@first-ch/tools-mcp 0.25.0 → 0.27.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +6 -2
- package/case-convert.mjs +390 -0
- package/csv-json.mjs +865 -0
- package/package.json +12 -3
- package/server.mjs +96 -0
package/csv-json.mjs
ADDED
|
@@ -0,0 +1,865 @@
|
|
|
1
|
+
/* CSV/TSV ⇄ JSON 相互変換(csv_convert)
|
|
2
|
+
* 変換コア(csvParseDelimited / csvParseJson / csvToJson / csvFromJson / csvJsonConvert)は
|
|
3
|
+
* FirstCHTools の site/csv-json/app.js と同一
|
|
4
|
+
* (2箇所ルール: site側が正本・片方を直したらもう片方も同じ内容で反映する)。
|
|
5
|
+
* 両ファイルからコアの開始/終了コメントに挟まれた範囲を sed で切り出し
|
|
6
|
+
* (site側は字下げ2文字を落とす・こちらは行頭の export を落とす)、
|
|
7
|
+
* diff が空になることで同期を機械的に確認できる。
|
|
8
|
+
*
|
|
9
|
+
* 使っているのは String/RegExp/JSON だけなので、ブラウザ版のコードをそのまま持ってこられる。
|
|
10
|
+
* 完全ローカル処理・ネットワーク送信なし。 */
|
|
11
|
+
import { readFile, writeFile } from 'node:fs/promises';
|
|
12
|
+
import { basename } from 'node:path';
|
|
13
|
+
|
|
14
|
+
/* ==================== ここから変換コア(site / MCP で同一) ==================== */
|
|
15
|
+
|
|
16
|
+
// 区切り文字の呼び名 ⇄ 実体
|
|
17
|
+
export const CSV_DELIMS = { comma: ',', tab: '\t', semicolon: ';', pipe: '|' };
|
|
18
|
+
export const CSV_DELIM_NAMES = { ',': 'comma', '\t': 'tab', ';': 'semicolon', '|': 'pipe' };
|
|
19
|
+
|
|
20
|
+
// 数値として読み替えてよい形。先頭の0(0123)や + 記号を弾くのは、
|
|
21
|
+
// 数値にすると表記が変わってしまう(電話番号・郵便番号・伝票番号が壊れる)ため。
|
|
22
|
+
export const CSV_NUM_RE = /^-?(0|[1-9]\d*)(\.\d+)?([eE][+-]?\d+)?$/;
|
|
23
|
+
// 数値に見えるが表記が保てないもの(先頭の0・+付き・桁あふれ)を見つけるための緩い形
|
|
24
|
+
export const CSV_NUMISH_RE = /^[+-]?\d[\d.eE+-]*$/;
|
|
25
|
+
// 表計算ソフトが数式として解釈しはじめる先頭文字(CSVインジェクション)
|
|
26
|
+
export const CSV_FORMULA_RE = /^[=+\-@\t\r]/;
|
|
27
|
+
|
|
28
|
+
/** 先頭20行を見て区切り文字を推定する。引用符の中は数に入れない */
|
|
29
|
+
export function csvDetectDelimiter(text) {
|
|
30
|
+
const head = String(text).replace(/^/, '').split(/\r\n|\n|\r/).slice(0, 20).join('\n');
|
|
31
|
+
const bare = head.replace(/"(?:[^"]|"")*"/g, '');
|
|
32
|
+
const lines = bare.split('\n').filter((l) => l !== '');
|
|
33
|
+
if (!lines.length) return 'comma';
|
|
34
|
+
let best = 'comma';
|
|
35
|
+
let bestScore = 0;
|
|
36
|
+
for (const name of Object.keys(CSV_DELIMS)) {
|
|
37
|
+
const d = CSV_DELIMS[name];
|
|
38
|
+
const counts = lines.map((l) => l.split(d).length - 1);
|
|
39
|
+
if (!counts.length || counts[0] === 0) continue;
|
|
40
|
+
// 「どの行でも同じ個数で並んでいるか」を重く見る(本文にたまたま混ざった記号を選ばない)
|
|
41
|
+
const same = counts.filter((c) => c === counts[0]).length / counts.length;
|
|
42
|
+
const score = counts[0] * (0.2 + same);
|
|
43
|
+
if (score > bestScore) {
|
|
44
|
+
bestScore = score;
|
|
45
|
+
best = name;
|
|
46
|
+
}
|
|
47
|
+
}
|
|
48
|
+
return best;
|
|
49
|
+
}
|
|
50
|
+
|
|
51
|
+
/**
|
|
52
|
+
* 区切りテキストを行×セルへ割る(RFC 4180)。
|
|
53
|
+
