文件列表:

LICENSE (35147, 2017-07-28)
clj/ (0, 2017-07-28)
clj/speech_tagger/ (0, 2017-07-28)
clj/speech_tagger/core.clj (1351, 2017-07-28)
clj/speech_tagger/pos.clj (3179, 2017-07-28)
docs/ (0, 2017-07-28)
docs/usage.png (22146, 2017-07-28)
emacs/ (0, 2017-07-28)
emacs/penn_treebank_tags.json (6303, 2017-07-28)
emacs/speech-tagger.el (24950, 2017-07-28)
project.clj (524, 2017-07-28)
resources/ (0, 2017-07-28)
resources/english-left3words-distsim.tagger (12409329, 2017-07-28)
test.txt (261, 2017-07-28)

speech-tagger ============= [](http://melpa.org/#/speech-tagger)  This is a program in two parts. The [clojure](clj/speech_tagger) section reads json in stdin (see [core.clj](clj/speech_tagger/core.clj)) and returns json characterizing parts of speech in that text using [corenlp](http://nlp.stanford.edu/software/corenlp.shtml). This is used in an [emacs extension](emacs/speech-tagger.el) ([here](https://melpa.org/#/speech-tagger) on [MELPA](https://melpa.org)) which allows you to highlight text and have it colored and tagged according to its part of speech as shown in the image above. # Dependencies Emacs and the `java` command on your PATH. For development, requires [lein](http://leiningen.org/) to build clojure. # Install For users: just install the `speech-tagger` [package from MELPA](https://melpa.org/#/speech-tagger). Downloads a pretty large jar file on first use, and also on updates, so don't be scared if it takes a while to download and looks like it's frozen the first time; it only happens once. For devs: `load` [the el file](emacs/speech-tagger.el), or add it to `load-path` and `(require 'speech-tagger)`. Run `lein uberjar` to produce the standalone jar, then move that to `emacs/speech-tagger.jar`, or alternatively just customize `speech-tagger-jar-path` to point to the compiled jar. (I could make a build system that does this automatically, but eh.) Set `speech-tagger-is-development` to non-nil to avoid re-downloading the jar file. # Usage The interactive functions exported by this extension follow a common protocol: if a region is active, then modify the region; otherwise modify the entire buffer. If a prefix argument is provided, they read in a buffer to modify the entirety of. A given region will be expanded to whitespace boundaries (so if region is around the `l` characters in ` he|ll|o `, the entirety of `|hello|` will be selected). ```emacs-lisp (defun speech-tagger-tag-dwim (pfx) ``` Tag parts of speech in the appropriate region. "dwim" is an abbreviation for "do what I mean;" hopefully what I the developer mean is close enough to what you the user mean. Tagging, as shown in the image above, colors a part of speech and adds a tooltip to it so that if you mouse over or move point over the part of speech, you get a description of the part of speech and example of that part of speech. ```emacs-lisp (defun speech-tagger-clear-tags-dwim (pfx) ``` As above, but clears the region of all such tags. # Customization ```emacs-lisp (defcustom speech-tagger-jar-path ``` Path to folder containing jar file used to tag parts of speech. Normal users won't need to touch this since it's included in the MELPA package, but developers may find it useful. # Utilities ```emacs-lisp (defun speech-tagger-clear-state () ``` Useful in the case that something screws up and you wish to debug. Should revert all lisp code back to the same as when first loaded. This isn't autoloaded, so you'll have to explicitly `(require 'speech-tagger)` or have already used one of the autoloaded functions. ```emacs-lisp (defun speech-tagger-force-refresh-jar () ``` Re-downloads jar file and md5sum. Useful for debugging jar downloads. # Why? Honestly? I was bored. I've taken enough language classes to be able to characterize parts of speech without a computer by my side. Maybe I'll think of a use for nlp integration in text editors at some point in the future. A friend of mine said it might be cool if I was to tag parts of speech in a separate language, then offer a translation of that text. It might be pretty cool to offer that, in addition to a google translation of the text. Tagging parts of speech in all the languages corenlp supports might be useful for teaching? We'll see. # License [GPL 3.0+](./LICENSE)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：