
As of 2026-10-24, the catalog contains 534 registered question slots across 21 exams in three terms. The T2 2026 end-term practice bank includes 175 MCQs, 25 MSQs, and 100 subjective prompts. Practice material is reference-only and is not an official paper or score guarantee.
The workspace uses four client-side subsystems for deterministic practice, modular loading, and interactive feedback. These mechanisms improve usability; they do not provide security or grading guarantees:
graph TD
A["User Interface (Glassmorphism Dark IDE)"] -->|"Normalized Input (Email + Token)"| B("app.js (Core Orchestrator)")
B -->|"Active Checkbox Guard"| C{"Academic Integrity Guard"}
C -->|"Acknowledged"| D["Dynamic ESM Import"]
C -->|"Not acknowledged"| A
D -->|"Import Registry"| E["Term Registry Map (T1/T2/T3)"]
E -->|"Load Target Solver"| F["Shared runtime.js"]
F -->|"Study-mode check"| G{"lockConfig.locked?"}
G -->|"Open"| H("Seeded Variations Engine (FNV-1a / PRNG)")
G -->|"Guide-only"| I["Guide + Rubric Coach"]
H -->|"Deterministic Solved Vector"| J(("Deterministic Case Solution"))
J -->|"Native DOM Mount"| K["Rendered Answer + Notes Panel"]
I -->|"Native DOM Mount"| L["Interactive Rubric Coach Evaluator"]
K --> A
L --> A
style A fill:#121214,stroke:#374151,stroke-width:2px,color:#fff
style C fill:#451a03,stroke:#ea580c,stroke-width:2px,color:#fbbf24
style G fill:#2e1065,stroke:#8b5cf6,stroke-width:2px,color:#c084fc
style H fill:#064e3b,stroke:#059669,stroke-width:2px,color:#34d399
style J fill:#1e1b4b,stroke:#4f46e5,stroke-width:2px,color:#818cf8
style L fill:#831843,stroke:#db2777,stroke-width:2px,color:#f472b6
Reproducible variants use two core functions:
normalizeEmail(): Standardizes user-supplied emails by trimming trailing whitespace and casting all characters to lowercase.rng() / createRng(): Deterministic seeded pseudo-random generation using seedrandom.js and FNV-1a hashing. This is for reproducible study variants, not cryptography. Identical inputs and runtime assumptions should produce identical results; it is not an authentication or secrecy mechanism.For open-ended qualitative case studies (Project 2), the system features a dynamic linguistic generator:
${email}:${questionId}). Variation does not guarantee originality or prevent plagiarism detection.Rather than bloating runtime memory with pre-cached assets, the engine employs a dynamic lazy loading architecture:
interface SolverResponse {
id?: string; // Canonical question identifier
title: string; // Display title
answer: string; // The literal computed solution / raw text
type: "solved" | "guide" | "bypass" | "error";
variant: string; // Variant identification metadata
answerDisplay: string; // Rich formatted Markdown preview
guide?: string; // Implementation walkthroughs for CLI steps
rubricCoachHtml?: string; // Interactive Rubric Coach terminal component
debug?: Record<string, any>; // Computational trace logs & benchmark stats
}
Every case study workspace embeds a dedicated Rubric Coach & Draft Evaluator:
## Judgment, ## Evidence Table, etc.), character count gates ([150–3,000] or [200–6,000]), and evidence matrix columns (| Claim | Source | Confidence |).DF-00020, Farah Iqbal, INV-17, ~17.5%, 9021.10.00, SUP-02, 02:10:00, $123,565.26).The workspace presents an academic-integrity acknowledgement and study-oriented guidance. These client-side checks are educational safeguards, not tamper-proof enforcement:
[!WARNING] MALPRACTICE ALERT: This sandbox is strictly designed for local educational research, study, algorithm verification, and rubric calibration. Using this portal during active, live, proctored, or graded exams is a direct violation of university rules and constitutes academic dishonesty.
lock-config.js): Per-exam client-side switches can change the study experience, but allowlists ship in the public browser bundle and are not a security boundary. Never commit personal email addresses; use server-side authentication for genuinely restricted content.localStorage.Discourse scraper caches (discourse-cache/ and discourse-target-cache/) and the legacy core-roe.js bundle are intentionally excluded from Git. The ignored cache directories may remain on a maintainer's machine as offline fallbacks for the scraper scripts; they are not required by the deployed portal. Do not commit scraped personal data or generated bundles.
