官方说明开源项目7 天前
Th0rgal/open-exoplanet-hunt
Open, falsification-first collaborative search for unreported exoplanets in public NASA TESS data (humans & AI agents welcome)
在 GitHub 查看 ↗直链 · 无中转页
- ★ 数取值
- 0 ★ · Th0rgal/open-exoplanet-hunt · 取自该仓库的 GitHub 页面 · （2026-10-16 20:00:20 核对）
- 来源明细
- 从该仓库的 GitHub 页面直接读取 · 条目类型：开源项目
- 收录与活跃
- 收录于 2026-10-10 · 最近见到 2026-10-10 · 共见到 3 次 · 被 1 个来源列出
- 评分构成
- 综合分 103 · 1 个来源 · repo
常见问题（答案都来自我们的记录）
怎么安装？
我们【没有】从该仓库取到安装命令（我们的 install 覆盖率只有约 4%）。请看它的 README 或 releases 页。
它适合谁 / 用来做什么？
我们把它归在这些标签下: 开源项目。标签来自条目自身，不是我们强加的分类。
谁维护它、还在更新吗？
仓库
Th0rgal/open-exoplanet-hunt；我们收录于 2026-10-10；最近见到 2026-10-10。这里没有"官方认证"或"质量评级"——我们只记录来源与时间。第三方质量标记
🪐 Open Exoplanet Hunt (open-exoplanet-hunt)
以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →
**An open, falsification-first collaborative search for unreported exoplanets in public NASA TESS data — humans & AI coding agents welcome.**
> *"The useful way to use an AI agent for science isn't to ask 'Is this a planet?' — it's to keep giving it ways to prove that it isn't."*
> — Inspired by Pavel Rabtsevich's TIC 4206066 discovery workflow (Oct 2026)
---
🔭 What Is This Repository?
Seven years of public photometry from NASA's **Transiting Exoplanet Survey Satellite (TESS)** contain millions of multi-epoch light curves. While single-pipeline alerts (TOI) and single-year FFI surveys catch high-SNR transits, **real small planets ($1.0\text{–}2.5\,R_\oplus$) around nearby K and M dwarfs still hide in the cross-pipeline multi-year blind spot**:
- In TESS Years 1–2 (2018–2020), many nearby K/M dwarfs were only recorded in 30-minute Full-Frame Images (HLSP/tess-spoc, HLSP/qlp).
- In Years 3–4 (2020–2022), some received a single sector of 2-minute SPOC cadence where 4–8 shallow transits fell just short of the multi-sector alert threshold.
- In Years 5–8 (2023–2026), they were observed again in 200-second FFIs (HLSP/qlp).
When stitched coherently across a 5-to-8-year baseline (TESS SPOC + HLSP/tess-spoc + HLSP/qlp), sub-threshold per-sector signals lock in phase at $\text{SNR} > 10\text{–}16\sigma$ — **provided we ruthlessly falsify instrumental artifacts, eclipsing binaries, background blends, and starspot modulation**.
This repository turns that search into an **open, git-native, machine-verified collaborative benchmark** where human researchers, citizen scientists, and AI coding agents (Claude Code, Codex, Antigravity, Cursor) collaborate via Pull Requests:
1. **Master Index of What Has Already Been Found (catalog/ & docs/EXISTING_DISCOVERIES_INDEX.md)**:
- Indexes **15,025 unique TIC stars** across **6,445 confirmed NASA exoplanets (pscomppars)**, **8,148 official NASA TOIs**, **5,138 community CTOIs (ExoFOP-TESS)**, and external AI/ML/citizen-science discoveries (TIC 4206066, Warwick RAVEN, NASA ExoMiner++, Princeton T16, Planet Hunters TESS, SHERLOCK).
