Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/rb81/ollama-feed-summarizer
A simple Python application that creates summaries of RSS feeds for easy reading. Uses Ollama for cheap and private inference.
https://github.com/rb81/ollama-feed-summarizer
content-summarizer data-management ollama rss-feed
Last synced: 2 months ago
JSON representation
A simple Python application that creates summaries of RSS feeds for easy reading. Uses Ollama for cheap and private inference.
- Host: GitHub
- URL: https://github.com/rb81/ollama-feed-summarizer
- Owner: rb81
- License: mit
- Created: 2024-07-17T14:55:05.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2024-10-16T20:37:27.000Z (3 months ago)
- Last Synced: 2024-10-18T20:00:55.335Z (3 months ago)
- Topics: content-summarizer, data-management, ollama, rss-feed
- Language: Python
- Homepage:
- Size: 529 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Ollama Feed Summarizer
Ollama Feed Summarizer is a Python application that loops through multiple RSS feeds, summarizing articles using a model running locally using [Ollama](https://github.com/ollama/ollama), and stores the summaries in a daily summary file for easy reading.
> Check out [Feed Summarizer](https://github.com/rb81/feed-summarizer) for a more powerful version that uses Claude for summarization, and a GitHub repository for storage.
## Features
- Fetches articles from multiple RSS feeds
- Summarizes articles using AI (Ollama)
- Compiles summaries into a single markdown file
- Handles unavailable feeds and updates the feed list
- Configurable via JSON file
- Outputs summaries with a date-stamped heading
- **NEW**: Converts summaries to audio using Text-to-Speech (TTS) with [OpenedAI Speech](https://github.com/matatonic/openedai-speech.git). (This script assumes that you have a local instance of OpenedAI Speech running.)
## Requirements
- Python 3.7+
- [Ollama](https://ollama.com/) installed and running locally
- [OpenedAI Speech](https://github.com/matatonic/openedai-speech.git) for TTS functionality
## Installation
1. Clone this repository:
```
git clone https://github.com/rb81/ollama-feed-summarizer.git
cd ollama-feed-summarizer
```
2. Install the required Python packages:
```
pip install -r requirements.txt
```
3. Create a `config.json` file in the project directory with the following structure:
```json
{
"feeds_file": "feeds.txt",
"removed_feeds_file": "removed_feeds.txt",
"output_folder": "/path/to/your/output/folder",
"num_articles": 5,
"ollama_model": "llama3.2:3b",
"ollama_ip": "localhost",
"ollama_port": "11434",
"text_to_speech": {
"enabled": true,
"endpoint_url": "http://localhost:8000/v1/audio/speech",
"model": "tts-1",
"voice": "alloy",
"response_format": "mp3",
"speed": 1.0
}
}
```
4. Create a `feeds.txt` file with one RSS feed URL per line.
## Usage
Run the script with:
```
python main.py
```
The script will:
1. Read RSS feeds from `feeds.txt`
2. Fetch and summarize the specified number of articles from each feed
3. Compile summaries into a markdown file in the specified output folder
4. Update `feeds.txt` and `removed_feeds.txt` if any feeds are unavailable or do not contain any content
5. **NEW**: Generate an audio summary of the articles if TTS is enabled
## Output
The script generates a markdown file named `YYYY-MM-DD_feed-summaries.md` in the specified output folder. The file contains:
- A heading with the current date (e.g., "News for Tuesday, January 1, 2024")
- Summaries of articles from the processed feeds
- **NEW**: A link to the audio summary if TTS is enabled
## Configuration
Adjust the `config.json` file to change:
- Input and output file locations
- Number of articles to summarize per feed
- Ollama model and connection details
- **NEW**: TTS settings such as model, voice, and speed
## License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## Transparency Disclaimer
[ai.collaboratedwith.me](https://ai.collaboratedwith.me) in creating this project.