* 引用符の中の改行・区切り記号・"" によるエスケープを扱い、
|
|
54
|
+
* 壊れた引用符は文字として読み進めたうえで指摘に残す。
|
|
55
|
+
* @returns {{rows: string[][], quirks: {stray: number[], newlineInCell: number}, error: object|null}}
|
|
56
|
+
*/
|
|
57
|
+
export function csvParseDelimited(text, delim) {
|
|
58
|
+
const src = String(text);
|
|
59
|
+
const rows = [];
|
|
60
|
+
const stray = [];
|
|
61
|
+
let newlineInCell = 0;
|
|
62
|
+
let row = [];
|
|
63
|
+
let field = '';
|
|
64
|
+
let i = 0;
|
|
65
|
+
let line = 1;
|
|
66
|
+
let col = 1;
|
|
67
|
+
let quoted = false;
|
|
68
|
+
let openLine = 1;
|
|
69
|
+
let openCol = 1;
|
|
70
|
+
let fieldStarted = false; // 引用符を開いてよいのは、そのセルの先頭にいるときだけ
|
|
71
|
+
|
|
72
|
+
while (i < src.length) {
|
|
73
|
+
const ch = src[i];
|
|
74
|
+
if (quoted) {
|
|
75
|
+
if (ch === '"') {
|
|
76
|
+
if (src[i + 1] === '"') { field += '"'; i += 2; col += 2; continue; }
|
|
77
|
+
quoted = false;
|
|
78
|
+
i += 1;
|
|
79
|
+
col += 1;
|
|
80
|
+
continue;
|
|
81
|
+
}
|
|
82
|
+
if (ch === '\r' || ch === '\n') {
|
|
83
|
+
// 引用符の中の改行はセルの中身。改行コードはLFへ揃える
|
|
84
|
+
field += '\n';
|
|
85
|
+
i += ch === '\r' && src[i + 1] === '\n' ? 2 : 1;
|
|
86
|
+
line += 1;
|
|
87
|
+
col = 1;
|
|
88
|
+
newlineInCell += 1;
|
|
89
|
+
continue;
|
|
90
|
+
}
|
|
91
|
+
field += ch;
|
|
92
|
+
i += 1;
|
|
93
|
+
col += 1;
|
|
94
|
+
continue;
|
|
95
|
+
}
|
|
96
|
+
if (ch === '"') {
|
|
97
|
+
if (!fieldStarted) {
|
|
98
|
+
quoted = true;
|
|
99
|
+
fieldStarted = true;
|
|
100
|
+
openLine = line;
|
|
101
|
+
openCol = col;
|
|
102
|
+
i += 1;
|
|
103
|
+
col += 1;
|
|
104
|
+
continue;
|
|
105
|
+
}
|
|
106
|
+
// セルの途中や閉じたあとに出た " は文字として扱う(Excelが書いた壊れた行を読めるように)
|
|
107
|
+
if (stray.indexOf(line) === -1) stray.push(line);
|
|
108
|
+
field += ch;
|
|
109
|
+
i += 1;
|
|
110
|
+
col += 1;
|
|
111
|
+
continue;
|
|
112
|
+
}
|
|
113
|
+
if (ch === delim) {
|
|
114
|
+
row.push(field);
|
|
115
|
+
field = '';
|
|
116
|
+
fieldStarted = false;
|
|
117
|
+
i += 1;
|
|
118
|
+
col += 1;
|
|
119
|
+
continue;
|
|
120
|
+
}
|
|
121
|
+
if (ch === '\r' || ch === '\n') {
|
|
122
|
+
row.push(field);
|
|
123
|
+
rows.push(row);
|
|
124
|
+
row = [];
|
|
125
|
+
field = '';
|
|
126
|
+
fieldStarted = false;
|
|
127
|
+
i += ch === '\r' && src[i + 1] === '\n' ? 2 : 1;
|
|
128
|
+
line += 1;
|
|
129
|
+
col = 1;
|
|
130
|
+
continue;
|
|
131
|
+
}
|
|
132
|
+
field += ch;
|
|
133
|
+
fieldStarted = true;
|
|
134
|
+
i += 1;
|
|
135
|
+
col += 1;
|
|
136
|
+
}
|
|
137
|
+
|
|
138
|
+
if (quoted) {
|
|
139
|
+
return { rows: [], quirks: { stray, newlineInCell }, error: { code: 'UNCLOSED_QUOTE', line: openLine, col: openCol, span: 1 } };
|
|
140
|
+
}
|
|
141
|
+
row.push(field);
|
|
142
|
+
rows.push(row);
|
|
143
|
+
// 末尾の改行が作る空の行は落とす
