Add a detailed handoff doc for future development sessions

Constraint: New sessions need a fast, durable understanding of the project state, open-dataset workflow, verified evidence, and next steps
Rejected: Rely on scattered docs and git history alone | Too slow for session handoff and easy to miss critical workflow context
Confidence: high
Scope-risk: narrow
Directive: Keep this handoff doc updated whenever a major workflow milestone or verified capability changes
Tested: existence checks for docs/session-handoff.md and docs/README.md, plus docs index link presence
Not-tested: Manual human review across multiple markdown renderers

@@ -69,6 +69,7 @@ flowchart TD
 
 ### C. 服务与工程
 - [服务接口](./service-api.md)
+- [持续开发交接文档](./session-handoff.md)
 - [更新记录](./CHANGELOG.md)
 
 ### D. 研究与路线

+# Session Handoff / 持续开发交接文档
+
+> 更新：2026-06-02  
+> 目的：让新 session / 新代理进入仓库后，能在最短时间内理解项目现状并继续开发。
+
+## 一页结论
+
+这是一个正在从原型向工业化推进的 **音乐 ACR / music retrieval** 项目。  
+当前已经完成：
+
+1. **原型可运行**
+   - synthetic 数据生成
+   - 训练
+   - 建索引
+   - 识别
+   - 评测
+
+2. **开放数据接入链路完整闭环**
+   - inspect-local / inspect-batch
+   - prepare-local
+   - validate-local
+   - train
+   - build-index
+   - evaluate
+   - generate_artifacts
+
+3. **文档已浓缩**
+   - docs 入口已分成 4 组
+   - 相对路径支持跳转
+   - 开放数据工作流有单页文档
+
+当前最重要的下一步：
+- 用真实本地 FMA / MTG-Jamendo 音频目录替换 synthetic stand-in
+- 跑真实开放数据 smoke
+- 继续优化准确率，尤其是 `confused` / `humming_like`
+
+---
+
+## 1. 项目是什么
+
+这是一个面向**音乐片段识别 / 音乐检索**的 ACR 引擎，核心路线是：
+
+- 指纹检索（Chromaprint-like）
+- embedding 检索（ECAPA-derived）
+- 可选 melody-aware 融合
+- retrieval-first 评测与优化
+
+它已经不是单纯的“分类模型训练脚本”，而是一个较完整的工程原型：
+- 数据层
+- 训练层
+- 索引层
+- 识别层
+- 评测层
+- 文档层
+- 开放数据接入层
+- 发布产物层
+
+---
+
+## 2. 你应该先看哪些文档
+
+### 核心 4 组入口
+- [docs/README.md](./README.md)
+- [docs/open-dataset-workflow.md](./open-dataset-workflow.md)
+- [docs/dataset-spec.md](./dataset-spec.md)
+- [docs/industrialization-roadmap.md](./industrialization-roadmap.md)
+
+### 如果你是算法/模型方向
+- [docs/dataset-spec.md](./dataset-spec.md)
+- [docs/sota-research-2026.md](./sota-research-2026.md)
+- [docs/industrial-benchmark-spec.md](./industrial-benchmark-spec.md)
+
+### 如果你是数据接入方向
+- [docs/open-dataset-workflow.md](./open-dataset-workflow.md)
+- [docs/dataset-sources-and-licensing.md](./dataset-sources-and-licensing.md)
+- [acr-engine/data/raw/README.md](../acr-engine/data/raw/README.md)
+
+### 如果你是工程/服务方向
+- [docs/service-api.md](./service-api.md)
+- [docs/CHANGELOG.md](./CHANGELOG.md)
+
+---
+
+## 3. 当前代码结构重点
+
+### 训练与评测主入口
+- [acr-engine/train.py](../acr-engine/train.py)
+- [acr-engine/evaluate.py](../acr-engine/evaluate.py)
+- [acr-engine/run_demo.py](../acr-engine/run_demo.py)
+
+### 数据层
+- [acr-engine/src/data/dataset.py](../acr-engine/src/data/dataset.py)
