KO
|
EN
gitlite — search
Search
#python
#java
#python3
#arduino
#golang
#machine-learning
#rust
#html
#flask
#javascript
#seismology
#nodejs
cst
★ 13
Open GitHub ↗
No description available.
Download README (.md)
Explore Similar Repositories
Flutter-Rounded-Design
:
Flutter Design and Development series
reaction-diffusion
:
Reference implementation and experiments for combining reaction-diffusion and tissue growth
python-webinar
:
Webinar slides and notebook
env-sanitize
:
Sanitization and verification of environment variables with typescript support.
nginx-proxy
:
Example using docker-compose with nginx-proxy and acme companion.
// repository documentation
Was this content helpful?
★ 0
(0 ratings)
Select Rating:
★
★
★
★
★
Submit Feedback
Recent Feedback
×
Download README
Do you want to download the
README.md
file for
cst
?
Download (.md)
# Chaṭṭha Saṅgāyana Tipiṭaka (CST) Reader CST Reader is a cross-platform application for reading and searching Pāli texts. The current main branch contains CST 5.0, built on .NET 10 and Avalonia UI — a ground-up rewrite of the Windows-only CST4. **Status: 5.0.0-beta.5.** This is the milestone release: CST 5 now matches or exceeds CST4 in features, apart from interface localization (see [Known Gaps](#known-gaps)). It is also the first release with Windows builds, alongside macOS. CST Reader presents the Tipiṭaka **in Pāli**, rendered in 14 scripts. It does not include translations of the texts; the built-in dictionaries give the meaning of individual words. ## Branch Overview This repository contains multiple branches representing different stages of CST development: - **`main`** (current): CST 5.0 — cross-platform reader on .NET 10 and Avalonia UI - **`cst_4_2`**: CST 4.2 development branch featuring Lucene.NET 4.8 upgrade work (never released, but provided the foundation for the current search system) - **`cst_4_1`**: CST 4.1 — the released Windows version (tagged `v4.1.0.3-2022-04-05`), .NET Framework with WinForms - **`cst_4_0`**: CST 4.0 — previous stable Windows release (tagged `v4.0.0.15-2020-05-07`), also .NET Framework and WinForms The legacy CST4 branches were Windows-only applications requiring Visual Studio 2022, WiX Toolset v3 for installer creation, and the separate tipitaka-xml repository for text data. Their sources are no longer on `main`; see [Legacy CST4 Development](#legacy-cst4-development) for how to read them. One further branch, `experimental/cef-controlrecycling-workarounds`, is a 2025 spike rather than a stage of development. It is contained in `main`'s history and needs no attention. ## Features ### Core Application - **Cross-Platform**: macOS (Apple Silicon and Intel) and Windows (x64 and ARM64) - **IDE-Style Interface**: dock-based layout with resizable panels, tab management, and persistent session state - **Session Restoration**: restores open books, search highlights, window positions, reading positions, the active tool tab, and the last dictionary lookup - **Floating Windows**: float a book by dragging its tab out of the main window, and drag it back to re-dock - **Dark Mode**: across all panels and book content, including color-inverted search highlights - **Native Packaging**: notarized macOS `.dmg` installers, and per-user Windows `setup.exe` installers plus portable zips ### Text Display & Scripts - **Multi-Script Support**: all 14 Pāli scripts for both display and search input (Devanagari, Latin, Bengali, Cyrillic, Gujarati, Gurmukhi, Kannada, Khmer, Malayalam, Myanmar, Sinhala, Telugu, Thai, Tibetan) - **Global and Per-Tab Script Selection**: changing the script re-renders every open book; each tab also remembers its own setting - **Font Management**: per-script UI font system with native font detection and real-time updates - **Script Conversion Quality**: lossless round-trip conversion for 13 of the 14 scripts, verified by a validation framework. The Cyrillic exceptions are inherent to that transliteration scheme — it cannot distinguish certain vowel sequences — not a converter defect. ### Search - **Full-Text Search**: Lucene.NET 4.8+ with position-based indexing across all 217 texts - **Query Types**: exact, phrase (quoted), proximity (all-within-a-window), mixed/multiple-phrase, wildcard (`*`/`?