See Privacy, Third-Party Notices, Contributing, and Security before using or changing the project.
Browse the full Tools in Data Science (TDS) solver and exam-practice catalog across T1, T2, and T3 2026. This matrix is generated from tds-config.json and every exam's registry.js: counts, terms, capability summaries, and expandable question IDs/titles stay synchronized with the actual code. Run npm run build after changing a registry; npm run check fails if this README index is stale.
Live coverage: 3 terms · 21 exams · 540 question solver slots. Generated from tds-config.json and every registered registry.js; expand a row to browse its complete question IDs and titles. Explore: Open the live TDS Portal · Browse the machine-readable catalog · View the solver source.
The home page lists every term, exam and question the portal supports, so students and search engines can see the whole project without running a solver.
tds-config.json (terms, their exams, the term label such as "T3 2026 (Sep–Dec)") plus each exam's registry.js. Nothing about terms or years is hard-coded in the generator.npm run build derives portal exam counts and question titles from tds-config.json and each registry. It updates this README's coverage matrix and count badge, the marker blocks in index.html (CATALOG, SEO, TERMS), and catalog.json, sitemap.xml, robots.txt, and generated descriptions. npm run check fails if any generated artifact is stale.tds-config.json, the exam menu (TERM_EXAMS in app.js) and create solvers/<term>/<exam>/registry.js.main deploys to production, dev gets a preview).The user interface is crafted to mirror an elite dark-themed IDE workspace:
backdrop-filter: blur(16px), reactive focus glows matching the active color theme, and sub-pixel typography..txt, and Focus Mode toggles.📋 Paste), template loading (📑 Insert Skeleton), draft resetting (🧹 Clear), and live character/word counters with reactive boundary indicators (Green for valid gates, Red for violations).tds-roe-solver/
├── 📂 .agents/ # Agent behavioral profiles and configurations
├── 📂 solvers/
│ ├── 📂 T12026/ # Term 1 2026 Exam Registry (GA7, GA8, ROE, P2)
│ │ ├── 📂 ga7/ # Midpoint color generators & standard RNG hooks
│ │ ├── 📂 ga8/ # Docker verification & Cloud Run triggers
│ │ ├── 📂 roe/ # Maze pathfinding & regex golf solvers
│ │ └── 📂 p2/ # Solana devnet QR & Discourse facts indexer
│ ├── 📂 T22026/ # Term 2 2026 Exam Registry (GA0–GA8 + ROE + P1 + P2)
│ │ ├── 📂 ga0/ ... 📂 ga8/ # Modular deterministic solvers (GA0 through GA8)
│ │ ├── 📂 p1/ # Project 1 AI Agent & GCS setups
│ │ ├── 📂 p2/ # Project 2 Case Study Engine & Rubric Coach
│ │ │ ├── 📄 case-specs.js # Case boundaries, headings & rubric criteria
│ │ │ ├── 📄 rubric-coach.js# Rubric Coach terminal & draft evaluator
│ │ │ ├── 📄 lock-config.js # Access control configuration & whitelist
│ │ │ ├── 📄 runtime.js # Solver execution wrapper & diagnostics
│ │ │ ├── 📄 variations-engine.js # PRNG variation & table generator
│ │ │ ├── 📄 registry.js # Ordered question registry (1-to-1 live exam match)
│ │ │ ├── 📄 q-case-1a.js # Case 1A: DTH Month-End Mystery
│ │ │ ├── 📄 q-case-1b.js # Case 1B: DTH Complaints Went Quiet
│ │ │ ├── 📄 q-case-2a.js # Case 2A: Solar Inverter Smell Test
│ │ │ ├── 📄 q-case-2b.js # Case 2B: Solar 31.6% Impact Claim
│ │ │ ├── 📄 q-case-3a.js # Case 3A: Swiss Mismatch Control
