DocxMatch is a Streamlit app that analyzes the similarity between Word files.
You can access the live tool at https://docxmatch.streamlit.app/
To run this app locally, you will need to install the following libraries:
- streamlit
- matplotlib
- networkx
- python-docx
- scikit-learn
- pandas
To use the app, upload two or more Word files. The app will analyze the similarity between the files and display the results.
When you upload two or more Word files to DocxMatch, the app will analyze their content and display a list of pairs of files that are considered similar based on a predefined similarity threshold. The default threshold is 0.9.
DocxMatch also displays a heatmap, a dendrogram, and a table showing the similarity scores between all pairs of uploaded files.
This allows you to quickly see which files are most similar to each other and how they are related.
DocxMatch can be used in a variety of scenarios where it is necessary to compare the content of multiple Word documents. Some potential use cases include:
- Organizing documents: DocxMatch can be used to group similar documents together, making it easier to organize and manage large collections of files.
- Finding duplicate content: DocxMatch can be used by content creators or editors to find and remove duplicate content from a large collection of documents.
- Detecting plagiarism: DocxMatch can be used by teachers to compare student assignments and detect potential cases of plagiarism.
When using the live tool, the uploaded files are temporarily stored on the server while the app is processing them. Once the analysis is complete and the results are displayed, the files are removed from the server. The speed of the tool in comparing large batches of files is due to the efficient algorithms used by the app:
- The TF-IDF algorithm assigns a weight to each word in the document based on its importance, allowing the app to accurately measure the similarity between different files without having to compare every single word.
- The cosine similarity metric used by the app is a fast and effective way to compare the similarity between pairs of documents.
Copyright (c) 2023 Nuno S. Osório. This project is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.