Enhance arXiv digest with plain-language Top3 summaries

org/cases/arxiv_digest/arxiv_daily_digest.py

@@ -105,6 +105,18 @@ def one_liner(p):
     return f"{t}（{p['primary']}）"
 
 
+def plain_summary(p):
+    s = p.get('summary', '')
+    s = re.sub(r'\s+', ' ', s).strip()
+    if not s:
+        return '这篇主要在提出一个新方法，并用实验验证它是否更好。'
+    # take first sentence-like chunk
+    chunk = re.split(r'(?<=[\.!?])\s+', s)[0]
+    if len(chunk) > 120:
+        chunk = chunk[:117] + '...'
+    return f"它在做的事：{chunk}"
+
+
 def build_digest(papers):
     # limit scoring set for speed/stability
     recent_pool = sorted(papers, key=lambda x: x['published'], reverse=True)[:40]
@@ -125,6 +137,12 @@ def to_md(hot, latest):
     lines = []
     lines.append(f"# ArXiv Daily Brief - {today}")
     lines.append('')
+    lines.append('## 🧠 今日 Top 3（普通人版概述）')
+    for i, p in enumerate(hot[:3], 1):
+        lines.append(f"{i}. **{p['title']}**")
+        lines.append(f"   - 一句话看懂: {plain_summary(p)}")
+        lines.append(f"   - 你可能会关心: {one_liner(p)}")
+    lines.append('')
     lines.append('## 🔥 今日热度 Top 5（新鲜度+关键词+HN提及+代码线索）')
     for i, p in enumerate(hot, 1):
         lines.append(f"{i}. **{p['title']}**")

org/cases/arxiv_digest/output/2026-03-08.md

@@ -1,25 +1,36 @@
 # ArXiv Daily Brief - 2026-03-08
 
+## 🧠 今日 Top 3（普通人版概述）
+1. **SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival Analysis**
+   - 一句话看懂: 它在做的事：Estimating heterogeneous treatment effects (HTEs) from right-censored survival data is critical in high-stakes applic...
+   - 你可能会关心: SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival An...（cs.LG）
+2. **Accelerating Text-to-Video Generation with Calibrated Sparse Attention**
+   - 一句话看懂: 它在做的事：Recent diffusion models enable high-quality video generation, but suffer from slow runtimes.
+   - 你可能会关心: Accelerating Text-to-Video Generation with Calibrated Sparse Attention（cs.CV）
+3. **Observing and Controlling Features in Vision-Language-Action Models**
+   - 一句话看懂: 它在做的事：Vision-Language-Action Models (VLAs) have shown remarkable progress towards embodied intelligence.
+   - 你可能会关心: Observing and Controlling Features in Vision-Language-Action Models（cs.RO）
+
 ## 🔥 今日热度 Top 5（新鲜度+关键词+HN提及+代码线索）
 1. **SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival Analysis**
    - arXiv: http://arxiv.org/abs/2603.05483v1
-   - 类别: cs.LG | HotScore: 33.06 | 作者: Shahriar Noroozizadeh, Xiaobin Shen, Jeremy C. Weiss
+   - 类别: cs.LG | HotScore: 32.79 | 作者: Shahriar Noroozizadeh, Xiaobin Shen, Jeremy C. Weiss
    - 速读: SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival An...（cs.LG）
 2. **Accelerating Text-to-Video Generation with Calibrated Sparse Attention**
    - arXiv: http://arxiv.org/abs/2603.05503v1
-   - 类别: cs.CV | HotScore: 28.17 | 作者: Shai Yehezkel, Shahar Yadin, Noam Elata
+   - 类别: cs.CV | HotScore: 27.9 | 作者: Shai Yehezkel, Shahar Yadin, Noam Elata
    - 速读: Accelerating Text-to-Video Generation with Calibrated Sparse Attention（cs.CV）
 3. **Observing and Controlling Features in Vision-Language-Action Models**
    - arXiv: http://arxiv.org/abs/2603.05487v1
-   - 类别: cs.RO | HotScore: 28.08 | 作者: Hugo Buurmeijer, Carmen Amo Alonso, Aiden Swann
+   - 类别: cs.RO | HotScore: 27.81 | 作者: Hugo Buurmeijer, Carmen Amo Alonso, Aiden Swann
    - 速读: Observing and Controlling Features in Vision-Language-Action Models（cs.RO）
 4. **Towards Provably Unbiased LLM Judges via Bias-Bounded Evaluation**
    - arXiv: http://arxiv.org/abs/2603.05485v1
-   - 类别: cs.AI | HotScore: 27.06 | 作者: Benjamin Feuer, Lucas Rosenblatt, Oussama Elachqar
+   - 类别: cs.AI | HotScore: 26.8 | 作者: Benjamin Feuer, Lucas Rosenblatt, Oussama Elachqar
    - 速读: Towards Provably Unbiased LLM Judges via Bias-Bounded Evaluation（cs.AI）
 5. **An interpretable prototype parts-based neural network for medical tabular data**
    - arXiv: http://arxiv.org/abs/2603.05423v1
-   - 类别: cs.LG | HotScore: 26.1 | 作者: Jacek Karolczak, Jerzy Stefanowski
+   - 类别: cs.LG | HotScore: 25.83 | 作者: Jacek Karolczak, Jerzy Stefanowski
    - 速读: An interpretable prototype parts-based neural network for medical tabular data（cs.LG）
 
