← Về thư mục
📄 / / projects / personal / Vedic-astro-mar / PROJECT_STATUS.md

Project Status: Vedic Astro Guide

📅 Session Log: 2026-04-02 (Memory Integration & Restructuring Prep)

Focus: Integrating Lob Brain with 4-Doc Rule and planning workspace restructuring.

✅ Major Updates

  1. Lob Brain Integration: Fixed lob-brain server (TS errors, transport layer) and automated its synchronization via v-sync-logger.
  2. Unified Memory Guide: Created docs/BRAIN_GUIDE.md to define the "Dual-Layer Memory" strategy (Human Truth vs AI Subconscious).
  3. Workspace Restructuring: Completed the move to a 2-layer folder hierarchy (.agents for shared logic, Projects for independent silos).

🔜 Next Steps

  1. Context Retrieval Tests: Validate brain_query and brain_context performance with real Vedic data.
  2. SAV Integration: Resume Sarvastakavarga UI work in Step0Input.tsx.

📅 Session Log: 2026-04-01 (SAV Data Integration & Phase 1 Finalization)

Focus: Implementing Sarvastakavarga (SAV) data capture and finalizing Step 0/0.5 fidelity.

✅ Major Updates

  1. Maintenance Execution: Successfully pruned legacy brain artifacts (pre-Mar 24) to optimize storage.
  2. SAV Prompt Design: Created docs/SAV_PROMPT.md for standardized vision-to-markdown extraction.
  3. Session Start: Resumed "Knowledge-Driven Interpreter" cycle with a focus on high-fidelity data inputs.

🔜 Next Steps

  1. SAV UI Integration: Add the SAV scoring grid to Step0Input.tsx to handle user-provided points.
  2. Phase 2 Implementation: Connect the SAV data to the Interpretation GEM (Step 1 analysis).
  3. Automated Verification: Run build and sync tests.

📅 Session Log: 2026-03-30 (Phase 1 Refinement & Monday Maintenance)

Focus: Finalizing Phase 1 with SAV (Sarvastakavarga) data capture and executing the Monday Morning Protocol.

✅ Major Updates

  1. Monday Maintenance Audit: Storage audit in ~/.gemini/antigravity initiated for legacy artifact pruning.
  2. Phase 1 Extension: Acknowledged new focus on SAV Tables using photo-to-markdown extraction workflow.

🔜 Next Steps

  1. Maintenance Execution: Delete confirmed legacy folders from March 20.
  2. SAV Extraction Workflow: Design a Vision-to-Table prompt for extracting Sarvastakavarga charts from user-provided photos.
  3. SAV UI Integration: Implement the storage and retrieval system for SAV tables in Step0Input.tsx.

📅 Session Log: 2026-03-29 (Table-First Sync & Phase 1 Completion)

Focus: Finalizing the Data Assembler with automated profile synchronization and Table-First architecture.

✅ Major Updates

  1. sync-user Automation: Created scripts/sync-user.ts to automatically populate Supabase profiles from JHora Markdown exports.
  2. Table-First Architecture: Refactored the data layer to prioritize Markdown tables over legacy ASCII grids, saving tokens and improving readability for NotebookLM.
  3. App-wide Sync Support: Updated Step0Input.tsx and importService.ts to seamlessly load charts and technical tables from the centralized database.
  4. Global Skill Bridge: Formalized the workspace-to-global connection in .agents/rules.md.
  5. v-sync-logger & Full 4-Doc Rule: Implemented automated synchronization for CHANGELOG, PROJECT_STATUS, APP_CONTEXT, and README.
  6. Operating Procedures: Codified "Save & Refresh" lifecycle guidance to SKILL_BRIDGE.md for maximum token efficiency.

🔜 Next Steps

  1. Phase 2 Implementation: Begin the interpretation layer build-out using the validated table-driven context.
  2. Skill Refinement: Audit additional local skills for "Pulse" compatibility.

📅 Session Log: 2026-03-28 (Token Optimization & Indexing Strategy)

Focus: Creating a scalable Logic Index to reduce AI context token usage.

✅ Major Updates

  1. The Indexing Strategy: Transitioned APP_CONTEXT.md into a "Logic Index Table". Replaced the idea of condensing the 36KB Original.md (which risked logic drift) with precise header-routing to preserve the Single Source of Truth.
  2. UI Logic Decoupling: Extracted UI behavior from the old workflow doc into a new tabular docs/UI_WORKFLOW.md.
  3. Model Stack Realignment: Directed technical stack to utilize Google AI Studio models, strictly prioritizing Flash/Lite/Gemma parameters to save tokens (using task-splitting). Pro models are excluded unless absolutely unavoidable.
  4. Agent Rules Updated: Modified @VedicGuru and @AG MDC rules to enforce reading the index rather than reading Original.md blindly.

🔜 Next Steps

  1. Dynamic Calibration: Implement a "Sync with Master" button in the dev tools.
  2. Phase 2 Execution: Begin implementing the specialized GEM interpretation layers using the validated dataset.

📅 Session Log: 2026-03-28 (Knowledge-Driven Refactoring)

Focus: Structural consolidation and "Single Source of Truth" implementation.

✅ Major Updates

  1. Source of Truth Mapping: Established docs/Vedic - Nhi data full - 19h43.md as the master data record and Vedic- How to_Original.md as the logic blueprint.
  2. Data Layer Synchronization: Updated src/constants.ts to match the 19:43:50 birth time snapshot (Longitudes, Shadbala, and Dashas).
  3. Architecture Modularization: Moved astrology-specific skills to .agents/skills/ to align with the AG framework standards.
  4. Knowledge Pivot: Shifted project focus from "Data Assembler" to "Knowledge-Driven Interpreter," prioritizing the use of curated artifacts over fragmented input.