`) and regular-expression — all working across the 14 scripts - **Two-Color Highlighting**: distinct colors for the match anchor vs. the remaining matched words, with occurrence-by-occurrence navigation - **Smart Filtering**: checkbox book filters with live counts and category management - **Incremental Indexing**: only changed files are reprocessed - **XML Updates**: GitHub integration with SHA-based change detection ### Dictionaries - **Several sources in one panel**, with a picker: the two bundled VRI dictionaries — Childers' *A Dictionary of the Pali Language* (1875) and a Pāli-Hindi dictionary — plus the **Digital Pāḷi Dictionary (DPD)** and the **Dictionary of Pāli Proper Names (DPPN)** - **Self-updating assets**: DPD and DPPN download and keep themselves current; this can be disabled - **User-controlled**: choose which dictionaries appear, and in what order - **Attribution**: each dictionary carries its own citation metadata - **Morphology**: DPD-backed resolution from an inflected form to its lemma, with a lemma report (etymology, root, paradigm, frequency) ### Printing Print a whole book, or print the current selection. ### Source Texts (View Source PDF) - **Burmese edition PDFs**: the 1957 and 2010 editions, plus the extra-canonical (Anya) texts, in dockable tabs — rendered via CEF's PDFium, downloaded on demand and cached locally - **Context-Aware Navigation**: opens the PDF at the page matching your position in the rendered book - Each book offers only the editions it actually has ### AI and Agent Access (optional, off by default) An optional loopback HTTP API and **MCP server** let an AI assistant search the corpus, read passages with their apparatus, use the dictionaries, resolve inflected forms to lemmas, and drive the reader's navigation — returning real references rather than recalled text. It binds only to the loopback interface, requires a per-session token, and stays off until enabled in Settings. ### Technical Architecture - **Stack**: .NET 10, Avalonia UI 11.3, ReactiveUI, Dock.Avalonia, dependency injection - **WebView Rendering**: WebViewControl-Avalonia (CEF) for book content and search highlighting - **Testing**: 1,400+ tests covering unit, integration, and performance scenarios - **Logging**: structured Serilog logging across all components ## Known Gaps - **The interface is English only.** CST4 offers 24 interface languages; that work is still ahead and needs both a localization system and the translations themselves. - **Book content fonts** are not yet user-configurable (UI fonts per script are). - **Elevated idle CPU on macOS** (~30%), inherent to Avalonia's macOS event loop and amplified by CEF rather than specific to CST Reader. ## Development Setup Development and testing are primarily on macOS; Windows builds and runs, and is validated per release. ### Prerequisites - .NET 10 SDK - Git access to this repository ### Build & Run ```bash cd src/CST.Avalonia dotnet build # build dotnet run # run ``` ```bash dotnet test src/CST.Avalonia.Tests # full suite dotnet test src/CST.Avalonia.Tests --filter "FullyQualifiedName~CstDockFactoryTests" # one class ``` ### macOS Packaging ```bash cd src/CST.Avalonia ./package-macos.sh arm64 # Apple Silicon ./package-macos.sh x64 # Intel ./notarize-macos.sh arm64 # code sign, notarize, staple ``` Produces self-contained app bundles and DMG installers (requires `brew install create-dmg`). ### Windows Packaging ```powershell cd src\CST.Avalonia .\package-windows.ps1 # x64 (default) .