PaperDelta
English
Review how experiment changes affect an existing research paper. PaperDelta connects declared CSV/JSON evidence to numbers, comparisons and figure provenance in LaTeX and optional Word/PDF manuscripts, then shows affected locations in an offline report.
Install with Python 3.11 or newer in a virtual environment:
python -m pip install paperdelta
paperdelta demo --out paperdelta-demo --open
The complete original demo ships in the core package. No model key, GPU, TeX, checkout or plotting dependency is required. The changed scenario intentionally reports outdated numbers and a false claim; the demo creates files but does not apply paper edits. Every run uses a new output directory.
Batch binding, explicit review scopes, traceable exclusions and file watching support repeated review. Reports and CLI prompts support English and Chinese. Verified numeric edits have a preview and recovery journal. Seventeen optional MCP tools are available through python -m pip install 'paperdelta[mcp]'.
Quick start · Workflows · Agent guide · Limits
Optional Word review: install paperdelta[docx] and run paperdelta demo --document docx --out word-demo --open. Paragraphs and ordinary tables use native positions; Word files are read-only and unsupported structures remain explicit.
Checks cover declared evidence and supported LaTeX/Word/PDF structures. They do not certify scientific truth or infer a correct mapping from matching numbers alone. The original implementation is MIT licensed; separately licensed paper evaluation sources are excluded from the Python distributions.
简体中文
检查实验结果变化影响了现有论文的哪些位置。PaperDelta 将明确声明的 CSV/JSON 证据关联到 LaTeX 及可选 Word/PDF 稿件中的数字、比较和图表来源,在离线报告中集中展示待复核内容。
使用 Python 3.11 以上版本,在虚拟环境安装:
python -m pip install paperdelta
paperdelta --lang zh-CN demo --out paperdelta-demo --open
核心包自带完整原创演示,无需模型密钥、GPU、TeX、源码仓库或绘图依赖。变化场景 会有意检出旧数字和失效结论;演示创建文件,不自动修改论文,每次需使用新输出目录。
批量绑定、明确审查范围、可追踪排除和文件监听支持持续使用。报告及命令行支持 中英文;验证后的数值修改可预览并通过事务恢复。安装 python -m pip install 'paperdelta[mcp]' 可使用十七个可选 MCP 工具。
可选 Word 检查:安装 paperdelta[docx] 后运行 paperdelta --lang zh-CN demo --document docx --out word-demo --open。 段落和普通表格使用原生位置;Word 文件保持只读,不支持的结构会明确报告。
检查限于已声明证据和受支持的 LaTeX/Word/PDF 结构,不认证科学正确性,也不以数字相同证明 映射正确。原创实现采用 MIT 许可;单独许可的论文评测源码不包含在 Python 发行包中。
Metadata
Release files for paperdelta 0.5.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| paperdelta-0.5.0.tar.gz | 14.4 MB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| paperdelta-0.5.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 14.7 MB
Release files / paperdelta-0.5.0.tar.gz
| Download URL | paperdelta-0.5.0.tar.gz |
|---|---|
| Size | 14.4 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
07979daac2afd0c897983efa861ae2e69d939b93b0f714b202e0ba68de7e2ee7
|
|
BLAKE2b-256 checksum How to use checksums |
9b647febf66cd07d77791ebf46db4dae4e54f701b66bfe16468b48ce07b01d57
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 3, 2026.
Transparency logRelease files / paperdelta-0.5.0-py3-none-any.whl
| Download URL | paperdelta-0.5.0-py3-none-any.whl |
|---|---|
| Size | 215.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
2bde0c05f4b4b3b10af0550d04239dfc60d8dc67ce60614760fb1069c8d2bcd3
|
|
BLAKE2b-256 checksum How to use checksums |
251fc30999db82c9c11a6805576e6c353ed0a40986238b50d8721aaf90115882
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 3, 2026.
Transparency log