|
|
144
|
+
if (rows.length && rows[rows.length - 1].length === 1 && rows[rows.length - 1][0] === '') rows.pop();
|
|
145
|
+
return { rows, quirks: { stray, newlineInCell }, error: null };
|
|
146
|
+
}
|
|
147
|
+
|
|
148
|
+
/** 列名を経路へ割る。a.b → ['a','b']、a[0].b / a.0.b → ['a',0,'b'] */
|
|
149
|
+
export function csvParsePath(key) {
|
|
150
|
+
const out = [];
|
|
151
|
+
for (const seg of String(key).replace(/\[(\d+)\]/g, '.$1').split('.')) {
|
|
152
|
+
if (seg === '') continue;
|
|
153
|
+
out.push(/^\d+$/.test(seg) ? Number(seg) : seg);
|
|
154
|
+
}
|
|
155
|
+
return out.length ? out : [String(key)];
|
|
156
|
+
}
|
|
157
|
+
|
|
158
|
+
/** 経路に沿って値を置く。数字の区間は配列になる。置けなければ false を返す */
|
|
159
|
+
export function csvSetPath(target, path, value) {
|
|
160
|
+
let node = target;
|
|
161
|
+
for (let i = 0; i < path.length - 1; i += 1) {
|
|
162
|
+
const seg = path[i];
|
|
163
|
+
const nextIsIndex = typeof path[i + 1] === 'number';
|
|
164
|
+
const cur = node[seg];
|
|
165
|
+
if (cur === undefined || cur === null) {
|
|
166
|
+
node[seg] = nextIsIndex ? [] : {};
|
|
167
|
+
} else if (typeof cur !== 'object') {
|
|
168
|
+
return false; // 同じ名前がスカラーで埋まっている(a と a.b が両方ある)
|
|
169
|
+
}
|
|
170
|
+
node = node[seg];
|
|
171
|
+
}
|
|
172
|
+
const last = path[path.length - 1];
|
|
173
|
+
if (node[last] !== undefined && typeof node[last] === 'object' && node[last] !== null) return false;
|
|
174
|
+
node[last] = value;
|
|
175
|
+
return true;
|
|
176
|
+
}
|
|
177
|
+
|
|
178
|
+
/** セル1つを値へ読み替える。読み替えないときは文字列のまま */
|
|
179
|
+
export function csvInferValue(raw, opts) {
|
|
180
|
+
const s = opts.trim ? raw.trim() : raw;
|
|
181
|
+
if (s === '') return opts.emptyNull ? null : '';
|
|
182
|
+
if (!opts.types) return s;
|
|
183
|
+
const lower = s.toLowerCase();
|
|
184
|
+
if (lower === 'true') return true;
|
|
185
|
+
if (lower === 'false') return false;
|
|
186
|
+
if (lower === 'null') return null;
|
|
187
|
+
// 「数値にして文字列へ戻したとき元通りか」で判定する。
|
|
188
|
+
// これだけで 0123・+1・1e999・9007199254740993 のような復元できない値を弾ける。
|
|
189
|
+
if (CSV_NUM_RE.test(s)) {
|
|
190
|
+
const num = Number(s);
|
|
191
|
+
if (Number.isFinite(num) && String(num) === s) return num;
|
|
192
|
+
}
|
|
193
|
+
return s;
|
|
194
|
+
}
|
|
195
|
+
|
|
196
|
+
/** 値の入れ子の深さ */
|
|
197
|
+
export function csvDepth(value) {
|
|
198
|
+
if (value === null || typeof value !== 'object') return 0;
|
|
199
|
+
let max = 0;
|
|
200
|
+
for (const key of Object.keys(value)) {
|
|
201
|
+
const d = csvDepth(value[key]);
|
|
202
|
+
if (d > max) max = d;
|
|
203
|
+
}
|
|
204
|
+
return max + 1;
|
|
205
|
+
}
|
|
206
|
+
|
|
207
|
+
/** JSONの値をCSVのセル1つへ平らにする。nest=true なら a.b / a.0 の列へ割る */
|
|
208
|
+
export function csvFlatten(value, prefix, out, nest, note) {