│ │ │ ├── 📄 q-case-4a.js # Case 4A: QC Queue Smell Test
│ │ │ ├── 📄 q-case-3b.js # Case 3B: Irish Preference Claim
│ │ │ └── 📄 q-case-4b.js # Case 4B: Spare-Parts Search
│ ├── 📂 T32026/ # Term 3 (GA1, GA2, GA3 Agent Safety & Trading Games, Project 1)
│ ├── 📂 ga1/ # GA1 16 official slots (Tangram SVG, Git forensics, Ledger)
│ ├── 📂 ga2/ # GA2 15 official slots (CORS, OAuth, Observability, Ollama)
│ ├── 📂 ga3/ # GA3 8 official slots (agent endpoints, face tools, game guides)
│ └── 📂 p1/ # Project 1 7 official slots (forecast, games, receipts)
│ └── 📂 shared/ # Cross-term helpers, including timeout cleanup
├── 📂 .github/ # CI, security guidance, and dependency automation
├── 📂 scripts/
│ └── 📄 build-catalog.mjs # Generates the README matrix, portal index, and SEO files
├── 📄 app.js # Central UI state machine & event orchestrator
├── 📄 check.mjs # Comprehensive multi-term test suite
├── 📄 CONTRIBUTING.md / SECURITY.md / CODE_OF_CONDUCT.md
├── 📄 index.html # Main portal shell (contains generated CATALOG/SEO/TERMS blocks)
├── 📄 catalog.json # Generated machine-readable solver index
├── 📄 sitemap.xml / robots.txt # Generated SEO files
├── 📄 tds-config.json # Terms, exams and intro text (source of truth for the home page)
├── 📄 vercel.json # Strict build command, cache policy, and baseline security headers
├── 📄 server.js # Traversal-safe local static server for development and checks
├── 📄 style.css # Master stylesheets & Rubric Coach terminal styles
├── 📄 sw.js # Progressive Web App offline Service Worker
├── 📄 verify.html # Solver verification dashboard
├── 📄 .editorconfig / .nvmrc # Editor defaults and recommended Node.js version
├── 📄 PRIVACY.md # Local storage, analytics, and external-request summary
└── 📄 THIRD_PARTY_NOTICES.md # Third-party content and attribution guidance
Use Node.js 22 or newer (Node.js 24 LTS is the recommended local version). Install from the lockfile, then start the local ESM server:
npm ci
npm run dev
🌐 Local Endpoint: http://localhost:3000/
With port 3000 available, run the registry, solver-output, mock-bank, catalog, and static-server checks:
npm run check
The same command is available as npm test. Validate the mock bank with npm run mock:check; npm run mock:generate intentionally rewrites generated mock artifacts.
The official May 2026 course index is mapped into a reproducible practice bank with exactly 175 MCQs, 25 MSQs, and 100 subjective prompts:
npm run mock:check
npm run mock:generate
Generated artifacts are written to mock-banks/T2-2026-End-Term-Mock.md, mock-banks/T2-2026-End-Term-Mock.json, and the browser data module. Select T2 2026 → End-Term Mock in the portal to use the interactive quiz. Use Jump to portion to practice MCQ, MSQ, or subjective questions separately. Progress, drafts, bookmarks, and rubric checks are saved locally per email.
Reference-only notice: this mock is original practice material, not an official IIT Madras paper, answer key, or prediction. It cannot guarantee a high end-term score and is intended only for revision and last-minute practice.
main triggers the Vercel production deployment; other branches create previews.tds-portal-v* caches and only removes caches owned by this app.127.0.0.1:3000 during npm test; the port must be free. CI runs the same checks on Node.js 22 and 24.