Civitai Companion: Streamlining AI Image Workflow with Automated Metadata Extraction and Resource Management

Civitai Companion is a powerful utility designed to streamline the process of working with AI-generated images from Civitai.com. This tool addresses the needs of AI artists and enthusiasts by automating the extraction of prompt metadata, downloading necessary resources, and formatting information for use in various AI image generation tools. Originally conceived as a companion for Dream Factory, Civitai Companion has evolved to support multiple platforms, including Automatic1111 and others, through its flexible template system.

Civitai Companion offers a robust set of features to enhance your AI image workflow: 1. Batch Metadata Extraction: Efficiently extracts prompts and associated metadata from Civitai images and Dream Factory creations. 2. Automatic Resource Downloads: Downloads models, LoRAs, and embeddings required for image generation, with customizable options for file types and sizes. 3. Customizable Output Formatting: Uses a template-driven system to output extracted information in user-specified formats. 4. LoRA Path Optimization: Automatically modifies LoRA references to match your local path structure. 5. Metadata Value Adjustment: Allows setting min/max values for certain prompt metadata, ensuring compatibility with your setup. 6. Content Filtering: Offers options to remove unwanted words or LoRAs from prompts automatically. 7. Image Dimension Optimization: Resizes images to the closest 'official' resolution for the base model while preserving aspect ratio. 8. Sampler Name Translation: Converts unsupported sampler names to compatible alternatives for tools like Automatic1111.

Setting up Civitai Companion is straightforward: 1. Requirements: A working Python environment (tested on 3.10). 2. Installation: - Clone the repository: `git clone https://github.com/rbbrdckybk/civitai-companion` - Navigate to the directory: `cd civitai-companion` - Install required packages: `pip install --no-input requests tqdm pillow` If you're already using Dream Factory, you may skip the package installation step.

To use Civitai Companion: 1. Run the script: `python civitai_reader.py --config_file config-example.txt` 2. The tool will scan sample images, extract metadata, download referenced LoRAs, and create an output file. 3. For full functionality, obtain a Civitai API key and add it to your config file. 4. Customize settings in the config file to suit your needs. The tool will process images, download resources, and generate output files based on your configuration and chosen template.

Civitai Companion uses a flexible template system for formatting output. Key tokens include: - [PROMPT]: The positive prompt used for the image - [NEG_PROMPT]: The negative prompt - [MODEL]: The model used - [SEED], [SAMPLER], [CLIP_SKIP], [WIDTH], [HEIGHT], [SCALE], [STEPS]: Various generation parameters - [BASE_MODEL]: The base model (e.g., SDXL 1.0, SD 1.5) - [FILENAME], [FILEPATH]: Original image file information Refer to the template-example.txt file for a complete list of available tokens and formatting options.

To get the most out of Civitai Companion: 1. Use command-line arguments to override config file settings for specific workflows. 2. Check the log.txt file in the logs folder for detailed run information and troubleshooting. 3. Manage resource blacklisting by editing the do_not_download.txt file in the cache folder. 4. Utilize the civitai_version_ids.txt file to find version IDs for specific resources. These advanced features allow for fine-tuned control and customization of your Civitai Companion workflow, enabling you to optimize your AI image generation process efficiently.