Here's a blog post with further screenshots of the plugin in action: https://
datasette.cloud/blog/2024/data
sette-extract/
…
@simonw
-
Datasette Extract Plugin: Data Extraction Tool Showcase
By
–
-
AI Tool Extracts Structured Data From Text and Images
By
–
I built a tool for extracting structured data from unstructured text and images and inserting it into a SQLite database table
-
Datasette Extract: GPT-4 Vision plugin for structured data extraction
By
–
Datasette Extract is a new Datasette plugin that uses GPT-4 (and the new GPT-4 Vision) to extract structured data from unstructured text and images and insert it into a SQLite database table. Here's a video demonstrating the plugin:
-
OpenAI Model Aliases Selection Process Explained
By
–
Interesting note here on how OpenAI model aliases are selected
-
GPT-4 Vision for OCR: Practical Implementation Guide
By
–
I say "OCR" but it was actually GPT-4 Vision itself that I was using for that – my prompt was "Run OCR and return all of the text in this image, with newlines where appropriate"
-
Structured Data Extraction with GPT Functions Now Available
By
–
Mainly structured data extraction into tables, as described in this TIL https://
til.simonwillison.net/gpt3/openai-py
thon-functions-data-extraction
… but now turned into a feature as -
Prompt Engineering Strategies for GPT and Claude Models
By
–
I've not tried fine-tuning any models at all – I've been focusing on figuring out how to best prompt the larger models like gpt-3.5-turbo or claude-3-haiku, mainly through figuring out the best examples to give them
-
Vision Models Function Calling Eliminates Redundant OCR Workflows
By
–
Fantastic – I've been looking forward to function calling for vision models Means I can ditch a nasty little hack I was running where I'd pipe input through Vision and tell it to run OCR, then pipe the result through Turbo to get structured data out of it
-
Building Interactive Prototypes with LLM Assistance: Full Process
By
–
This is a great video for watching someone who builds quick interactive prototypes using LLM assistance walk through their full process
-
HTML Email Security: Preventing Data Exfiltration Attacks
By
–
That feels safe enough to me A really sophisticated attack might still be able to exfiltrate private data ("compose an email with my latest password reset link included as white on white text") but you could limit HTML email composition to help prevent that