+- [acr-engine/src/data/synthetic.py](../acr-engine/src/data/synthetic.py)
+- [acr-engine/src/data/manifest_tools.py](../acr-engine/src/data/manifest_tools.py)
+- [acr-engine/src/data/external_adapters.py](../acr-engine/src/data/external_adapters.py)
+
+### 检索与模型层
+- [acr-engine/src/engines/hybrid_engine.py](../acr-engine/src/engines/hybrid_engine.py)
+- [acr-engine/src/engines/ecapa_embedder.py](../acr-engine/src/engines/ecapa_embedder.py)
+- [acr-engine/src/engines/chromaprint_matcher.py](../acr-engine/src/engines/chromaprint_matcher.py)
+- [acr-engine/src/models/ecapa_tdnn.py](../acr-engine/src/models/ecapa_tdnn.py)
+- [acr-engine/src/models/losses.py](../acr-engine/src/models/losses.py)
+
+### 服务层
+- [acr-engine/src/service/app.py](../acr-engine/src/service/app.py)
+
+---
+
+## 4. 已经完成的关键能力
+
+### 4.1 原型与 synthetic 数据
+- synthetic dataset 可生成
+- `train.py --dry-run` 可通过
+- 可训练出 checkpoint
+- 可 build-index
+- 可 recognize
+- 可 evaluate
+
+### 4.2 开放数据接入
+已经具备以下命令：
+
+- `inspect-local`
+- `inspect-batch`
+- `prepare-local`
+- `validate-local`
+- `smoke-local`
+
+这些都在：
+- [acr-engine/src/data/external_adapters.py](../acr-engine/src/data/external_adapters.py)
+
+### 4.3 文档与发布产物
+开放数据 smoke 也能生成：
+- benchmark report
+- model card
+- release checklist
+- artifact manifest
+
+---
+
+## 5. 开放数据当前的实际工作方式
+
+### 真实音频应该放到哪里
+- [acr-engine/data/raw/fma_small_audio/](../acr-engine/data/raw/fma_small_audio/)
+- [acr-engine/data/raw/mtg_jamendo_audio/](../acr-engine/data/raw/mtg_jamendo_audio/)
+
+说明文件：
+- [acr-engine/data/raw/README.md](../acr-engine/data/raw/README.md)
+
+### 当前最推荐的命令
+
+#### FMA
+```bash
+/usr/local/miniconda3/bin/python src/data/external_adapters.py smoke-local fma data/raw/fma_small_audio --output-root data/external_smoke --eval-ratio 0.2 --query-duration 8.0 --train-epochs 1 --batch-size 2
+```
+
+#### MTG-Jamendo
+```bash
+/usr/local/miniconda3/bin/python src/data/external_adapters.py smoke-local mtg_jamendo data/raw/mtg_jamendo_audio --output-root data/external_smoke --eval-ratio 0.2 --query-duration 8.0 --train-epochs 1 --batch-size 2
+```
+
+### 当前 smoke-local 已验证能力
+`smoke-local` 会自动跑：
+1. inspect-local
+2. prepare-local
+3. validate-local
+4. train
+5. build-index
+6. evaluate
+7. generate_artifacts
+
+---
+
+## 6. 目前最重要的验证证据
+
+### 6.1 synthetic-as-open-fixed（开放数据 stand-in）
+已成功验证：
+- `prepare-local`
+- `validate-local`
+- `train.py`
+- `build-index`
+- `evaluate.py`
+- `generate_artifacts.py`
+
+关键结果：
+- `num_queries=8`
+- `top1=1.0`
+- `topk=1.0`
+
+相关目录：
+- [acr-engine/data/external_ingested/synthetic_as_open_fixed/](../acr-engine/data/external_ingested/synthetic_as_open_fixed/)
