fix(translate_only): ensure MD and HTML outputs use consistent directories

Problem:
- MD output was saved to input file's parent directory
- HTML output used final_dirs.html from config
- This caused MD/HTML files to be out of sync when input came from
  version history (outputs/versions/v1/)

Solution:
- Both outputs now use final_dirs from config
- MD → final_dirs.markdown (docs/mindmaps/)
- HTML → final_dirs.html (docs/pages/mindmaps/)

Impact:
- translate_only.py --html now produces synchronized outputs
- Consistent with main pipeline behavior
- Files are always in expected final directories

Author: lufftw

tools/ai-markmap-agent/outputs/versions/v1/neetcode_ontology_agent_evolved_en.html

@@ -2,42 +2,97 @@
 
 Translate the following Markmap content to **Traditional Chinese (Taiwan)**.
 
-## CRITICAL: Use Taiwan's Algorithm & Data Structure Terminology
-
-### ⚠️ Taiwan vs Mainland China Terminology (MUST use Taiwan terms)
-
-The following terms differ between Taiwan (台灣) and Mainland China (中國大陸). 
-**You MUST use the Taiwan column. NEVER use Mainland China terms.**
-
-| English | 台灣 (USE THIS) | 中國大陸 (NEVER USE) |
-|---------|-----------------|---------------------|
-| Pointer | 指標 | ~~指針~~ |
-| Two Pointers | 雙指標 | ~~雙指針~~ |
-| Fast-Slow Pointers | 快慢指標 | ~~快慢指針~~ |
-| In-place | 原地 | ~~就地~~ |
-| Enumerate | 列舉 | ~~枚舉~~ |
-| Boolean | 布林 / Boolean | ~~布爾~~ |
-| Function | 函式 | ~~函數~~ |
-| Variable | 變數 | ~~變量~~ |
-| Parameter | 參數 | ~~參數~~ (same) |
-| Memory | 記憶體 | ~~內存~~ |
-| Program | 程式 | ~~程序~~ |
-| Object | 物件 | ~~對象~~ |
-| Interface | 介面 | ~~接口~~ |
-| Implementation | 實作 | ~~實現~~ |
-| Information | 資訊 | ~~信息~~ |
-| Data | 資料 | ~~數據~~ |
-| Network | 網路 | ~~網絡~~ |
-| Software | 軟體 | ~~軟件~~ |
-| Hardware | 硬體 | ~~硬件~~ |
-| Default | 預設 | ~~默認~~ |
-| Support | 支援 | ~~支持~~ |
-| Recursive | 遞迴 | ~~遞歸~~ |
-| Iterate | 迭代 | ~~迭代~~ (same) |
-| Loop | 迴圈 | ~~循環~~ |
-| Execute | 執行 | ~~執行~~ (same) |
-
-### Standard Taiwan CS Terminology
+## ⚠️ CRITICAL: Taiwan DSA Terminology Standards
+
+You are translating for **Taiwan's Computer Science community**. Taiwan uses different terminology from Mainland China. Using Mainland terms will immediately mark the document as "非台灣體系" (non-Taiwan system).
+
+---
+
+## 🚨 A-Level: ZERO TOLERANCE (Must Replace)
+
+These terms will **100% be identified as Mainland Chinese** by Taiwan CS readers. **NEVER use the left column.**
+
+| ❌ 禁用 (NEVER USE) | ✅ 台灣標準 (USE THIS) | English |
+|---------------------|------------------------|---------|
+| 字符串 | **字串** | String |
+| 字符 | **字元** | Character |
+| 指针 / 指針 | **指標** | Pointer |
+| 就地 | **原地** | In-place |
+| 枚举 / 枚舉 | **列出 / 逐一產生** (動詞); **窮舉** (名詞) | Enumerate |
+| 搜索 | **搜尋** | Search |
+| 修剪 | **剪枝** | Prune/Pruning |
+| 映射 | **對應表 / 對照表** | Mapping |
+| 窗口 | **視窗** | Window |
+| 運行 | **執行** | Run/Execute |
+| 單元格 | **格子** | Cell (grid) |
+| 前沿 | **frontier / 邊界** | Frontier |
+| 链表 / 鏈表 | **鏈結串列** | Linked List |
+| 数组 / 數組 | **陣列** | Array |
+| 哈希 / 哈希表 | **雜湊 / 雜湊表** | Hash / Hash Table |
+| 堆栈 | **堆疊** | Stack |
+| 布尔 / 布爾 | **布林** | Boolean |
+| 函数 / 函數 | **函式** | Function |
+| 变量 / 變量 | **變數** | Variable |
+| 内存 / 內存 | **記憶體** | Memory |
+| 程序 | **程式** | Program |
+| 对象 / 對象 | **物件** | Object |
+| 接口 | **介面** | Interface |
+| 实现 / 實現 | **實作** | Implementation |
+| 信息 | **資訊** | Information |
+| 数据 / 數據 | **資料** | Data |
+| 网络 / 網絡 | **網路** | Network |
+| 软件 / 軟件 | **軟體** | Software |
+| 硬件 / 硬件 | **硬體** | Hardware |
+| 默认 / 默認 | **預設** | Default |
+| 支持 | **支援** | Support |
+| 递归 / 遞歸 | **遞迴** | Recursive |
+| 循环 / 循環 | **迴圈** | Loop |
+| 调用 / 調用 | **呼叫** | Call (function) |
+
+---
+
+## ⚠️ B-Level: SHOULD REPLACE (Taiwan Preference)
+
+These won't break the document but will make it "sound like Mainland notes." **Prefer Taiwan terms.**
+
+| 🔶 中國偏用 (Avoid) | ✅ 台灣慣用 (Prefer) | English |
+|---------------------|----------------------|---------|
+| 遍历 / 遍歷 (as noun) | **走訪 / 逐一處理** | Traversal |
+| 搜索树 / 搜索樹 | **搜尋樹** | Search Tree |
+| 子串 | **子字串** | Substring |
+| 区间 / 區間 | **區間** (OK, but 範圍 also works) | Interval |
+| 前缀 / 前綴 | **前綴** | Prefix |
+| 后缀 / 後綴 | **後綴** | Suffix |
+| 队列 / 隊列 | **佇列** | Queue |
+| 入队 / 入隊 | **加入佇列 / enqueue** | Enqueue |
+| 出队 / 出隊 | **移出佇列 / dequeue** | Dequeue |
+| 权重 / 權重 | **權重 / weight** | Weight |
+| 覆盖 / 覆蓋 (cover) | **涵蓋 / 包含** | Cover |
+| 边界情况 / 邊界情況 | **邊界情況 / edge case** | Edge Case |
+| 节点 / 節點 | **節點** (OK, ensure consistent) | Node |
+
+---
+
+## ⚠️ C-Level: 語感問題 (Sounds Like Mainland Teaching Materials)
+
+These are not "wrong" but will make Taiwan readers feel the text is "not local." **Strongly recommend replacing.**
+
+| 🔶 陸系語感 (Avoid) | ✅ 台灣自然說法 (Prefer) | Context |
+|---------------------|-------------------------|---------|
+| 變體 | **變形 / 延伸題 / 變化題 / 進階題** | Problem variants |
+| 列舉 (名詞化) | **列出 / 找出** | "列舉所有解" → "列出所有解" |
+| 系統映射 | **系統對應 / 系統對照** | System mapping |
+| 防護欄 | **注意事項 / 限制 / 實作注意** | Guardrails |
+| 有效性 | **成立條件 / 判定條件** | Validity |
+| 有效 (狀態) | **成立 / 合法** | "當有效時" → "當成立時" |
+| 無效 (狀態) | **不成立 / 不合法** | Invalid state |
+| 取捨 | **權衡** | Trade-offs |
+| 目標 (列表式) | **求解目標 / 要求** | "目標：存在" → "求解目標：存在" |
+| 實作不變量 | **實作時的不變量** | Implementation invariant |
+
+---
+
+## ✅ Taiwan Standard CS Terminology Reference
 
