🚀📢 BookNLP-fr is here!

A major step for french computational literary studies by the Lattice Lab team—Frédérique, Olga, Martial, Clément, @tpoibeau.bsky.social and I!

🌍 The French expansion of @dbamman.bsky.social’s BookNLP project

📖 doi.org/10.48694/jcl...

#ComputationalHumanities #NLP

BookNLP-fr, the French Versant of BookNLP. A Tailored Pipeline for 19th and 20th Century French Literature doi.org/10.48694/jcl... #CHR2024

Since the #DH community has grown a lot lately: Announcing the official release of Brahe, a fully multilingual analytic LLM for analyzing literature works. Brahe works like BookNLP and outputs as much as 20 annotations per literary excerpts. huggingface.co/Pclanglais/B...

This week, we announce another article from JCLS 3 (1): Mélanie-Becquet et al. "BookNLP-fr, the French Versant of BookNLP. A Tailored Pipeline for 19th and 20th Century French Literature" (10.48694/jcls.3924). Check it out at: jcls.io/issue/109/in... #CLS #JCLS #DigitalHumanities #CCLS24 #BookNLP

My girlfriend's mom has written her life story (107,203 words), and I'm going to weave some of it as data for a Christmas present. She has low vision so I'm trying to lean on yarn thickness/texture rather than colors. But first: @dbamman.bsky.social BookNLP for extracting names/coreference. #DHmakes

I am so so excited for this! We were just thinking of a #DataSittersClub on BookNLP; maybe @readywriting.bsky.social and I should start that one as a Multilingual Mystery with our French translations!

Then @jbarre.bsky.social , Olga Semnick, Frédérique Mélanie, and @tpoibeau.bsky.social led us through a series of questions about gender and adventure in French fiction, using a range of different NLP methods.

Mélanie-Becquet et al. introduce BookNLP-fr, a French adaptation of the #BookNLP pipeline by @dbamman.bsky.social et al., enhancing #genre classification with interpretable features and expanding tools for #distantreading in #CLS. doi.org/10.48694/jcl... #CCLS24 #NLP #DigitalHumanities #DH

#DH2025: Xiaoyan Yang & @fpianz.bsky.social present a #multilingual #CLS pipeline comparing #English & #Chinese #fanfiction. Using #BookNLP, #ARMCC & #LLMs, they address #CoreferenceResolution, dialogue speaker & character feature extraction 🔥 hot topics at this year's conference. 📚🌏

What’s our paper about? We explore how semantic verb classes can be used for thematic analysis of German literary texts. While tools like BookNLP exist for English, a comparable solution for German is still missing. (2/3)

Guessing that running BookNLP is part of the fun, but if you want to start w/output files, your friends at HTRC have run ~200k English-lang fiction vols through the pipeline already, and released all non-expressive data: htrc.atlassian.net/wiki/spaces/.... Unsure if any BSC vols are included though!

Had fun with bookNLP this evening

Maybe a question for @tedunderwood.com ? I'm planning to benchmark this approach against BookNLP, but am working on a broader text extraction / analysis tool.

I can't remember if Japanese was one of the language @dbamman.bsky.social was working on for his Multilingual BookNLP annotation project? If so, he might have some useful data here.

Has anyone doing computational social science used bookNLP for (1) non-fiction or (2) non-English language works that wouldn't mind letting me know how it went?

I have a corpus of books I'd like to work on, but I'm not sure if there's a better way than trying to hack something onto bookNLP.

We build a 🕵️‍♂️ Detective Detector 🕵️‍♂️ based on manual annotation of major French detectives across the corpus.
We used
• character/Coref resolution from the Propp-fr pipeline (child of BookNLP-fr)
• Contextual embeddings of character attributes using CamemBERT

Definitely going to need to tackle BookNLP for a future #DataSittersClub, including what to expect in terms of data cleanup/normalization to get good results. A personal narrative that references family members named (in part) after one another is hard mode, w/ false clustering across generations. 🫠

literature. We have created a system specifically for the French language, based on BookNLP-fr (the French counterpart of BookNLP) and NetworkX (a Python package for the manipulation and visualization of complex networks). This allows us to establish [3/6 of https://arxiv.org/abs/2503.13449v1]

Sigh of relief that we are talking about the technical elements too - including BERT, BookNLP, prediction+++
Can it go harder? It's not NeurIPs or FAccT. Still interesting I'll say