AthDGC, system demonstration

Open PROIEL-style treebank for the entire Greek language

Nikolaos Lavidas

NKUA · Athens Digital Glossa Chronos Research Network

June 30, 2026

AthDGC

Open PROIEL (the dependency-treebank (a collection of sentences whose grammatical structure has been analysed and stored) standard for early Indo-European languages, developed at Oslo)-style dependency treebank, Homer to Modern Greek - with NT verse-level cross-alignment to Latin / Gothic / Old Church Slavonic.

20-minute LREC-COLING system demo · Athens-PROIEL

Demo agenda (20 min)

  1. 3 min, Overview + funding context
  2. 6 min, Live browse of the showcase: PROIEL tree + arg-structure cards
  3. 5 min, Live cross-alignment query (Neo4j (a graph database that stores data as nodes and connections rather than as tables)) across Greek/Latin/Gothic
  4. 3 min, Editor workflow: flag-for-review toolbar + corpus-side fix workflow
  5. 3 min, Q&A

Live demo step 1, browse

Open https://athdgc.github.io in the browser.

Select Classical period tab. Click into a Plato tlg0059.tlg022 sample.

Observe:

  • Greek sentence in Cardo / Old Standard TT
  • PROIEL dependency tree (root at top, daughters indented)
  • Per-verb argument-structure card (subj / obj / oblique / voice / aspect)
  • Raw PROIEL XML (collapsible)
  • Mark-for-review toolbar (state persisted to localStorage)

Live demo step 2, cross-alignment

Open the public Neo4j endpoint (https://athdgc.github.io/graph).

Run:

MATCH (gk:Token {language:"grc", upos:"VERB"})-[:TRANSLATED_AS]->(la:Token {language:"lat"})
WHERE gk.feats CONTAINS "Voice=Act" AND gk.feats CONTAINS "Aspect=Perf"
  AND la.feats CONTAINS "Voice=Pass"
RETURN gk.form, gk.lemma, la.form, la.lemma LIMIT 25

Returns Greek aorist active verbs whose Vulgate counterpart is passive.

Live demo step 3, editor workflow

In the showcase: click Mark for review on a problematic sample.

The flag is stored in localStorage and persists across reloads. Export current review state via the console:

athdgcExportReview()

This JSON (a plain-text data format for structured records) is then handed to the corpus-side fix_corpus_data.py which applies bulk corrections on ARIS.

Architecture

Layer Implementation
Discovery 43_discover_sources.py - daily cron on ARIS
Filtering 44_filter_candidates.py - Greek-script + apparatus filters
OCR (Optical Character Recognition, the conversion of scanned page images into searchable text) / Conversion 45_ocr_candidates.sbatch + 46_to_proiel_xml.py
Annotation 47_annotate_inbox.sbatch - Stanza (Stanford’s open-source Python workflow that automatically tags, lemmatises, and parses sentences) on A100 GPU
Merge 48_merge_xml_to_corpus.py - JSONL (a plain-text data format that stores one record per line) partitioning
Corpus fix fix_corpus_data.py - TLG author override + Stanza error fixes
Showcase build 51_build_showcase_site.py - regenerates HTML from JSONL
Deploy deploy_athdgc_v2.ps1 - scp + git push

Reuse

  • Per-period JSONL partitions for fine-tuning Greek diachronic transformers
  • PROIEL XML 2.0 (the file format that stores each sentence as a tree of word-by-word grammatical relations, developed at Oslo) export compatible with Oslo PROIEL tools
  • Neo4j alignment graph (a database that records which token in language A corresponds to which token in language B) for cross-linguistic queries
  • Public Hugging Face (a public hosting platform for machine-learning models and datasets) dataset (forthcoming)

Open + cite

  • Source code: github.com/AthDGC/Diachronic-Linguistics-Platform (Apache-2.0 (a permissive open-source licence widely used for software))
  • Dataset: 10.5281/zenodo.20439182 (CC-BY-4.0 (the Creative Commons Attribution 4.0 licence, which permits reuse with credit))

Lavidas, N., Nikiforidou, K., Haug, D., Kulikov, L., Geka, V., Symeonidis, V., Michalareas, T., Chionidi, S., Tsiropina, A., Plakoutsi, E., and Argyropoulos, E. (2026). AthDGC: Athens Digital Glossa Chronos. Zenodo.

Thanks

Funded by HFRI (Project No. 20577) · Greece 2.0 NRRP Compute: GRNET ARIS (the Greek national high-performance computing cluster, run by GRNET) (pa260305)

athdgc.github.io 10.5281/zenodo.20439182

Funding

Funded by the Hellenic Foundation for Research and Innovation (HFRI) under the 3rd Call for HFRI Research Projects to support Post-Doctoral Researchers, Project No. 20577; with complementary support from the Greece 2.0 National Recovery and Resilience Plan. Compute supplied by GRNET ARIS (Greek national HPC (high-performance computing, i.e. a cluster of fast machines used for heavy computation)), allocation pa260305.