The dataset viewer is not available for this dataset.
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Ask your favourite LLM who killed JFK
Parsing the JFK Assassination Records Release 2025 with OCR to extract text files that can be used for Retrieval Augmented Generation (RAG) in LLMs.
Download text_output.tar.gz (49 MB) from this dataset repository which are the text files extracts via OCR from the JFK pdf files.
Then download for example https://www.nomic.ai/gpt4all and use LocalDocs to import text_output into some high dimensional space and then chat with your favourite LLM model.
Computation time
OCR processing of text_ouput took about 1 days of computation time on my AMD Ryzen Threadripper 1950X. Same for enhancing the PDFs with text layer, see optional step 5. Embedding the dataset into GPT4All-LocalDocs takes another day of computation time on the same CPU hardware, but shortens to less than half an hour on the RTX 3090.
Credits:
Source is from National Archives https://www.archives.gov/research/jfk/release-2025 All scripts below to scrape and extract text data created with vibe programming and help of ChatGPT 4-5
Steps to recreate the dataset:
- Retrieve first PDFs with jdk_scraper.sh All PDFs will be stored in directory jfk_documents
- Then install requirements with requirements.sh
- Parse retrieved PDFs to extract text with extracttextfrompdf.sh using tesseract. If the script hangs from some reason you can restart it several times. This script extract text only in directory text_output.
- Optional: if you want to enhance your PDFs so that they have a text layer and you can highlight text on the PDF, then launch also enhancepdf.sh. Script will use ocrmypdf to add the invisible text layer. Also this script can be relaunched, in case processing hangs the machine.
- Downloads last month
- 30