feat:Added goal-oriented learning with constraints and high-frequency key point extraction.

[SCW5370]

Project Positioning

This submission is based on the upstream DeepTutor dev branch and delivers a final derivative customization focused on student users and exam-prep scenarios.
The goal is to preserve upstream core architecture compatibility while strengthening learning-loop execution, exam efficiency, and product consistency for education workflows.

Final-State Change Overview (Baseline: upstream/dev)

This PR is submitted in a “final-state direct alignment” form, representing the current complete state of the derivative edition.

1) Student & Exam-Prep Capability Enhancements

• Added a Goal planning workflow with day-level execution (plan-execute-feedback-replan).
• Added Exam Radar related paths for extracting high-frequency knowledge points/question patterns from exam materials and supporting prioritization.
• Improved session context continuity for smoother learning flow transitions.

2) Product UX and Language Consistency Improvements

• Cleaned and refined key learner-facing pages and guidance copy for consistency and readability.
• Unified TutorBot-related naming and presentation for a more consistent user journey.
• Updated default templates and prompts to better match student/exam-prep usage context.

3) Engineering and Structural Synchronization

• Synchronized module-level additions/removals and route adjustments in the derivative edition (including goal and exam-radar related modules).
• Added corresponding tests and submission documentation package for review and regression reference.

Validation & Notes

• This is a final-state alignment PR, covering the full derivative change surface.
• Basic usability checks and key build validation were completed against the local final edition.
• No private secret material was introduced by this submission (no private key-type content included).

Suggested Review Order

1. Entry points and routes (goal / exam-radar / tutorbot related pages)
2. Learning workflow (planning, feedback, replanning, session continuity)
3. Language and copy consistency
4. Tests and submission documents

Derivative Intent

This edition is positioned as an education-focused derivative distribution of DeepTutor:

• Daily student learning
• Exam-prep score-improvement practice
• Practical learning flow with executable plans and feedback loops

[SCW5370]

Thanks for reviewing. To make evaluation easier, here is a focused functional breakdown:

1. Student / Exam-Prep Workflow Enhancements (Primary Scope)

• Added a new Goal learning-path module with multi-page flow:
  • Entry page: define learning objective, time budget, target level, and strategy
  • Day plan page: day-level task cards, time allocation, and learn/practice/review steps
  • Interactive day page: follow-up Q&A grounded in the current day plan
• Added a new Exam Radar page:
  • Supports pasting exam text and uploading exam materials for structured analysis
  • Produces high-frequency knowledge points and question-type signals for exam prioritization
• Added feedback-driven replanning behavior:
  • Users can submit task feedback (done/partial/skipped, etc.)
  • The system adjusts upcoming plans based on real execution feedback
• Established a full plan-execute-feedback-replan loop:
  • From goal generation to day-level execution to feedback-based adaptation

What this enables in practice:

• Converts vague learning goals into executable day-level plans
• Converts exam materials into actionable high-frequency topic/question priorities
• Keeps the plan adaptive to real progress, rather than static one-time planning

2. Localization and UX Consistency

• Refined key learner-facing copy and guidance text
• Reduced mixed-language output in core user flows
• Improved reading and interaction consistency for student exam-prep use

Quick validation path:

• Open Goal page and create a learning session
• Enter Day detail page and verify task decomposition and execution steps
• Open interactive day page and test follow-up Q&A
• Open Exam Radar, upload/paste materials, and verify high-frequency analysis output
• Submit feedback on Goal board and trigger replan to verify adaptive plan updates

Scope note:

• This PR is submitted as a derivative final-state alignment against upstream/dev
• Core architecture compatibility is preserved; no structural rewrite of upstream core runtime

[SCW5370]

I would like to apologize for reaching out again. In my recent work on a study path planning feature, I attempted to implement an interactive learning content generator inspired by your guided learning approach. However, I am aware that my implementation is still far from the level of professionalism and structure demonstrated in your project.

I truly admire the quality and rigor of the interactive learning content your project produces. At this stage, I am still learning and improving, and I would greatly appreciate any opportunity to contribute, learn, and help refine this aspect of the system.

[pancacake]

Thanks for your contribution! That is a HUGE pr though lol. The design insights behind this pr is highly valued, and I'll try to do something great base on your pr. Thanks

[pancacake]

我这边看您是中文开发者，我就用中文接着回复了哈，您这个pr我仔细看过了，可能没有办法给您直接合并，因为里面有一些逻辑和内容看起来对系统的整体结构破坏性有一点点大。不过这个目标导向的交互式学习是很棒的一个想法，我觉得他是可以进一步和tutorbot相结合，来做一些事情的。后面我也会看看怎么去具体的实现这些内容。感谢您的贡献～

[SCW5370]

感谢您的回复，您的项目很有学习价值希望看到它越来越好 发自我的iPhone

…

------------------ Original ------------------ From: Bingxi Zhao ***@***.***> Date: Sun,Apr 12,2026 9:42 PM To: HKUDS/DeepTutor ***@***.***> Cc: 向阳 ***@***.***>, Author ***@***.***> Subject: Re: [HKUDS/DeepTutor] feat:Added goal-oriented learning with constraints and high-frequency key point extraction. (PR #298) pancacake left a comment (HKUDS/DeepTutor#298) 我这边看您是中文开发者，我就用中文接着回复了哈，您这个pr我仔细看过了，可能没有办法给您直接合并，因为里面有一些逻辑和内容看起来对系统的整体结构破坏性有一点点大。不过这个目标导向的交互式学习是很棒的一个想法，我觉得他是可以进一步和tutorbot相结合，来做一些事情的。后面我也会看看怎么去具体的实现这些内容。感谢您的贡献～ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>