forked from
standard.site/standard.site
Standard.site landing page built in Next.js
Add MDX-to-markdown parser with table and component transforms
+269
-2
+269
-2
app/api/docs/markdown/[...slug]/route.ts
+269
-2
app/api/docs/markdown/[...slug]/route.ts
···
1
1
import { NextResponse } from 'next/server'
2
2
import { readFile } from 'node:fs/promises'
3
-
import { join, resolve } from 'node:path'
3
+
import { resolve } from 'node:path'
4
4
5
5
// Map of slugs to their file paths
6
6
const docsFiles: Record<string, string> = {
···
36
36
}
37
37
38
38
const filePath = resolve(process.cwd(), 'content', 'docs', fileName)
39
-
const markdown = await readFile(filePath, 'utf-8')
39
+
const raw = await readFile(filePath, 'utf-8')
40
+
const markdown = await mdxToMarkdown(raw)
40
41
41
42
return new NextResponse(markdown, {
42
43
headers: {
···
48
49
return new NextResponse('Internal server error', { status: 500 })
49
50
}
50
51
}
52
+
53
+
async function mdxToMarkdown(content: string): Promise<string> {
54
+
let result = content
55
+
56
+
// Remove import lines
57
+
result = result.replace(/^import\s+.*$/gm, '')
58
+
59
+
// Replace <StandardSite /> with plain text
60
+
result = result.replace(/<StandardSite\s*\/>/g, 'Standard.site')
61
+
62
+
// Replace <LinkCard> components with markdown links, fetching OG titles
63
+
const linkCardRegex = /<LinkCard\s+([\s\S]*?)\/>/g
64
+
const linkCards: { match: string; url: string; title: string }[] = []
65
+
66
+
for (const m of content.matchAll(linkCardRegex)) {
67
+
const attrs = m[1]
68
+
const url = attrs.match(/url=["']([^"']*)["']/)?.[1] ?? ''
69
+
const title = attrs.match(/title=["']([^"']*)["']/)?.[1] ?? ''
70
+
linkCards.push({ match: m[0], url, title })
71
+
}
72
+
73
+
// Fetch OG titles in parallel for cards without explicit titles
74
+
const titles = await Promise.all(
75
+
linkCards.map(async (card) => {
76
+
if (card.title) return card.title
77
+
return await fetchOgTitle(card.url)
78
+
})
79
+
)
80
+
81
+
for (let i = 0; i < linkCards.length; i++) {
82
+
result = result.replace(linkCards[i].match, `[${titles[i]}](${linkCards[i].url})`)
83
+
}
84
+
85
+
// Replace <Table> components with markdown tables
86
+
// The closing /> is always on its own line, so we anchor to that to avoid
87
+
// matching /> inside JSX fragments like </>
88
+
result = result.replace(/<Table\s+([\s\S]*?)\n\/>/g, (_match, body) => {
89
+
return parseTable(body)
90
+
})
91
+
92
+
// Clean up extra blank lines
93
+
result = result.replace(/\n{3,}/g, '\n\n')
94
+
95
+
return result.trim() + '\n'
96
+
}
97
+
98
+
async function fetchOgTitle(url: string): Promise<string> {
99
+
try {
100
+
const res = await fetch(url, {
101
+
next: { revalidate: 86400 },
102
+
headers: {
103
+
'User-Agent': 'Mozilla/5.0 (compatible; Standard.site/1.0; +https://standard.site)'
104
+
},
105
+
signal: AbortSignal.timeout(5000),
106
+
})
107
+
108
+
if (!res.ok) return url
109
+
110
+
const html = await res.text()
111
+
const ogTitle =
112
+
html.match(/<meta[^>]*property=["']og:title["'][^>]*content=["']([^"']*)["']/i)?.[1] ||
113
+
html.match(/<meta[^>]*content=["']([^"']*)["'][^>]*property=["']og:title["']/i)?.[1] ||
114
+
html.match(/<title[^>]*>([^<]*)<\/title>/i)?.[1]
115
+
116
+
return ogTitle?.trim() || url
117
+
} catch {
118
+
return url
119
+
}
120
+
}
121
+
122
+
function parseTable(body: string): string {
123
+
const headersMatch = body.match(/headers=\{(\[.*?\])}/)
124
+
125
+
const rowsMatch = body.match(/rows=\{(\[[\s\S]*)\}\s*$/)
126
+
127
+
if (!headersMatch || !rowsMatch) return body
128
+
129
+
let headers: string[]
130
+
try {
131
+
// Convert single quotes to double quotes for JSON parsing
132
+
headers = JSON.parse(headersMatch[1].replace(/'/g, '"'))
133
+
} catch {
134
+
return body
135
+
}
136
+
137
+
const rows = parseRows(rowsMatch[1])
138
+
139
+
// Build markdown table
140
+
const headerRow = '| ' + headers.join(' | ') + ' |'
141
+
const separator = '| ' + headers.map(() => '---').join(' | ') + ' |'
142
+
const dataRows = rows.map(row =>
143
+
'| ' + row.map(cell => cell.replace(/\|/g, '\\|')).join(' | ') + ' |'
144
+
)
145
+
146
+
return [headerRow, separator, ...dataRows].join('\n')
147
+
}
148
+
149
+
function parseRows(raw: string): string[][] {
150
+
const rows: string[][] = []
151
+
152
+
// Match each row array: ['col1', 'col2', ...]
153
+
// Rows can contain strings or JSX expressions (<>...</>, <code>...</code>, etc.)
