ba-auto-daily/graphify-out/GRAPH_REPORT.md
Nik Afiq 0ee1d495c8 Add manifest.json and implementation plan for ba-auto-daily script
- Created manifest.json to track file metadata for the project.
- Added plan.md detailing the implementation strategy for the ba-auto-daily automation script, including feature prioritization and prerequisites.
2026-07-05 05:29:30 +09:00

12 KiB

Graph Report - . (2026-07-05)

Corpus Check

  • Large corpus: 21 files · ~974,160 words. Semantic extraction will be expensive (many Claude tokens). Consider running on a subfolder.

Summary

  • 159 nodes · 176 edges · 23 communities (16 shown, 7 thin omitted)
  • Extraction: 90% EXTRACTED · 10% INFERRED · 1% AMBIGUOUS · INFERRED: 17 edges (avg confidence: 0.8)
  • Token cost: 320,012 input · 0 output

Community Hubs (Navigation)

God Nodes (most connected - your core abstractions)

  1. do_cafe() - 17 edges
  2. detect_and_click.py (color-masked template-match sparkle detector/clicker) - 12 edges
  3. ba-auto-daily Implementation Plan - 12 edges
  4. Cafe Room 2 Reference Screenshot - 11 edges
  5. Event content (Activities) sweep feature - 8 edges
  6. do_mailbox() - 7 edges
  7. ba_dailies.sh (entry-point automation script) - 6 edges
  8. Cafe Income Popup (カフェの収入) - 6 edges
  9. Cafe Room 1 Reference Screenshot - 6 edges
  10. Blue Archive Main/Home Screen Screenshot - 6 edges

Surprising Connections (you probably didn't know these)

  • Lesson/Schedule (affection farming via classes) feature --semantically_similar_to--> do_cafe() [INFERRED] [semantically similar] plan.md → ba_dailies.sh
  • Cafe Reward Menu Popup Screenshot --rationale_for--> do_cafe() [INFERRED] screenshots/cafe/cafe_reward/02_cafe_reward_menu_popup.png → ba_dailies.sh
  • Receive Button (受取 / Enter) --rationale_for--> do_cafe() [INFERRED] screenshots/cafe/cafe_reward/02_cafe_reward_menu_popup.png → ba_dailies.sh
  • Cafe Button (Bottom Navigation) --references--> do_cafe() [INFERRED] screenshots/main/main.png → ba_dailies.sh
  • Blue Archive Main/Home Screen Screenshot --references--> focus_game() [INFERRED] screenshots/main/main.png → ba_dailies.sh

Import Cycles

  • None detected.

Hyperedges (group relationships)

  • Tier 1 — high-priority, no-OCR-needed features — plan_stamina_ap_sweep, plan_club_group_ap_claim, plan_momo_talk [EXTRACTED 1.00]
  • Tier 2 — sweep-style AP consumption features sharing the buy-tickets/sweep-N-times/color-detect shape — plan_commissions, plan_scrimmage, plan_bounty, plan_story_ap_sweep [EXTRACTED 1.00]
  • Cafe sparkle-click automation pipeline (focus -> loop-detect-click -> claim) — ba_dailies_script, ba_dailies_focus_game, ba_dailies_do_cafe, scripts_detect_and_click_script, assets_cafe_sparkle_image [EXTRACTED 1.00]

Communities (23 total, 7 thin omitted)

Community 0 - "Detection Pipeline & Architecture"

Cohesion: 0.10 Nodes (26): cafe_sparkle.png (template image for sparkle detection), cafe_sparkle.png Template Image, scp/rsync + ssh deployment of scripts/assets to nik-gpu, Single local screenshot->detect->click pipeline on nik-gpu (avoids SSH round-trip for the fast-moving sparkle), nik-gpu (Linux box running the Blue Archive client and X display), nik-macbookair (macOS development machine), numpy dependency, opencv-python dependency (+18 more)

Community 1 - "ba_dailies.sh Script & Cafe Claim"

Cohesion: 0.18 Nodes (17): CAFE_MAX_CLICKS_PER_ROOM constant, click(), DISPLAY, do_cafe(), do_cafe_room(), do_mailbox(), focus_game(), press_enter() (+9 more)

