starred/minne
1
0
Fork 0
mirror of https://github.com/perstarkse/minne.git synced 2026-04-25 02:08:30 +02:00
Code Issues Packages Projects Releases 10 Wiki Activity

feat: completed image ingestion with system settings

Browse Source 
styling
This commit is contained in:
Per Stark
2025-06-27 21:40:44 +02:00
parent f1548d18db
commit 43263fa77e
11 changed files with 157 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
ingestion-pipeline/src/utils/image_parsing.rs
View File

@@ -22,17 +22,12 @@ pub async fn extract_text_from_image(
let image_url = format!("data:image/png;base64,{}", base64_image);
let request = CreateChatCompletionRequestArgs::default()
.model(system_settings.processing_model)
.model(system_settings.image_processing_model)
.max_tokens(6400_u32)
.messages([ChatCompletionRequestUserMessageArgs::default()
.content(vec![
ChatCompletionRequestMessageContentPartTextArgs::default()
.text(r#"Analyze this image and respond based on its primary content:
- If the image is mainly text (document, screenshot, sign), transcribe the text verbatim.
- If the image is mainly visual (photograph, art, landscape), provide a concise description of the scene.
- For hybrid images (diagrams, ads), briefly describe the visual, then transcribe the text under a "Text:" heading.
Respond directly with the analysis."#)
.text(system_settings.image_processing_prompt)
.build()?
.into(),
ChatCompletionRequestMessageContentPartImageArgs::default()