natsume-server

Natsume writing assistance system data processor and server source code.

For those that wish to use the Natsume system, please go to the official page:

Developing

Environment

To setup a development environment, start a REPL (with or without cider):

boot dev
boot cider dev

Run go to initiate and start the system.

user=> (go)
:started

By default this creates a web server at http://localhost:3000.

When you make changes to your source files, use reset to reload any modified files and restart the server.

user=> (reset)
:reloading (...)
:started

Prerequisites

natsume-server is written in Clojure, and thus requires a working Java environment to run. All development is done on the IcedTea 8 JVM (OpenJDK), though other versions (7) should work.

A PostgreSQL database is required as well. Currently all testing is done on the 9.5 release.

natsume-server additionally makes use of MeCab and CaboCha.

Free-floating jars

Unfortunately, the following free-floating jars need to be downloaded and installed into a local maven repo:

mvn install:install-file -DgroupId=org.chasen -DartifactId=cabocha -Dpackaging=jar -Dversion=0.69 -Dfile=/usr/share/java/cabocha/CaboCha.jar -DgeneratePom=true

Usage

Running with lein

The simplest way to run natsume-server is to use lein in the project directory.

lein run /path/to/corpus/dir/

Corpus processing: Extraction (for word embedding training etc.)

boot run --extract --unit suw --out corpus-file.tsv

Planned Features

Integration with automatic error correction/feedback system Nutmeg

provide an API from Natsume to Nutmeg to ease register (genre) error detection
- at the token level
- at the collocation level
- for both levels include the option of sending back example sentences with query included (this should also be offered as a separate API)
provide fulltext search API (perhaps using Elastic Search (Elastisch in Clojure); maybe even generalize it so that register error detection could be possible)

Integration with reading assistance system Asunaro

add functionality to help read example sentences in Natsume
- tree view or dependency link view in JavaScript (by translating the existing Java program; or from scratch for the dependency view)

Learner dashboard

provide a dashboard displaying everything the user did, using cookies and totally optional (make sure to save old Apache logs!)
- need to record all user input and clicks (an expansion of the users/history tables in Asunaro)
- provide a user login system (OAuth?/standard)
provide review/learning tasks based on users input and level
- help build up vocabulary by using cloze test-like quizzes
- integrate with Natane learner database to train against common errors (per native language)

Roadmap

0.2.0 - initial release

0.3.0 - basic collocation extraction

0.4.0 - token and collocation information offered as API

0.6.0 - pattern matching and transformation rule engine/state machine inspired by CQL/MQL

...

1.0.0 - working release ready to replace current main site

License

Distributed under the Eclipse Public License, the same as Clojure.

borh / natsume-server