Skip to content

parse.md

Latest commit

 

History

History
55 lines (39 loc) · 1.37 KB

parse.md

File metadata and controls

55 lines (39 loc) · 1.37 KB
Raw

How ASTR Parse Raw Story Text

Arknights Raw Story Text

The Arknights raw story text are constructed by lines containing the following parts:

[prop(attr1=value1, attr2=value2, ...)] content

Those prop and attr are case-insensitve. Sometimes the prop can be ignored, left only name attribute in it:

[name="..."] content

or ignore the name too:

content

in the case above, ASTR will convert them into:

[name(name="...")] content
[name(name="")] content

So for every story line, the reader funtion in jsonconvert.py will parse them into:

{
    "id": 1,     // the original line index in raw file
    "prop": "...",
    "attributes": {
        "name": "...",  
        "content": "...",
        "attr1": "value1",
        "attr2": "value2"
        // ...
    }
}

Wordcount Rule

To get accurate wordcount, ASTR won't use the size of raw file as wordcount. Instead, ASTR will only count valid strings in the following attributes:

  • all content & text attributes
  • option attribute in decision prop with semicolomn removed

Note that name attribute are shown in ASTR but won't be counted, since the name are mostly duplicated.

For CN&JP&KR servers, the wordcount or character count, are simply the length of strings. While for EN server, the wordcount are computed by counting the size of string array splited by space.