 | English | 台灣繁體中文 |
 |---------|-------------|
@@ -55,17 +110,17 @@ The following terms differ between Taiwan (台灣) and Mainland China (中國大
 | Sorting | 排序 |
 | Sliding Window | 滑動視窗 |
 | Dynamic Programming | 動態規劃 |
-| Backtracking | 回溯 |
+| Backtracking | 回溯法 |
 | Greedy | 貪婪法 |
 | Divide and Conquer | 分治法 |
-| BFS (Breadth-First Search) | 廣度優先搜尋 (BFS) |
-| DFS (Depth-First Search) | 深度優先搜尋 (DFS) |
+| BFS | 廣度優先搜尋 (BFS) |
+| DFS | 深度優先搜尋 (DFS) |
 | Traversal | 走訪 |
 | Node | 節點 |
 | Edge | 邊 |
 | Vertex | 頂點 |
 | Index | 索引 |
-| Invariant | 不變量 |
+| Invariant | 不變量 / 不變式 |
 | Complexity | 複雜度 |
 | Time Complexity | 時間複雜度 |
 | Space Complexity | 空間複雜度 |
@@ -80,20 +135,25 @@ The following terms differ between Taiwan (台灣) and Mainland China (中國大
 | Frequency | 頻率 |
 | Counter | 計數器 |
 | Window | 視窗 |
+| Sliding Window | 滑動視窗 |
 | Shrink | 收縮 |
 | Expand | 擴展 |
-| Valid | 有效 |
-| Invalid | 無效 |
+| Cell (grid) | 格子 |
+| Frontier | frontier / 邊界 |
+| Run/Execute | 執行 |
+| Valid | 有效 / 合法 |
+| Invalid | 無效 / 不合法 |
 | Target | 目標 |
 | Template | 模板 |
 | Pattern | 模式 |
 | State Machine | 狀態機 |
-| Wavefront | 波前 |
-| Streaming | 流式 |
+| Pointer | 指標 |
+| Two Pointers | 雙指標 |
+| Fast-Slow Pointers | 快慢指標 |
 
