EditaPlot is an AI-guided local scientific plotting tool for Windows that uses Codex to understand data columns, recommend editable Origin figures, and export OPJU, PNG, PDF, and TIF files while maintaining scientific rigor by requiring user confirmation at each step.

Stars

82

7-day growth

No data

Forks

3

Open issues

0

License

Apache-2.0

Last updated

2026-07-28

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It bridges AI assistance with real Origin projects for fully editable figures, emphasizes transparency in data column usage, and refuses to make unwarranted inferences or modifications to the data.

Who it is for

  • Researchers and scientists who need editable Origin plots from raw data
  • Material scientists and chemists working with XRD, XPS, UV-Vis, and similar spectroscopy data
  • Medical and deep learning researchers generating ROC, box plots, and statistical graphs
  • Origin users seeking to automate plotting without losing editability

Use cases

  • Generate editable Origin figures from CSV/TXT/XLS/XLSX data with automatic column role identification
  • Create publication-ready plots for material characterization (XRD, XPS, DSC, NMR, etc.)
  • Use a reference image to guide plot style without copying data or text
  • Produce consistent visual outputs (OPJU, PNG, PDF, TIF) for lab reports and papers

Strengths

  • Transparent column classification (main evidence, auxiliary, computation-only, reserved) preventing data misrepresentation
  • Generates native Origin OPJU files that remain fully editable, not static previews
  • Covers over 40 chart types across materials, statistics, and biomedical domains
  • Strict scientific boundaries: no automatic baseline correction, peak fitting, model learning, or data imputation

Considerations

  • Currently limited to Windows 10/11 x64 physical machines; no macOS, Linux, WSL, or virtual machine support
  • Requires a locally installed Origin/OriginPro 2021–2026b (only 2024b fully verified) and Python 3.10–3.12
  • Reference image feature extracts only visual grammar, cannot replicate data, text, or complex layouts one-to-one

README quick start

EditaPlot · 艾迪图 AI 驱动的可编辑科研绘图工作流AI-guided editable scientific figures

English · 中文为主要说明语言

我把 EditaPlot 做成了一个面向 Codex 的 Windows 本地科研绘图 Skill。你把自己的实验数据交给它后,它会依次理解数据、逐列说明用途、推荐图形、请你确认图形元素、调用 Origin 并验证结果,最后生成可编辑 OPJU,同时导出 PNG、PDF、TIF。

我不希望它只是一套“替换数字”的静态模板,也不会让 Python 预览图冒充 Origin 成图。科学含义和最终选择始终由你决定;遇到把握不足的数据,EditaPlot 会把不确定的列单独列出来请你确认,不会擅自补列、拟合或推断结论。

[!WARNING] 我目前只完成了 Windows 10/11 x64 实体电脑上的完整验证。 因此 V1 暂未提供 macOS(Intel 与 Apple Silicon)、Linux、WSL、Wine/CrossOver、Parallels 或其他虚拟机版本。如果你使用 Mac,这一版暂时还不能完成 Origin 全流程;当前请换用 Windows 实体电脑,后续支持情况以 release 说明为准。

[!IMPORTANT] 我已按 Apache License 2.0 开源 EditaPlot。当前兼容目标是 Origin/OriginPro 2021–2026b;你不必提前打开它,EditaPlot 会在绘图前自动启动一个专用实例。我不会替你安装或修改 Origin。

一眼看懂

flowchart LR
    A["你的数据CSV / TXT / XLS / XLSX"] --> B["读取表格识别每一列的作用"]
    B --> C["推荐 1–3 种图并给出配色"]
    C --> D["列用途与图形元素清单画 / 辅助 / 保留 / 待确认"]
    D --> E{"还有关键歧义?"}
    E -- 有 --> F["只追问必要信息列义、单位、误差或变换"]
    F --> D
    E -- 没有 --> G["你确认科学目的和最终图形元素"]
    R["可选参考图"] --> S["只提取图形语法与风格不复制数据和文字"]
    S --> G
    G --> H["在专用 Origin 实例中绘图"]
    H --> I["可编辑 OPJUPNG + PDF + TIF"]
    I --> J["反读对象并人工检查"]

当我说“已经画好”时,你会拿到可继续编辑的 Origin 项目和 PNG、PDF、TIF;我还会检查原始数据没有被改动、坐标轴和文字完整、每个文件都能正常打开。

先理解数据,再决定画什么

很多科研表格并不是“每个数字都要画”。我会先把每一列放进下面一种用途,并用大白话请你确认:

用途在图中怎么处理
主要证据作为实测点、主曲线、柱体等核心元素绘制
可见辅助作为背景、拟合线、残差、参考线或物相刻线绘制
仅用于计算或验证保留用于权重、筛选、坐标或布局,不画成曲线
保留但不绘制留在数据映射和可编辑项目中,图上不显示
仍不确定暂停规划,先问清楚用途,不能自动猜成新曲线

确认时你会看到“这是什么数据、哪些列会画、哪些列不会画、会出现哪些图形元素、哪些计算不会自动做”。只要源文件、列映射或理解结果发生变化,这次确认就会失效,需要重新核对。

GSAS / GSAS-II XRD Rietveld 示例

我已为普通 XRD、GSAS-II Powder CSV 和 Publication CSV 加入专门的理解规则。以精修表为例,EditaPlot 可以把 Observed 识别为实测点、Calculated 识别为计算线,并按文件实际提供的内容加入 Background、Difference 和具备明确身份的 Phase 刻线;weightQUseddiff/sigmaAxis-limits 等列会保留为辅助或控制数据,不会被误画成强度曲线。

Publication CSV 中已经带显示位置的 Diff 会按源值直接绘制,不会再次偏移。我也不会自动计算背景、差值、Rwp、χ²,或替你识别物相和峰归属。仓库内提供了 example_gsas_powder.csvexample_gsas_publication.csv,可以先拿它们熟悉格式。

用参考图告诉我“想要这种表达”

你也可以上传一张 PNG、JPEG 或 TIFF 参考图。我会先把它理解为“图形简报”:提取面板、插图、点线柱等图形元素、数据编码和有限的视觉风格,再把适合当前模板且有用户数据支撑的部分列出来请你单独确认。

这条路线不会从像素反推实验数据,不复制参考图中的数值、文字、拟合结果、物相、Logo 或水印,也不会把参考图片塞进 OPJU。它的目标是安全借鉴图形语法,而不是承诺任意图片 1:1 复刻;当前

Description

AI-guided editable scientific figures with Codex and local Origin/OriginPro

Related repositories

Similar projects matched by category, topics, and programming language.

powerycy
Featured
powerycy GitHub avatar

goutoujunshi

An AI relationship coach that provides emotional support, relationship science, and actionable strategies through a multidisciplinary knowledge base.

Python
1,174
makecindy
Featured
makecindy GitHub avatar

cindy

Cindy is an open-source AI agent that runs locally on your machine, integrates multiple AI harnesses and models, and provides memory, skills, and automation to perform real work in your projects and apps.

AI & Machine LearningLarge Language Models
958
l0ng-ai
Featured
l0ng-ai GitHub avatar

tty7

tty7 is a fast, GPU-accelerated terminal workbench with persistent sessions, built-in input tools, SSH support, and coding agent awareness.

Developer ToolsCLI & Terminal
359