Community 2 - "Combat/Story Feature Roadmap"

Cohesion: 0.12 Nodes (19): baas module/arena.py (173 lines), baas module/explore_tasks/explore_task.py, baas module/main_story.py (auto_fight/enter_battle), baas module/momo_talk.py (183 lines, getConversationState), Known gap: unconfirmed rotated-student popup / zoom-center / rank-up popup handling in do_cafe, Arena (PvP ladder) feature, Auto-fight primitive (shared start-fight/max-speed/auto-mode/detect-result routine), baas reference project (pur1fying/blue_archive_auto_script, GPL-3.0, read-only research reference) (+11 more)

Community 3 - "AP/Task Sweep Features"

Cohesion: 0.13 Nodes (15): baas module/activities/activity_utils.py, baas module/clear_special_task_power.py (148 lines), baas module/collect_daily_free_power.py, baas module/collect_daily_task_power.py, baas module/explore_tasks/sweep_task.py, baas module/explore_tasks/task_utils.py, baas module/rewarded_task.py (204 lines), baas module/scrimmage.py (166 lines) (+7 more)

Community 4 - "Cafe Room 2 UI & Detector"

Cohesion: 0.13 Nodes (11): Green Swirl Markers Above Two Students (possible status/affection indicator), Bottom Edit Toolbar (編集モード / 贈り物 / 家具情報 / プリセット / すべて収納), Comfort Level Indicator (快適度 5500), Top HUD: AP / Gold / Gems Counters, Cafe Furnishings Layout (arcade cabinets, crane machine, sofas, plants), Cafe Income Rate Display (カフェの収入 4.1%), Invitation Ticket Purchase Prompt with Countdown Timer, "Move to Store 1" Button (1号店に移動) (+3 more)

Community 5 - "Cafe Income Popup UI"

Cohesion: 0.29 Nodes (7): Cafe Reward Menu Popup Screenshot, Cafe Income Popup (カフェの収入), Popup Close Button (X), Receive Button (受取 / Enter), Store 1 AP Income Panel (1号店 AP), Store 1 Credits Income Panel (1号店 Credits), Store 2 Credits Income Panel (2号店 Credits)

Community 6 - "Cafe Room 1 UI"

Cohesion: 0.29 Nodes (7): Cafe Edit-Mode Bottom Toolbar (edit/gift/furniture-info/preset/store-all), Cafe Income (カフェの収入) and Comfort Level (快適度) Display, Cafe Room 1 Reference Screenshot, "Move to Shop 2" (2号店に移動) Navigation Button, Cafe Room 1 Furniture Layout (train set, animal pen, swing, car, music stand), Student Characters Scattered in Cafe Room 1 (potential sparkle click targets), Designated Visit / Random Visit Icons (指定訪問・ランダム訪問)

Community 7 - "Cafe Income Reference Capture"

Cohesion: 0.33 Nodes (6): Cafe Income Percentage Display (カフェの収入 8.3%), Cafe Room Interior Scene (furniture, students, train set), Comfort Level Meter (快適度 5500), Invitation Ticket Buttons (招待券 / 追加招待券) with Countdown Timer, Cafe Top Navigation Bar (AP / Gold / Gems, Back Button, Visit Buttons), Cafe Reward Screen Screenshot (Cursor on Top)

Community 8 - "Game Main Screen UI"

Cohesion: 0.33 Nodes (6): Bottom Navigation Bar (Cafe, Schedule, Students, Formation, Social, Manufacture, Shop, Recruit, Work), Cafe Button (Bottom Navigation), Currency/Resource Display (AP, Credits, Gems), Mailbox/Envelope Icon (Top Bar), Blue Archive Main/Home Screen Screenshot, nik-gpu Window Title Bar

Community 9 - "Student Rotation Popup"

Cohesion: 0.40 Nodes (5): Popup Close (X) Button, top-right corner, Student Rotation Popup Screenshot, OK Button (Enter-bound) below student list, Visited Student List Popup ("案内"), 訪問した生徒リスト (Visited Student List row of 5 avatars)