\package-windows.ps1 -Arch arm64 # ARM64 ``` Produces a portable `.zip` and an Inno Setup `setup.exe` (per-user install, no UAC prompt). Requires Inno Setup 6.3+. Beta builds are unsigned, so Windows SmartScreen shows a warning on first run. Both architectures cross-build from a single host; only *testing* requires a machine of the target architecture. See [docs/development/RELEASE_PROCESS.md](docs/development/RELEASE_PROCESS.md). ### Project Structure ``` src/CST.Avalonia/ # Main application (the working directory) ├── ViewModels/ # ReactiveUI ViewModels ├── Views/ # Avalonia XAML views ├── Services/ # Core services, incl. the local API and MCP surface ├── Resources/ # App resources ├── xsl/ # Per-script book stylesheets └── dictionaries/ # Bundled dictionary data src/CST.Avalonia.Tests/ # Test suite src/CST.Core/ # Book catalog, source-PDF mappings, shared contracts src/CST.Lucene/ # Search engine library src/CST.Lexicon/ # Dictionary asset format src/CST.Lemma/ # Lemma / morphology support src/CST.ScriptValidation/ # Round-trip validation harness for the 14 script converters src/CST.CharacterAnalysis/ # Finds characters in the corpus outside the standard Pāli set docs/ # Architecture, features, research, release process ``` The last two are standalone command-line tools, run by hand rather than referenced by the app. `CST.ScriptValidation` converts Devanagari → IPE → target script → IPE and requires the two IPE forms to match, which is the evidence behind the round-trip claim above; it has its own [README](src/CST.ScriptValidation/README.md). `CST.CharacterAnalysis` scans the XML for characters outside the expected Pāli inventory. The legacy CST4 sources are no longer in the working tree — see [Legacy CST4 Development](#legacy-cst4-development). ## Text Data The application uses Pāli text data from the separate [tipitaka-xml](https://github.com/VipassanaTech/tipitaka-xml) repository, downloaded automatically on first run and updated as needed. ## Legacy CST4 Development The WinForms CST4 sources, the Visual Studio solution, the WiX installer and its font payload were removed from `main` in August 2026. They were 40% of the tracked files in the repository and no part of the CST 5 build, so every search and review had to step around them. Nothing is lost — they remain in history, and are tagged for direct access: | Tag | What it holds | |---|---| | **`cst4-final`** | **The final CST4** — tip of `cst_4_1` (2022-04-08). Use this one. | | `cst4-2-final` | The 4.2 line — a partial Lucene.NET 4.8 port and single-word highlighting. **The only CST4 code not reachable from `main`'s history.** | | `cst4-main-final` | The exact files removed from `main`. Provenance for the removal commit, *not* authoritative — see below. | | `cst-maui-final` | The MAUI/Blazor proof of concept (`src/CST.MAUI`) | Read a file without checking anything out: ```bash git show cst4-final:src/Cst4/FormBookDisplay.cs git show cst4-final:src/Cst4/Reference/en/pali-english-dictionary.txt > dict.txt git checkout cst4-final -- src/Cst4 # restore into the working tree if you need to build it ``` **Why `cst4-final` and not the copy that was on `main`.** No work on CST4 was intended after 4.1. The copy carried on `main` nevertheless picked up three files of drift during CST5 work — including a *functional* edit to two transliteration-table keys in `src/CST/Conversion/Latn2Deva.cs` — so it is not the shipped source. `cst4-main-final` records what was deleted; `cst4-final` records what CST4 is. CST4 documentation stays in the tree — [docs/reference/cst4/](docs/reference/cst4/), the [parity checklist](docs/testing/CST4_PARITY_CHECKLIST.md), and the feature specs that cite CST4 behaviour. The `src/Cst4/...` paths they reference resolve through the tags above. The branches those tags sit on are listed under [Branch Overview](#branch-overview); `cst4-final` is the tip of `cst_4_1`. Building CST4 needs Visual Studio 2022 and WiX Toolset v3 — see the README on the branch itself. ## Documentation & Roadmap - **Documentation index:** [docs/README.md](docs/README.md) — architecture, implementation notes, feature specs, research, and the release process. - **Roadmap / planned work:** tracked as [GitHub issues](https://github.com/fsnow/cst/issues) (filter by the `feature` / `enhancement` labels); detailed specs for several features live in [docs/features/planned/](docs/features/planned/). ## License The texts are provided by the Vipassana Research Institute (VRI). See individual text files for specific attribution and licensing information.
