An online archive of historical issues of the Daily Californian.
This repository contains tools to process raw microfilm images, generate PDFs and host them using Amazon S3. It also contains the Daily Cal's archival website, also hosted in a separate S3 bucket.
This project is under heavy development. While we continue to add more images, we're working on refining our image processing pipeline and web interface. Specifically, we want to
- Perform automatic image correction on the raw TIF files
- Perform OCR using PyTesseract
- Automatically tag pages using a list of Daily Cal and Berkeley keywords
- Create a full-text search for the entire archive
- Detect articles and advertisements
- Crowd-source corrections to our OCR text
If you're interested in contributing to this project -- perhaps if you're another student newspaper interested in digitizing your archives without spending the hundreds of thousands of dollars required for a private service -- please drop us a line at firstname.lastname@example.org.
This project interacts with S3 to host the processed files and to host the "baked" website using django-bakery. You'll need to set the following environment variables:
ARCHIVE_BUCKET_NAME: The name of the Amazon S3 bucket where the processed files will go
RAW_BUCKET_NAME: The name of the bucket where raw scans will be pulled for processing
AWS_BUCKET_NAME: The name of the bucket where the baked website will go
AWS_S3_REGION_NAME: For example,
You can move images through the processing and archival pipeline with a series of management commands.
updatedatabase: Sync the database of pages and issues to the files already in the archival bucket
downloadrawscans: Download any new items from the raw S3 bucket
processrawscans: Check if the raw scans match our expected name format, and if so, move them to the archived files directory
Publish the website
The site uses django-bakery to generate static pages for all the pages and issues. Eventually, as we add the ability to crowdsource text corrections and tags, this will need to change. For now, though, deploy the site using
python manage.py build and
python manage.py publish.