+- [acr-engine/reports/open-smoke-fixed/fma/](../acr-engine/reports/open-smoke-fixed/fma/)
+
+### 6.2 一键 smoke-local
+已验证：
+```bash
+/usr/local/miniconda3/bin/python src/data/external_adapters.py smoke-local fma data/synthetic_v2/songs --output-root data/external_smoke --eval-ratio 0.2 --query-duration 5.0 --train-epochs 1 --batch-size 2
+```
+
+关键结果：
+- `num_audio_files=24`
+- `catalog=24`
+- `train_queries=16`
+- `test_queries=8`
+- `top1=1.0`
+- `topk=1.0`
+
+相关目录：
+- [acr-engine/data/external_smoke/](../acr-engine/data/external_smoke/)
+
+---
+
+## 7. 当前最重要的待办
+
+### 优先级 A：真实开放数据替换
+目标：
+- 用真实本地 FMA / MTG-Jamendo 音频替换 synthetic stand-in
+
+操作：
+1. 把真实音频放进：
+   - `acr-engine/data/raw/fma_small_audio/`
+   - 或 `acr-engine/data/raw/mtg_jamendo_audio/`
+2. 直接运行 `smoke-local`
+3. 记录：
+   - inspect 规模
+   - train/test query 数
+   - top1/topk
+   - artifact bundle
+
+### 优先级 B：hard-case 精度继续优化
+当前历史结论：
+- naive oversampling：失败
+- type-aware weighting：部分有效
+- sample-level weighting：提升 `confused`
+- retrieval fusion tuning：更稳定有效
+
+下阶段重点：
+- `confused`
+- `humming_like`
+- 真实开放数据上的 hard-case bucket
+
+### 优先级 C：foundation model / SOTA baseline
+已经在文档中记录：
+- MERT
+- MuQ
+- 更强 retrieval-first 路线
+
+后续可以做：
+- frozen backbone baseline
+- adapter fine-tune
+
+---
+
+## 8. 最新关键提交（便于新 session 快速定位）
+
+近几次关键提交建议优先看：
+
+- `d221852` Add explicit drop zones for real open-music corpora
+- `eee15ac` Automate the full open-dataset smoke workflow behind one command
+- `8795907` Generate release artifacts for the open-dataset smoke path
+- `dc9ef1b` Close the open-dataset smoke loop through evaluation
+- `b766c74` Make open-dataset manifests trainable end to end
+- `fa23144` Add a single-page open dataset workflow for training prep
+- `af33be3` Condense docs and add manifest validation before training
+
+这些 commit 基本覆盖了当前开放数据与文档演进主线。
+
+---
+
+## 9. 新 session 接手时的推荐动作
+
+如果你是新的 session，建议顺序：
+
+1. 读：
+   - [docs/README.md](./README.md)
+   - [docs/open-dataset-workflow.md](./open-dataset-workflow.md)
+   - [docs/session-handoff.md](./session-handoff.md)
+
+2. 检查真实数据是否已落位：
+   - `acr-engine/data/raw/fma_small_audio/`
+   - `acr-engine/data/raw/mtg_jamendo_audio/`
+
+3. 如果已有真实音频：
+   - 直接跑 `smoke-local`
+
+4. 如果还没有真实音频：
+   - 继续优化 synthetic-as-open-fixed
+   - 或继续补开放数据下载/清洗自动化
+
+5. 每完成一个阶段：
+   - 更新 [docs/CHANGELOG.md](./CHANGELOG.md)
+   - `git commit`
+   - `git push`
+
+---
+
+## 10. 注意事项
+
+- 这个仓库里存在已跟踪的 `__pycache__` 文件；提交时要小心不要让它们污染变更。
+- 当前最稳定的改进方向不是盲目调训练权重，而是：
+  - retrieval-time fusion
+  - 更真实开放数据
+  - 更真实评测
+- 开放数据布局现在依赖“自包含输出根”：
+  - `audio/`
+  - `manifests/`
+  这一点后续不要破坏。
+
+---
+
+## Sources
+- [README.md](./README.md)
+- [open-dataset-workflow.md](./open-dataset-workflow.md)
+- [CHANGELOG.md](./CHANGELOG.md)