Practical overview
When it fits
Use Pythonrouge when a summarization evaluation workflow needs ROUGE scores against reference summaries and can provide the required ROUGE-1.5.5 script and data paths.
Who it is for
- NLP researchers evaluating document summarization systems
- Developers who want ROUGE score comparisons in a Python-based text evaluation workflow
Documented uses
- Compare generated summaries with one or more reference summaries.
- Choose ROUGE metrics and report recall, F-measure, or both.
- Score summaries provided directly as text or organized in files.
Requirements & limits
- Evaluation requires paths to ROUGE-1.5.5.pl and RELEASE-1.5.5/data.
- The README documents both in-memory text inputs and a directory-based format with matching system and reference filenames.
- The MIT license is identified in repository metadata.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install Pythonrouge and prepare its ROUGE paths
Install the wrapper and identify the ROUGE files required for evaluation.
Before you start
- Python environment
- Access to the Pythonrouge Git repository
- ROUGE-1.5.5.pl and the RELEASE-1.5.5/data directory
Steps
Install from the repository with pip, or clone the repository and run python setup.py install.
Set the paths to ROUGE-1.5.5.pl and RELEASE-1.5.5/data when configuring evaluation.
Expected outcome
Pythonrouge is installed, and the required ROUGE script and data paths are identified for evaluation.
Guide limitations
- This guide covers installation and required path preparation; evaluation also requires system summaries and reference summaries.
Prepare summary files for directory-based evaluation
Arrange system and reference summaries in the naming format Pythonrouge documents for file-based scoring.
Before you start
- System-generated summary files
- One or more reference summary files for each system summary
Steps
Name each system summary {NAME}.txt.
Name each reference summary {NAME}.{SUMMARY_ID}.txt, keeping {NAME} consistent with its system file.
Configure Pythonrouge for file-based input and provide the system and reference directory paths.
Expected outcome
System and reference summaries are arranged for ROUGE metric evaluation through Pythonrouge's file-input mode.
Guide limitations
- The README specifies the directory naming pattern and file-input configuration; it does not provide a broader evaluation workflow.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 17 October 2026
- Original website title
- tagucci/pythonrouge · Official homepage title · retained in its source language