|
|
209
|
+
if (value === null || value === undefined) {
|
|
210
|
+
out.push([prefix, '']);
|
|
211
|
+
return;
|
|
212
|
+
}
|
|
213
|
+
if (typeof value !== 'object') {
|
|
214
|
+
out.push([prefix, typeof value === 'string' ? value : String(value)]);
|
|
215
|
+
return;
|
|
216
|
+
}
|
|
217
|
+
const keys = Array.isArray(value) ? value.map((_, i) => i) : Object.keys(value);
|
|
218
|
+
if (!nest || !keys.length) {
|
|
219
|
+
// 展開しない指定・空の {} と [] は、意味を落とさないようJSONの文字列として1セルに入れる
|
|
220
|
+
out.push([prefix, JSON.stringify(value)]);
|
|
221
|
+
if (keys.length) note.stringified += 1;
|
|
222
|
+
return;
|
|
223
|
+
}
|
|
224
|
+
if (Array.isArray(value)) note.arrays += 1;
|
|
225
|
+
for (const key of keys) {
|
|
226
|
+
csvFlatten(value[key], prefix === '' ? String(key) : prefix + '.' + key, out, nest, note);
|
|
227
|
+
}
|
|
228
|
+
}
|
|
229
|
+
|
|
230
|
+
/** セルを区切りテキストの1項目へ書く(RFC 4180 の引用規則) */
|
|
231
|
+
export function csvQuoteCell(s, delim, quoteAll) {
|
|
232
|
+
const need = quoteAll
|
|
233
|
+
|| s.indexOf(delim) !== -1
|
|
234
|
+
|| s.indexOf('"') !== -1
|
|
235
|
+
|| /[\r\n]/.test(s)
|
|
236
|
+
|| /^\s|\s$/.test(s);
|
|
237
|
+
return need ? '"' + s.replace(/"/g, '""') + '"' : s;
|
|
238
|
+
}
|
|
239
|
+
|
|
240
|
+
/* ---- JSONの読み取り(行・桁つき。ブラウザ差のあるメッセージに頼らない) ---- */
|
|
241
|
+
|
|
242
|
+
export function csvJsonError(code, line, col, found) {
|
|
243
|
+
return { code, line, col, span: 1, found: found === undefined ? '' : found };
|
|
244
|
+
}
|
|
245
|
+
|
|
246
|
+
/**
|
|
247
|
+
* JSONを読む。仕様どおりの厳密な読み取りで、失敗したときは行と桁を返す。
|
|
248
|
+
* @returns {{ok: true, value: *, duplicates: string[]}|{ok: false, error: object}}
|
|
249
|
+
*/
|
|
250
|
+
export function csvParseJson(text) {
|
|
251
|
+
const src = String(text);
|
|
252
|
+
let i = 0;
|
|
253
|
+
let line = 1;
|
|
254
|
+
let col = 1;
|
|
255
|
+
const duplicates = [];
|
|
256
|
+
|
|
257
|
+
const fail = (code, found) => {
|
|
258
|
+
const err = csvJsonError(code, line, col, found);
|
|
259
|
+
throw err;
|
|
260
|
+
};
|
|
261
|
+
const step = (n) => {
|
|
262
|
+
for (let k = 0; k < n; k += 1) {
|
|
263
|
+
if (src[i] === '\n') { line += 1; col = 1; } else { col += 1; }
|
|
264
|
+
i += 1;
|
|
265
|
+
}
|
|
266
|
+
};
|
|
267
|
+
const ws = () => {
|
|
268
|
+
while (i < src.length && (src[i] === ' ' || src[i] === '\t' || src[i] === '\n' || src[i] === '\r')) step(1);
|
|
269
|
+
};
|
|
270
|
+
const readString = () => {
|
|
271
|
+
step(1); // 開きの "
|
|
272
|
+
let out = '';
|
|
273
|
+
while (true) {
|
|
274
|
+
if (i >= src.length) fail('UNEXPECTED_END');
|
|
275
|
+
const ch = src[i];
|
|
276
|
+
if (ch === '"') { step(1); return out; }
|
|
277
|
+
if (ch === '\n' || ch === '\r') fail('NEWLINE_IN_STRING');