 ---
 
-## DO NOT Translate (Keep in English)
+## 🔒 DO NOT Translate (Keep in English)
 
 ### 1. API Kernel Names (Class-style identifiers)
 Keep these EXACTLY as-is:
@@ -116,26 +176,11 @@ Keep these EXACTLY as-is:
 - `sliding_window_cost_bounded`
 - `two_pointer_opposite_maximize`
 - `two_pointer_three_sum`
-- `two_pointer_opposite_palindrome`
-- `two_pointer_writer_dedup`
-- `two_pointer_writer_remove`
-- `two_pointer_writer_compact`
-- `fast_slow_cycle_detect`
-- `fast_slow_cycle_start`
-- `fast_slow_midpoint`
-- `fast_slow_implicit_cycle`
 - `dutch_flag_partition`
-- `two_way_partition`
 - `quickselect_partition`
 - `merge_two_sorted_lists`
-- `merge_two_sorted_arrays`
-- `merge_sorted_from_ends`
-- `merge_k_sorted_heap`
-- `merge_k_sorted_divide`
 - `heap_kth_element`
-- `linked_list_k_group_reversal`
-- `backtracking_n_queens`
-- `grid_bfs_propagation`
+- `fast_slow_cycle_detect`
 - Any other `snake_case` pattern identifiers
 
 ### 3. Code Elements
@@ -153,27 +198,45 @@ Keep these EXACTLY as-is:
 - Keep link text that contains problem names: "[LeetCode 3 - Longest Substring...]"
 
 ### 6. Table Headers with Technical Terms
-- Keep column headers like "Invariant", "State", "Goal" in the pattern tables
+- Keep column headers like "Invariant", "State", "Goal" in pattern tables
 - These are technical terms that match code concepts
 
 ---
 
 ## Translation Rules
 
 1. **Preserve Formatting**: Keep ALL Markdown formatting exactly (headers, lists, links, checkboxes, code blocks, tables)
-2. **Translate**:
-   - Section headings (but keep API Kernel names in English)
-   - Descriptive text and explanations
-   - Emoji labels are fine to keep
-3. **Hybrid Headers**: For headers like "### SubstringSlidingWindow — *1D window state machine*"
+2. **Hybrid Headers**: For headers like "### SubstringSlidingWindow — *1D window state machine*"
    - Keep `SubstringSlidingWindow` in English
    - Translate the description part: "一維視窗狀態機"
-4. **Preserve Structure**: Maintain the same tree structure and indentation
-5. **Style**: Use Taiwan's technical documentation style - concise and professional
+3. **Preserve Structure**: Maintain the same tree structure and indentation
+4. **Style**: Use Taiwan's technical documentation style - concise, professional, academic tone
 
 ---
 
-## Output
+## Self-Check Before Output
 
-Output ONLY the translated Markdown content. No explanations, no code fence wrappers.
+Scan your translation for these terms. If ANY appear, you have failed:
+
+**A-Level (零容忍):**
+```
+字符串, 字符, 指针, 指針, 就地, 枚举, 枚舉, 搜索, 修剪, 
+映射, 数组, 數組, 链表, 鏈表, 哈希, 堆栈, 布尔, 布爾,
+函数, 函數, 变量, 變量, 内存, 內存, 程序, 对象, 對象,
+接口, 实现, 實現, 信息, 数据, 數據, 网络, 網絡, 
+软件, 軟件, 硬件, 默认, 默認, 支持, 递归, 遞歸, 循环, 循環,
+窗口, 運行, 單元格, 前沿
+```
+
+**C-Level (語感問題 - 強烈建議避免):**
+```
+變體, 系統映射, 防護欄, 有效性, 取捨
+```
+- 「列舉」只能當動詞用，不要名詞化
+- 「有效/無效」改用「成立/不成立」或「合法/不合法」
+
+---
+
+## Output
 
