kallewesterling / defoe

Hi there 👋

I am Research Application Manager at the Alan Turing Institute, where my work revolves around the democratisation of access to digital twin technology. I will help in providing open and reproducible computational and social tools that are freely accessible to the UK research and innovation communities. As a Research Application Manager, my role prioritises the user perspective in the development of research software by researchers and engineers and helps stakeholders understand the outputs from their research in a larger perspective.

I am currently seconded to The Alan Turing Institute's Data/Culture: Building sustainable communities around Arts and Humanities datasets and tools project, where he is focusing on building sustainable communities around Arts and Humanities datasets and tools. My work there addresses particular challenges in research software development, ensuring that tools and data reach and engage audiences beyond the original project team. By fostering collaboration and accessibility, I aim to create lasting impact and wider adoption of these resources within the broader research community.

I am a performance scholar and historian by training, holding a Ph.D. in Theatre and Performance from The Graduate Center, CUNY. My Digital Humanities dissertation project analyzed networks of itinerant drag performers around New York and along the Eastern Seaboard in the 1930s as part of a collective art form that brought queerness with them to towns and cities around the area. It was awarded a Connect New York grant for its innovative use of digital methods in Theatre and Performance.

In my previous position, I was a Research Software Engineer at the British Library and The Alan Turing Institute on the Living with Machines research project, which combines large-scale digitized historical collections with data science and computational methods. Before that, I was responsible for building nation-wide infrastructure and community for digital humanities across several higher education institutions across the U.S. My conference participation has been centered around topics such as the Digital Humanities, the Public Humanities, and on how to leverage digital technology for effective pedagogy.

I have written on topics ranging from queer indie wrestlers, boylesque and male-identified strippers in neo-burlesque, and the influence of RuPaul’s Drag Race on the livelihood and aesthetics of contemporary New York City drag queens. You can also find me on the Steering Committee of HASTAC, the Humanities, Arts, Science, and Technology Alliance and Collaboratory.

• 📫   Reach me via email or Twitter
• 😄   Pronouns: they/them or he/him

---

defoe's Issues

Create new query structure for zooniverse

Adding logging when running in Notebook

It seems important and easy enough through the command line (turns out that's just a > bash command though)...

Add `max_height` back in after cropping

After chat with @mialondon today, add max_height back in to the script. If file size is too big after cropping:

We can drop max pixel height (for now) but keep resize if image filesize after cropping is greater than 1mb

Speed up image processing

Look into Pillow-SIMD. "The latest versions of Pillow-SIMD with AVX2 are resizing stuff from 15 to 30 times faster than the original PIL."[1]

Ensure that results are not full pages

From chat with @mialondon on Feb 16:

If defoe returns whole pages and not segmented articles then we still need to crop to the article

Extend documentation

Big bug: yaml does not load unsafely anymore

Add DeezyMatch

See issue https://github.com/alan-turing-institute/Living-with-Machines/issues/2861.

For code, https://github.com/Living-with-machines/DeezyMatch

Clean-up of code

• Format all code with black.
• Correct spellings/singulars where plurals should be used.
• Structure imports + remove unnecessary ones.
• Remove references to rosa's home directory.
• Remove unnecessary keywords = [].
• Remove deprecated code.
• Clean up documentation (although this calls for a more extensive issue opened with more)
• Drop the u"" string syntax. (The package requires Python 3.x, where all strings are already unicoded for convenience)

Make highlighting optional

Add highlighting as option in settings.yml file.

Add missing docstrings

Choose whether to use `os.path` or `pathlib.Path`

Currently, the script/queries use both. I think, go with pathlib is a better idea. It is the modern way (since Python 3.4) and since we're already asking for 3.7+ here, there's no reason to stick with the os library.