|
|
278
|
+
if (ch === '\\') {
|
|
279
|
+
const esc = src[i + 1];
|
|
280
|
+
if (esc === undefined) fail('UNEXPECTED_END');
|
|
281
|
+
if ('"\\/'.indexOf(esc) !== -1) { out += esc; step(2); continue; }
|
|
282
|
+
if (esc === 'b') { out += '\b'; step(2); continue; }
|
|
283
|
+
if (esc === 'f') { out += '\f'; step(2); continue; }
|
|
284
|
+
if (esc === 'n') { out += '\n'; step(2); continue; }
|
|
285
|
+
if (esc === 'r') { out += '\r'; step(2); continue; }
|
|
286
|
+
if (esc === 't') { out += '\t'; step(2); continue; }
|
|
287
|
+
if (esc === 'u') {
|
|
288
|
+
const hex = src.slice(i + 2, i + 6);
|
|
289
|
+
if (!/^[0-9a-fA-F]{4}$/.test(hex)) fail('BAD_ESCAPE', '\\u' + hex);
|
|
290
|
+
out += String.fromCharCode(parseInt(hex, 16));
|
|
291
|
+
step(6);
|
|
292
|
+
continue;
|
|
293
|
+
}
|
|
294
|
+
fail('BAD_ESCAPE', '\\' + esc);
|
|
295
|
+
}
|
|
296
|
+
out += ch;
|
|
297
|
+
step(1);
|
|
298
|
+
}
|
|
299
|
+
};
|
|
300
|
+
const readValue = () => {
|
|
301
|
+
ws();
|
|
302
|
+
if (i >= src.length) fail('UNEXPECTED_END');
|
|
303
|
+
const ch = src[i];
|
|
304
|
+
if (ch === '{') {
|
|
305
|
+
step(1);
|
|
306
|
+
const obj = {};
|
|
307
|
+
ws();
|
|
308
|
+
if (src[i] === '}') { step(1); return obj; }
|
|
309
|
+
while (true) {
|
|
310
|
+
ws();
|
|
311
|
+
if (src[i] !== '"') fail('EXPECTED_KEY', src[i]);
|
|
312
|
+
const key = readString();
|
|
313
|
+
if (Object.prototype.hasOwnProperty.call(obj, key) && duplicates.indexOf(key) === -1) duplicates.push(key);
|
|
314
|
+
ws();
|
|
315
|
+
if (src[i] !== ':') fail('EXPECTED_COLON', src[i]);
|
|
316
|
+
step(1);
|
|
317
|
+
obj[key] = readValue();
|
|
318
|
+
ws();
|
|
319
|
+
if (src[i] === ',') { step(1); continue; }
|
|
320
|
+
if (src[i] === '}') { step(1); return obj; }
|
|
321
|
+
fail(i >= src.length ? 'UNEXPECTED_END' : 'EXPECTED_COMMA', src[i]);
|
|
322
|
+
}
|
|
323
|
+
}
|
|
324
|
+
if (ch === '[') {
|
|
325
|
+
step(1);
|
|
326
|
+
const arr = [];
|
|
327
|
+
ws();
|
|
328
|
+
if (src[i] === ']') { step(1); return arr; }
|
|
329
|
+
while (true) {
|
|
330
|
+
arr.push(readValue());
|
|
331
|
+
ws();
|
|
332
|
+
if (src[i] === ',') { step(1); continue; }
|
|
333
|
+
if (src[i] === ']') { step(1); return arr; }
|
|
334
|
+
fail(i >= src.length ? 'UNEXPECTED_END' : 'EXPECTED_COMMA', src[i]);
|
|
335
|
+
}
|
|
336
|
+
}
|
|
337
|
+
if (ch === '"') return readString();
|
|
338
|
+
if (src.startsWith('true', i)) { step(4); return true; }
|
|
339
|
+
if (src.startsWith('false', i)) { step(5); return false; }
|
|
340
|
+
if (src.startsWith('null', i)) { step(4); return null; }
|
|
341
|
+
const num = /^-?(0|[1-9]\d*)(\.\d+)?([eE][+-]?\d+)?/.exec(src.slice(i));
|
|
342
|
+
if (num) {
|
|
343
|
+
