Skip to content

Latest commit

 

History

History
63 lines (60 loc) · 3.67 KB

FILES.md

File metadata and controls

63 lines (60 loc) · 3.67 KB
Raw

This list is a WIP/TODO.

  • ./TODO.md: rough roadmap and ticketing-system-in-a-file

  • ./DONE.md: contains closed entries from TODO.md. This has been started quite late, so not all the history is in here.

  • ./modules/*: modules, one per file.

    • ./modules/data.js: Wraps access to various data submodules. We have a few different data types:
      • modules/data/dict/*: dictionaries
      • modules/data/decomp/*: character decomposition tables
      • modules/data/big5/: big5 related encoding table
      • modules/data/book/*: some books which can be studied with zhongmu For each data types, the submodules correspond to individual parsers and parsing utilities.
    • modules/cut.js: relies on data files to identify words in a chunk of Chinese text (there is no spaces between words in Chinese, hence we can't just split on spaces to extract words);
    • modules/db.js: for now, just a JSON/JavaScript dump of the SQL database. Later, we'll perform requests to a backend to get the database.
    • modules/view.js: this is the core of the UI. We use a simple, peculiar JS pattern which avoids us the need for React or any modern framework, which seems to be as expressive: to each (DOM) node type, we associate a "constructor": a function which takes some paramaters, and return a DOM node. Naturally, complex nodes (trees) are built by calling "constructors" within other "constructors". We can then use various mechanism to send data back and forth within the DOM (e.g. keep pointers to specific nodes, event bubbling).
    • modules/view/*: specific, per-page UI code;
    • modules/dom.js: DOM manipulation utilities;
    • modules/move.js: state machine handling user motion through a piece of Chinese text;
    • modules/node.js: Node.js related code. We rely on Node.js to run a few JS scripts, mainly to prepare data.
    • ./modules/utils.js: small, independent utilities;
    • ./modules/bookmark.js: contains code to dump/load some of a page's state (e.g. user location within a piece of Chinese) to/from the URL's hash;
    • ./modules/links.js: tools and ad-hoc databases of available external links (e.g. external dictionaries);
    • ./modules/user.js: user preferences management (not the UI code, just the internal stuff);
    • ./modules/assert.js: dumb wrapper to provide assertions;
    • ./modules/stack.js: on the UI, we can have a stack of decomposed words, allowing to switch back and forth between different words: this module implements that stack;
    • ./modules/log.js: for now just a dumb debugging wrapper;
    • ./modules/main.js: bootstrap the creation of the UI;
    • ./modules/enums.js: automatically generated enums from typescript data;
    • ./modules/classes.js: names corresponding to CSS classes;
    • ./modules/config.js: automatically generated; allows to set a deployment version (for (broken btw) caching) and the HTTP path root of the deployment (e.g. GET /$x vs. GET /zm/$x)

  • ./bin/

    • ./bin/mkabout.js: builds the about.html page (the goal is to use the DB to properly reference and credits all data sources);
    • ./bin/mkshuowen.js: builds a markdown version of the Shuowen
    • ./bin/mksite.js: ad-hoc static site builder (~temporary);
    • ./bin/check-data.js: ensure that the data files can be loaded successfully.
    • ./bin/mktestsjs.sh: builds tests.js;

  • ./tests/*: all our tests; tests/$x.js contains the tests for the module ./modules/$x.js. Tests are automatically ran on import of a test file.