2. **10-Gate Adversarial Falsification Engine (scripts/vet_candidate.py)**:
- Every candidate must survive 10 quantitative falsification gates directly from raw MAST FITS files (including **Blind Leave-One-Epoch-Out (LOEO) temporal prediction**, odd/even depth consistency, phase-$0.5$ secondary eclipse null test, in-transit pixel centroid shift $ 0$) |
| **6** | gate6_secondary_eclipse_null | Self-luminous companions & phase-$0.5$ sub-harmonic crests | Phase $0.5 P$ anomaly has $\|\text{SNR}\| 0$, **Jackknife SNR** $\ge 6.5\,\sigma$, and median transit depth $\ge 55\%$ of weighted depth |
| **10** | gate10_rotational_modulation_guard | Starspot rotational modulation, rational sub-harmonics & multi-harmonic spot profiles | Out-of-transit detrended amplitude across $P, \frac{2}{3}P, \frac{1}{2}P, \frac{2}{5}P, \frac{1}{3}P, \frac{1}{4}P, \frac{1}{5}P$ & joint 4-harmonic Fourier profile $
> *"The useful way to use an AI agent for science isn't to ask 'Is this a planet?' — it's to keep giving it ways to prove that it isn't."*
> — Inspired by Pavel Rabtsevich's TIC 4206066 discovery workflow (Oct 2026)
---
🔭 What Is This Repository?
Seven years of public photometry from NASA's **Transiting Exoplanet Survey Satellite (TESS)** contain millions of multi-epoch light curves. While single-pipeline alerts (TOI) and single-year FFI surveys catch high-SNR transits, **real small planets ($1.0\text{–}2.5\,R_\oplus$) around nearby K and M dwarfs still hide in the cross-pipeline multi-year blind spot**:
- In TESS Years 1–2 (2018–2020), many nearby K/M dwarfs were only recorded in 30-minute Full-Frame Images (HLSP/tess-spoc, HLSP/qlp).
- In Years 3–4 (2020–2022), some received a single sector of 2-minute SPOC cadence where 4–8 shallow transits fell just short of the multi-sector alert threshold.
- In Years 5–8 (2023–2026), they were observed again in 200-second FFIs (HLSP/qlp).
When stitched coherently across a 5-to-8-year baseline (TESS SPOC + HLSP/tess-spoc + HLSP/qlp), sub-threshold per-sector signals lock in phase at $\text{SNR} > 10\text{–}16\sigma$ — **provided we ruthlessly falsify instrumental artifacts, eclipsing binaries, background blends, and starspot modulation**.
This repository turns that search into an **open, git-native, machine-verified collaborative benchmark** where human researchers, citizen scientists, and AI coding agents (Claude Code, Codex, Antigravity, Cursor) collaborate via Pull Requests:
1. **Master Index of What Has Already Been Found (catalog/ & docs/EXISTING_DISCOVERIES_INDEX.md)**:
- Indexes **15,025 unique TIC stars** across **6,445 confirmed NASA exoplanets (pscomppars)**, **8,148 official NASA TOIs**, **5,138 community CTOIs (ExoFOP-TESS)**, and external AI/ML/citizen-science discoveries (TIC 4206066, Warwick RAVEN, NASA ExoMiner++, Princeton T16, Planet Hunters TESS, SHERLOCK).
2. **10-Gate Adversarial Falsification Engine (scripts/vet_candidate.py)**:
- Every candidate must survive 10 quantitative falsification gates directly from raw MAST FITS files (including **Blind Leave-One-Epoch-Out (LOEO) temporal prediction**, odd/even depth consistency, phase-$0.5$ secondary eclipse null test, in-transit pixel centroid shift $ 0$) |
| **6** | gate6_secondary_eclipse_null | Self-luminous companions & phase-$0.5$ sub-harmonic crests | Phase $0.5 P$ anomaly has $\|\text{SNR}\| 0$, **Jackknife SNR** $\ge 6.5\,\sigma$, and median transit depth $\ge 55\%$ of weighted depth |
| **10** | gate10_rotational_modulation_guard | Starspot rotational modulation, rational sub-harmonics & multi-harmonic spot profiles | Out-of-transit detrended amplitude across $P, \frac{2}{3}P, \frac{1}{2}P, \frac{2}{5}P, \frac{1}{3}P, \frac{1}{4}P, \frac{1}{5}P$ & joint 4-harmonic Fourier profile $
- 来源
- github · https://github.com/Th0rgal/open-exoplanet-hunt
- 按领域浏览
- 开源项目
- 收录于
- 7 天前
更多开源项目
所属领域与标签
每个标签都能点进去看同类的全部条目。数字是当前库里的真实条数。