+Output ONLY the translated Markdown content. No explanations, no code fence wrappers around the output.

tools/ai-markmap-agent/outputs/versions/v1/neetcode_ontology_agent_evolved_en.md

@@ -271,16 +271,35 @@ def _parse_adoption_list(self, response: str) -> AdoptionList:
         """Parse adoption list from discussion response."""
         adopted_ids = []
 
-        # Look for adoption list section
-        adoption_section = re.search(
-            r'(?:Final Adoption List|My Final Adoption|I recommend adopting).*?(?=##|$)',
-            response,
-            re.IGNORECASE | re.DOTALL
-        )
-        
-        if adoption_section:
-            section_text = adoption_section.group(0)
+        # Strategy 1: Look for explicit adoption section
+        # The regex was failing because "###" contains "##" 
+        # Use a more robust pattern: find the adoption header and take everything after it
+        adoption_patterns = [
+            r'(?:^|\n)#+\s*(?:My\s+)?Final\s+Adoption\s+List.*',  # "### My Final Adoption List"
+            r'I\s+recommend\s+adopting\s+(?:these\s+)?suggestions?:?\s*\n.*',  # "I recommend adopting..."
+            r'(?:^|\n)#+\s*Part\s*2\s*:?\s*Final\s+Adoption.*',  # "## Part 2: Final Adoption..."
+        ]
+        
+        section_text = ""
+        for pattern in adoption_patterns:
+            match = re.search(pattern, response, re.IGNORECASE | re.DOTALL)
+            if match:
+                # Take from match position to end of response
+                section_text = response[match.start():]
+                break
+        
+        # Strategy 2: If no explicit section found, look for all ✅ Agree votes
+        if not section_text:
+            # Fallback: collect IDs from ✅ Agree vote lines
+            agree_pattern = r'\*\*Vote\*\*:\s*✅\s*Agree.*?(?:^|\n)#+\s*([APE]\d+)'
+            agrees = re.findall(agree_pattern, response, re.IGNORECASE | re.DOTALL | re.MULTILINE)
+            if agrees:
+                adopted_ids = list(dict.fromkeys(agrees))
+        
+        # Extract IDs from section text
+        if section_text:
             # Find all suggestion IDs (A1, P2, E3, etc.)
+            # Match IDs that appear in list items or bold text
             ids = re.findall(r'\b([APE]\d+)\b', section_text)
             adopted_ids = list(dict.fromkeys(ids))  # Remove duplicates, preserve order
 

tools/ai-markmap-agent/outputs/versions/v1/neetcode_ontology_agent_evolved_zh-TW.html

@@ -184,6 +184,9 @@ def main() -> int:
                 return 1
             print(f"\n📂 Found latest output: {input_path}")
 
+        # Create converter for output path resolution
+        converter = MarkMapHTMLConverter(config)
+        
         # Determine output file
         if args.output:
             output_path = Path(args.output)
@@ -196,7 +199,9 @@ def main() -> int:
                 new_stem = stem[:-len(suffix)] + f"_{args.target}"
             else:
                 new_stem = f"{stem}_{args.target}"
-            output_path = input_path.parent / f"{new_stem}.md"
+            
+            # Use final_dirs.markdown from config for consistency with HTML output
+            output_path = converter.md_output_dir / f"{new_stem}.md"
 
         # Determine model
         model = args.model
@@ -219,14 +224,12 @@ def main() -> int:
         # Generate HTML if requested
         if args.html:
             print("\n📊 Generating HTML...")
-            converter = MarkMapHTMLConverter(config)
             html_content = converter.convert(
                 translated,
                 title=f"NeetCode Agent Evolved Mindmap ({args.target.upper()})"
             )
             # Use correct HTML output directory from config
-            html_dir = converter.html_output_dir
-            html_path = html_dir / f"{output_path.stem}.html"
+            html_path = converter.html_output_dir / f"{output_path.stem}.html"
             html_path.write_text(html_content, encoding="utf-8")
             print(f"   ✓ Saved: {html_path}")