154
+
const rowRegex = /\[([^\]]*(?:\[[^\]]*\])*[^\]]*(?:<[^>]*>[^<]*<\/[^>]*>)*[^\]]*)\]/g
155
+
156
+
// Strip outer brackets
157
+
const inner = raw.trim().replace(/^\[/, '').replace(/\]$/, '')
158
+
159
+
// Split into individual row arrays by matching balanced brackets
160
+
const rowStrings = splitRows(inner)
161
+
162
+
for (const rowStr of rowStrings) {
163
+
const cells = splitCells(rowStr)
164
+
if (cells.length > 0) {
165
+
rows.push(cells.map(cell => jsxToPlainText(cell.trim())))
166
+
}
167
+
}
168
+
169
+
return rows
170
+
}
171
+
172
+
function splitRows(content: string): string[] {
173
+
const rows: string[] = []
174
+
let depth = 0
175
+
let current = ''
176
+
let inString = false
177
+
let stringChar = ''
178
+
let inJsx = false
179
+
180
+
for (let i = 0; i < content.length; i++) {
181
+
const char = content[i]
182
+
const prev = content[i - 1]
183
+
184
+
if (inString) {
185
+
current += char
186
+
if (char === stringChar && prev !== '\\') {
187
+
inString = false
188
+
}
189
+
continue
190
+
}
191
+
192
+
// Track JSX fragments: <> and </>
193
+
if (char === '<' && content[i + 1] === '>') {
194
+
inJsx = true
195
+
current += char
196
+
continue
197
+
}
198
+
if (char === '<' && content[i + 1] === '/' && content[i + 2] === '>') {
199
+
inJsx = false
200
+
current += char
201
+
continue
202
+
}
203
+
204
+
// Only treat quotes as string delimiters outside JSX content
205
+
if (!inJsx && (char === "'" || char === '"' || char === '`')) {
206
+
inString = true
207
+
stringChar = char
208
+
current += char
209
+
continue
210
+
}
211
+
212
+
if (char === '[') {
213
+
if (depth > 0) current += char
214
+
depth++
215
+
} else if (char === ']') {
216
+
depth--
217
+
if (depth === 0) {
218
+
rows.push(current.trim())
219
+
current = ''
220
+
} else {
221
+
current += char
222
+
}
223
+
} else if (depth > 0) {
224
+
current += char
225
+
}
226
+
}
227
+
228
+
return rows
229
+
}
230
+
231
+
function splitCells(row: string): string[] {
232
+
const cells: string[] = []
233
+
let current = ''
234
+
let depth = 0
235
+
let inString = false
236
+
let stringChar = ''
237
+
let angleBracketDepth = 0
238
+
let inJsx = false
239
+
240
+
for (let i = 0; i < row.length; i++) {
241
+
const char = row[i]
242
+
const prev = row[i - 1]
243
+
244
+
if (inString) {
245
+
current += char
246
+
if (char === stringChar && prev !== '\\') {
247
+
inString = false
248
+
}
249
+
continue
250
+
}
251
+
252
+
// Track JSX fragments: <> and </>
253
+
if (char === '<' && row[i + 1] === '>') {
254
+
inJsx = true
255
+
angleBracketDepth++
256
+
current += char
257
+
continue
258
+
}
259
+
if (char === '<' && row[i + 1] === '/' && row[i + 2] === '>') {
260
+
inJsx = false
261
+
current += char
262
+
continue
263
+
}
264
+
265
+
// Only treat quotes as string delimiters outside JSX content
266
+
if (!inJsx && (char === "'" || char === '"' || char === '`')) {
267
+
inString = true
268
+
stringChar = char
269
+
current += char
270
+
continue
271
+
}
272
+
273
+
if (char === '{') { depth++; current += char; continue }
274
+
if (char === '}') { depth--; current += char; continue }
275
+
if (char === '<') { angleBracketDepth++; current += char; continue }
276
+
if (char === '>') { angleBracketDepth--; current += char; continue }
277
+
278
+
if (char === ',' && depth === 0 && angleBracketDepth === 0) {
279
+
cells.push(current.trim())
280
+
current = ''
281
+
continue
282
+
}
283
+
284
+
current += char
285
+
}
286
+
287
+
if (current.trim()) cells.push(current.trim())
288
+
289
+
return cells
290
+
}
291
+
292
+
function jsxToPlainText(cell: string): string {
293
+
// Remove wrapping quotes from plain strings
294
+
if (/^['"].*['"]$/.test(cell)) {
295
+
return cell.slice(1, -1)
296
+
}
297
+
298
+
// Handle JSX expressions wrapped in {<>...</>} or <>...</>
299
+
let text = cell
300
+
301
+
// Strip outer JSX expression wrapper: {<>...</>}
302
+
text = text.replace(/^\{?<>/, '').replace(/<\/>\}?$/, '')
303
+
304
+
// Convert <code>...</code> to backtick-wrapped text
305
+
text = text.replace(/<code>([^<]*)<\/code>/g, '`$1`')
306
+
307
+
// Convert <a href="...">...</a> to markdown links
308
+
text = text.replace(/<a\s+href=["']([^"']*)["'][^>]*>([^<]*)<\/a>/g, '[$2]($1)')
309
+
310
+
// Strip any remaining JSX tags
311
+
text = text.replace(/<[^>]+>/g, '')
312
+
313
+
// Clean up whitespace
314
+
text = text.replace(/\s+/g, ' ').trim()
315
+
316
+
return text
317
+
}