How datadrivenhumanities/ai-in-classics works
AI in Classics combines a text-analysis interface and backend with language-specific search and lemmatization tools; the tree also contains a separate Latin OCR application. The principal documented workflow is querying texts and receiving analysis, with chat, sentiment, feedback, model controls, and retraining as additional capabilities. The OCR upload-to-result path is directly evidenced by the sampled frontend and service. Several README-described features have unsampled implementations, so their nodes are included without speculative call edges.
Main entry points: app_functions.py Query handler · Analyzer.tsx Analyzer UI · query.py Query page · db.py Latin database access
Components 25
Analysis application
- Analyzer UI
src/frontend/src/components/Analyzer.tsx - Query page
src/app/query.py - Query handler
src/app/app_functions.py - Chat routes
src/app/routers/chat_router.py - Ollama client
src/app/ollama_client.py - Model registry
src/app/model_registry.py - Sentiment analysis
src/app/routers/sentiment_router.py - Feedback
src/app/routers/feedback_router.py - Model presets
src/app/routers/presets_router.py - Model status
src/app/routers/probing_router.py - Model training
src/app/routers/train_router.py
Language tools
- Greek lemmatization
src/Lemmatizer-GRK-CLTK/GreekCLTKLemmatizer.py - Greek dictionary queries
src/Lemmatizer-GRK/database/query.py - Greek dictionary store
src/Lemmatizer-GRK/database/models.py - Latin lemmatizer API
src/Lemmatizer-LTN/api/routes.py - Latin database access
src/Lemmatizer-LTN/api/db.py - Latin dictionary store
src/Lemmatizer-LTN/api/db.py
Text search and data
- Text collection
src/pipelines/data_collection - Query preprocessing
src/pipelines/preprocessing/preprocess_root_query - Text search indexes
src/pipelines/text_search
Latin OCR
- OCR interface
Codex-Continuus/frontend/src/App.jsx - OCR service
Codex-Continuus/backend/ocr_service/app.py - OCR engines
Other components
- User
- OCR user
Connections 17
- User → Analyzer UI: uses
README.md - Analyzer UI → Query page: queries
- Query page → Query handler: calls
src/app/query.py - Query handler → Model registry: uses
src/app/app_functions.py - Query handler → Ollama client: calls
src/app/app_functions.py - Chat routes → Ollama client: streams via
src/app/routers/chat_router.py - Sentiment analysis → Query handler: uses
src/app/routers/sentiment_router.py - User → Feedback: submits
README.md - User → Model presets: adjusts
README.md - User → Model status: checks
README.md - User → Model training: requests
README.md - Greek dictionary queries → Greek dictionary store: queries
src/Lemmatizer-GRK/database/query.py - Latin lemmatizer API → Latin database access: calls
src/Lemmatizer-LTN/api/routes.py - Latin database access → Latin dictionary store: accesses
src/Lemmatizer-LTN/api/db.py - User → OCR interface: uploads text
Codex-Continuus/frontend/src/App.jsx - OCR interface → OCR service: posts file
Codex-Continuus/frontend/src/App.jsx - OCR service → OCR engines: recognizes text