https://github.com/codelibs/fess-webapp-multimodal
https://github.com/codelibs/fess-webapp-multimodal
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/codelibs/fess-webapp-multimodal
- Owner: codelibs
- License: apache-2.0
- Created: 2024-06-20T13:15:44.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-06T01:31:45.000Z (4 months ago)
- Last Synced: 2025-03-06T01:31:59.998Z (4 months ago)
- Language: Java
- Size: 211 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Fess for Multimodal Search
[](https://github.com/codelibs/fess-webapp-multimodal/actions/workflows/maven.yml)
## Overview
This is a multimodal-search plugin for Fess, enabling the crawling and indexing of various media formats such as text, images, audio, and video (support for audio and video will be available in future updates).
## Download
See [Maven Repository](https://repo1.maven.org/maven2/org/codelibs/fess/fess-webapp-multimodal/).
## Installation
See [Plugin](https://fess.codelibs.org/14.15/admin/plugin-guide.html) of Administration guide.
## Usage
After installing the plugin, follow these steps to configure and use it:
1. **Start Fess**: Launch Fess and log in to the administration console.
2. **Configure System Properties**: Add the following properties to the general settings under system properties:
```
fess.multimodal.content.field=content_vector
fess.multimodal.content.dimension=512
fess.multimodal.content.method=hnsw
fess.multimodal.content.engine=lucene
fess.multimodal.content.space_type=cosinesimil
fess.multimodal.min_score=0.5
```
3. **Update Settings**: Navigate to the scheduler page and execu te Config Reloader.
4. **Re-indexing**: Navigate to the maintenance page and execute re-indexing.
5. **Setup CLIP as Service**: For image embedding, use CLIP as Service. Start the CLIP API using Docker:
```sh
git clone https://github.com/codelibs/fess-webapp-multimodal.git
cd fess-webapp-multimodal/docker
docker compose up -d
```
This will make the CLIP API accessible at `localhost:51000`.
6. **Crawl Directories**: Crawl directories containing image files.
7. **Test Data**: If you need test data, you can download the Open Images Dataset. For example:
```sh
pip install fiftyone
fiftyone zoo datasets load open-images-v7 --split validation --kwargs max_samples=1000 -d fiftyone
```
This will download 1000 images into the `fiftyone` directory.
## Contributing
We welcome contributions to enhance the functionality of this plugin. Please fork the repository and submit pull requests.
## License
This project is licensed under the Apache License 2.0 - see the [LICENSE](LICENSE) file for details.