 ## 🆕 最新上新 Top 10

org/cases/arxiv_digest/output/latest.md

@@ -1,25 +1,36 @@
 # ArXiv Daily Brief - 2026-03-08
 
+## 🧠 今日 Top 3（普通人版概述）
+1. **SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival Analysis**
+   - 一句话看懂: 它在做的事：Estimating heterogeneous treatment effects (HTEs) from right-censored survival data is critical in high-stakes applic...
+   - 你可能会关心: SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival An...（cs.LG）
+2. **Accelerating Text-to-Video Generation with Calibrated Sparse Attention**
+   - 一句话看懂: 它在做的事：Recent diffusion models enable high-quality video generation, but suffer from slow runtimes.
+   - 你可能会关心: Accelerating Text-to-Video Generation with Calibrated Sparse Attention（cs.CV）
+3. **Observing and Controlling Features in Vision-Language-Action Models**
+   - 一句话看懂: 它在做的事：Vision-Language-Action Models (VLAs) have shown remarkable progress towards embodied intelligence.
+   - 你可能会关心: Observing and Controlling Features in Vision-Language-Action Models（cs.RO）
+
 ## 🔥 今日热度 Top 5（新鲜度+关键词+HN提及+代码线索）
 1. **SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival Analysis**
    - arXiv: http://arxiv.org/abs/2603.05483v1
-   - 类别: cs.LG | HotScore: 33.06 | 作者: Shahriar Noroozizadeh, Xiaobin Shen, Jeremy C. Weiss
+   - 类别: cs.LG | HotScore: 32.79 | 作者: Shahriar Noroozizadeh, Xiaobin Shen, Jeremy C. Weiss
    - 速读: SurvHTE-Bench: A Benchmark for Heterogeneous Treatment Effect Estimation in Survival An...（cs.LG）
 2. **Accelerating Text-to-Video Generation with Calibrated Sparse Attention**
    - arXiv: http://arxiv.org/abs/2603.05503v1
-   - 类别: cs.CV | HotScore: 28.17 | 作者: Shai Yehezkel, Shahar Yadin, Noam Elata
+   - 类别: cs.CV | HotScore: 27.9 | 作者: Shai Yehezkel, Shahar Yadin, Noam Elata
    - 速读: Accelerating Text-to-Video Generation with Calibrated Sparse Attention（cs.CV）
 3. **Observing and Controlling Features in Vision-Language-Action Models**
    - arXiv: http://arxiv.org/abs/2603.05487v1
-   - 类别: cs.RO | HotScore: 28.08 | 作者: Hugo Buurmeijer, Carmen Amo Alonso, Aiden Swann
+   - 类别: cs.RO | HotScore: 27.81 | 作者: Hugo Buurmeijer, Carmen Amo Alonso, Aiden Swann
    - 速读: Observing and Controlling Features in Vision-Language-Action Models（cs.RO）
 4. **Towards Provably Unbiased LLM Judges via Bias-Bounded Evaluation**
    - arXiv: http://arxiv.org/abs/2603.05485v1
-   - 类别: cs.AI | HotScore: 27.06 | 作者: Benjamin Feuer, Lucas Rosenblatt, Oussama Elachqar
+   - 类别: cs.AI | HotScore: 26.8 | 作者: Benjamin Feuer, Lucas Rosenblatt, Oussama Elachqar
    - 速读: Towards Provably Unbiased LLM Judges via Bias-Bounded Evaluation（cs.AI）
 5. **An interpretable prototype parts-based neural network for medical tabular data**
    - arXiv: http://arxiv.org/abs/2603.05423v1
-   - 类别: cs.LG | HotScore: 26.1 | 作者: Jacek Karolczak, Jerzy Stefanowski
+   - 类别: cs.LG | HotScore: 25.83 | 作者: Jacek Karolczak, Jerzy Stefanowski
    - 速读: An interpretable prototype parts-based neural network for medical tabular data（cs.LG）
 
 ## 🆕 最新上新 Top 10