Community 10 - "Shop Automation (Planned)"

Cohesion: 0.50 Nodes (4): baas module/shop/common_shop.py, baas module/shop/shop_utils.py, baas module/shop/tactical_challenge_shop.py, Common Shop + Tactical (Arena) Shop feature

Community 11 - "Sparkle Zoomed-Out Calibration"

Cohesion: 0.50 Nodes (4): Cafe HUD (currency, comfort/kaitekido 5500, income 0.0%, timer 17:41:20), Cafe Room Interior (Zoomed Out, Full Layout), Yellow/White Affection Sparkle (visible near plush character), Cafe Zoomed-Out Calibration Screenshot

Community 12 - "Student Rank-Up Popup Stage 1"

Cohesion: 0.67 Nodes (4): Bond Rank Heart Indicator (Rank 4), Conditional Rank-Up Popup Close (Automation Gap), Student Bond Rank-Up Popup Screenshot, "Attack +12" Stat Boost Banner

Community 13 - "Sparkle Template Asset"

Cohesion: 1.00 Nodes (3): Cafe Affection Sparkle UI Element, Cafe Sparkle Template Image, Yellow/White Sparkle Color Signature

Community 14 - "Group/Mini Story (Planned)"

Cohesion: 0.67 Nodes (3): baas module/group_story.py, baas module/mini_story.py, Group Story / Mini Story cleanup feature

Community 15 - "Student Rank-Up Popup Stage 2"

Cohesion: 1.00 Nodes (3): Bond Rank-Up Popup UI (絆ランクアップ! banner over portrait, stat-gain footer), Depicted Student (white/silver-haired, halo motif, off-shoulder outfit, Max HP +144), Rank-Up Popup Screenshot (Student B, Bond Rank 4, Max HP +144)

Ambiguous Edges - Review These

  • Cafe Room 2 Reference ScreenshotGreen Swirl Markers Above Two Students (possible status/affection indicator) [AMBIGUOUS] screenshots/cafe/room2.png · relation: conceptually_related_to

Knowledge Gaps

  • 78 isolated node(s): DISPLAY, XAUTHORITY, nik-macbookair (macOS development machine), opencv-python dependency, numpy dependency (+73 more) These have ≤1 connection - possible missing edges or undocumented components.
  • 7 thin communities (<3 nodes) omitted from report — run graphify query to explore isolated nodes.

Suggested Questions

Questions this graph is uniquely positioned to answer:

  • What is the exact relationship between Cafe Room 2 Reference Screenshot and Green Swirl Markers Above Two Students (possible status/affection indicator)? Edge tagged AMBIGUOUS (relation: conceptually_related_to) - confidence is low.
  • Why does do_cafe() connect ba_dailies.sh Script & Cafe Claim to Detection Pipeline & Architecture, Game Main Screen UI, Combat/Story Feature Roadmap, Cafe Income Popup UI? High betweenness centrality (0.216) - this node is a cross-community bridge.
  • Why does ba-auto-daily Implementation Plan connect Combat/Story Feature Roadmap to ba_dailies.sh Script & Cafe Claim? High betweenness centrality (0.148) - this node is a cross-community bridge.
  • Why does detect_and_click.py (color-masked template-match sparkle detector/clicker) connect Detection Pipeline & Architecture to ba_dailies.sh Script & Cafe Claim? High betweenness centrality (0.114) - this node is a cross-community bridge.
  • Are the 5 inferred relationships involving do_cafe() (e.g. with Lesson/Schedule (affection farming via classes) feature and Cafe Reward Menu Popup Screenshot) actually correct? do_cafe() has 5 INFERRED edges - model-reasoned connections that need verification.
  • What connects DISPLAY, XAUTHORITY, nik-macbookair (macOS development machine) to the rest of the system? 86 weakly-connected nodes found - possible documentation gaps or missing edges.
  • Should Detection Pipeline & Architecture be split into smaller, more focused modules? Cohesion score 0.10461538461538461 - nodes in this community are weakly interconnected.