🔜 Next Steps

  1. Dynamic Calibration: Implement a "Sync with Master" button in the dev tools to automatically refresh constants from the Markdown master if it changes.
  2. Phase 2 Execution: Begin implementing the specialized GEM interpretation layers using the validated 19:43 dataset.

📅 Session Log: 2026-03-27 (Vedic Data Interface Refinement)

Focus: High-fidelity space optimization and user-centric data capture.

✅ Major Updates

  1. Birth Time Precision: Fixed synchronization issues in VedicDateTimePicker ensuring birth time initializes correctly (19:44:50) from props.
  2. Premium Varga UX:
  3. Replaced bulky Varga grid with an "Add Varga" flyout menu.
  4. Added global "All/None" controls with dummy data auto-filling for rapid testing.
  5. Integrated custom hover tooltips for astrological chart education.
  6. Transit Synthesis Redesign:
  7. Condensed Section 4 into a sidebar-weighted layout (fixed Date/Time column + flexible ASCII chart area).
  8. Standardized "Natal Ref" selection UI to match the Varga component.
  9. Enhanced Data Export:
  10. Centralized "Copy" buttons in section headers for cleaner context aggregation.
  11. Verified 5-column Markdown table generation for transit analysis.

🔜 Next Steps

  1. Dasha Synthesis: Connect the dasha list to real calculation logic (Phase 1).
  2. Interpretation Flow: Begin testing the "Master Context" with the high-fidelity interpreted GEMs.

📅 Session Log: 2026-03-25 (Workflow & Complexity Governance)

Focus: Structuring planning protocols and modularizing domain rules.

✅ Major Updates

  1. Unified Global Flow: Created the /global-flow workflow to handle the entire meta-lifecycle (Research, Rules, Skills) with a focus on Gemini 2.5 Flash and a trilateral analysis model (Inherit vs Missing vs Rejected).
  2. Knowledge Pivot: Decided to use the Official Gemini Grounding API as the professional bridge to replicate NotebookLM's "Truth of Search Knowledge" in Phase 2.

📅 Session Log: 2026-03-25 (Global Maintenance System)

Focus: Storage optimization and autonomous maintenance rituals.

✅ Major Updates

  1. Maintenance Ritual: Implemented a mandatory start-of-session storage scan for the @AG Lead Coordinator.
  2. Maintenance Workflow: Created the /maintenance command for structured artifact cleanups.
  3. Legacy Cleanup: Successfully deleted the folder 286342f8-ae9d-4f81-ac6f-0bb7e637d39d (Created March 17) to free up storage.
  4. Governance Mirroring: Updated all primary documents to reflect the new retention policies.

📅 Session Log: 2026-03-25 (Phase 0 & @AG Alignment)

🔜 Next Steps

  1. Step 0 Technical Audit: Perform a deep audit of the asciiParser.ts for accuracy across varied JHora output formats (delegated to @VedicGuru).
  2. UI Polish: Ensure the "Copy Master Context" button is prominent and foolproof (delegated to @AstroDev).

📅 Session Log: 2026-03-23 (5-Screen AI Workflow Implementation)

📅 Session Log: 2026-03-21 (Architecture & Step 0.5 Refinement)

Focus: Granular Technical Roadmaps and GEM Architecture.

✅ Major Updates

  1. Technical Roadmap Refactor (Step 0.5):
  2. Split the one-click distillation into 4 distinct sections (Step 1, 2, 3, 4).
  3. Each section now has a specialized "Copy Interpretation Prompt" button.
  4. These buttons automatically inject the relevant Vedic Guidance from Vedic- How to_Original.md into the prompt.
  5. Guidance Integration:
  6. Extracted core rules from the original markdown and added them to src/services/astroWorkflow.ts as VEDIC_GUIDANCE.
  7. Testing Documentation:
  8. Updated README.md with explicit localhost:3000 instructions.
  9. Architecture Strategy:
  10. Agreed on a Hybrid 2-GEM Strategy:
    • Vedic Astrology Interpretation by Nhi Vo: Vedic Astrology Interpretation by Nhi Vo (Steps 1, 2, 3).
    • GEM 2: Event Timing & Trigger Analyst (Step 4).

🔜 Next Steps (Future Session)

  1. Step 4 Context Injection: Implement the "Foundation Context" bundle for the Step 4 copy prompt.
  2. Workflow Discussion: Discuss the end-to-end workflow from App to GEM.

📅 Session Log: 2026-03-20 (Refinement v1.3.0)

Focus: Quota-Savvy Technical Distillation.

✅ Major Updates

  1. Technical Roadmap (Step 0.5):
  2. One-click distillation of ALL astrological data into a single technical audit.
  3. Saves 4+ API requests per session.
  4. Step-Based Formatting:
  5. Extraction prompt now explicitly organizes data by Step 1, 2.1-2.4, 3, and 4.
  6. Model Flex:
  7. Prepared VITE_GEMINI_MODEL for easy switching to high-volume models (Gemma 3).

📅 Session Log: 2026-03-19 (UI Stability)

Focus: Error handling and rate limit resilience.

✅ Major Updates

  1. API Fallback: Initial implementation of Gemini fallback logic.
  2. UI Polish: Framer Motion transitions for step navigation.

🛠️ Global Development & Testing