Nim中文教程
我们可以使用 Nim 语言，来提取文章的元数据。
文章的元数据限定在文章开头，且使用 --- 与正文分隔。author 为关键词，flywind 为值。关键词，是我们限定的属性，有以下这些："title", "layout", "id", "author", "datetime", "excerpt", "tags" 。
---
author: flywind
title: it's a test file.
id: 12
tags: id, bae, nae
---
导入的标准库
我们主要使用 parseutils 模块来解析文章的元数据。
import sets, parseutils, strutils, strformat, times
# 关键词列表
let lookupSet = ["title", "layout", "id", "author", "datetime", "excerpt", "tags"].toHashSet
定义元数据的数据结构
元数据就是我们希望提取的属性。
type
MetaData* = tuple
title: string
layout: string
id: string
author: string
dateTime: string
excerpt: string
tags: seq[string]
Article* = object
data*: MetaData
content*: string
title: bool
id: bool
author: bool
dateTime: bool
excerpt: bool
tags: bool
ParseError* = Exception
解析元数据
我们使用 parseutils 按照步骤解析文章。
proc parseHeader*(s: string): Article =
result.dateTime = true
result.data.dateTime = $local(now())
template assign(attr, value: untyped): untyped =
result.attr = true
result.data.attr = value
var
pos: int
key, value: string
let length = s.len
if skip(s, "---", pos) != 3:
raise newException(ParseError, "No `---` in the head")
pos += 3
pos += skipWhitespace(s, pos)
if skipUntil(s, {'\n'}, pos) == 0:
raise newException(ParseError, "can't have words after `---`")
while true:
pos += skipWhitespace(s, pos)
pos += parseUntil(s, key, {':'}, pos) + 1
key = normalize(key)
if key notin lookupSet:
raise newException(KeyError, fmt"key should be in {lookupSet}")
pos += skipWhitespace(s, pos)
pos += parseUntil(s, value, {'\n'}, pos)
case key
of "title":
assign(title, value.strip)
of "id":
assign(id, value.strip)
of "author":
assign(author, value.strip)
of "datetime":
assign(datetime, value.strip)
of "excerpt":
assign(excerpt, value.strip)
of "tags":
result.tags = true
result.data.tags = value.strip.split(", ")
else: discard
pos += skipWhitespace(s, pos)
pos += skipWhile(s, {'\n'}, pos)
if pos >= length or s[pos .. pos + 2] == "---":
pos += 3
break
result.content = s[pos ..< s.len]
proc parseHeader*(s: File): Article =
result = parseHeader(s.readAll())
测试程序
程序只是为了学习的目的，所以不保证正确性与健壮性。
const tmp = """
---
author: flywind
title: it's a test file.
id: 12
tags: id, bae, nae
---
## title
Let's play with Nim lang.
"""
echo parseHeader(tmp)
输出:
(data: (title: "it\'s a test file.", layout: "", id: "12", author: "flywind", dateTime: "2019-12-09T19:39:38+08:00", excerpt: "", tags: @["id", "bae", "nae"]), content: "\n\n## title\nLet\'s play with Nim lang.\n", title: true, id: true, author: true, dateTime: true, excerpt: false, tags: true)