const value = Number(num[0]);
|
|
344
|
+
step(num[0].length);
|
|
345
|
+
return value;
|
|
346
|
+
}
|
|
347
|
+
fail('UNEXPECTED_CHAR', ch);
|
|
348
|
+
return null;
|
|
349
|
+
};
|
|
350
|
+
|
|
351
|
+
try {
|
|
352
|
+
const value = readValue();
|
|
353
|
+
ws();
|
|
354
|
+
if (i < src.length) fail('TRAILING_TEXT', src[i]);
|
|
355
|
+
return { ok: true, value, duplicates };
|
|
356
|
+
} catch (e) {
|
|
357
|
+
if (e && e.code) return { ok: false, error: e };
|
|
358
|
+
throw e;
|
|
359
|
+
}
|
|
360
|
+
}
|
|
361
|
+
|
|
362
|
+
/** 1行1JSON(JSON Lines / NDJSON)として読めるか試す */
|
|
363
|
+
export function csvParseJsonLines(text) {
|
|
364
|
+
const lines = String(text).split(/\r\n|\n|\r/).map((l) => l.trim()).filter((l) => l !== '');
|
|
365
|
+
if (lines.length < 2 || !lines.every((l) => l.startsWith('{') || l.startsWith('['))) return null;
|
|
366
|
+
const rows = [];
|
|
367
|
+
for (const l of lines) {
|
|
368
|
+
const r = csvParseJson(l);
|
|
369
|
+
if (!r.ok) return null;
|
|
370
|
+
rows.push(r.value);
|
|
371
|
+
}
|
|
372
|
+
return rows;
|
|
373
|
+
}
|
|
374
|
+
|
|
375
|
+
/** エラーの前後2行を抜き出す(画面で ^ を立てるため) */
|
|
376
|
+
export function csvExcerpt(text, error) {
|
|
377
|
+
if (!error) return null;
|
|
378
|
+
const lines = String(text).split(/\r\n|\n|\r/);
|
|
379
|
+
const from = Math.max(1, error.line - 2);
|
|
380
|
+
const to = Math.min(lines.length, error.line + 2);
|
|
381
|
+
const rows = [];
|
|
382
|
+
for (let n = from; n <= to; n += 1) rows.push({ line: n, text: lines[n - 1] || '', error: n === error.line });
|
|
383
|
+
return { rows, col: error.col, span: error.span || 1 };
|
|
384
|
+
}
|
|
385
|
+
|
|
386
|
+
/* ---- 変換の本体 ---- */
|
|
387
|
+
|
|
388
|
+
/**
|
|
389
|
+
* CSV/TSV → JSON
|
|
390
|
+
* @param {string} text
|
|
391
|
+
* @param {object} o delimiter/header/nest/types/trim/emptyNull/skipEmpty/indent
|
|
392
|
+
*/
|
|
393
|
+
export function csvToJson(text, o) {
|
|
394
|
+
const notes = [];
|
|
395
|
+
const raw = String(text);
|
|
396
|
+
const hadBom = raw.charCodeAt(0) === 0xfeff;
|
|
397
|
+
const body = hadBom ? raw.slice(1) : raw;
|
|
398
|
+
if (hadBom) notes.push({ code: 'BOM', level: 'info' });
|
|
399
|
+
if (/\r\n/.test(body)) notes.push({ code: 'CRLF', level: 'info' });
|
|
400
|
+
|
|
401
|
+
const auto = !o.delimiter || o.delimiter === 'auto';
|
|
402
|
+
const name = auto ? csvDetectDelimiter(body) : o.delimiter;
|
|
403
|
+
const delim = CSV_DELIMS[name] || ',';
|
|
404
|
+
if (auto) notes.push({ code: 'DELIM_AUTO', level: 'info', delimiter: name });
|
|
405
|
+
|
|
406
|
+
const parsed = csvParseDelimited(body, delim);
|
|
407
|
+
if (parsed.error) return { ok: false, error: parsed.error, notes, excerpt: csvExcerpt(body, parsed.error) };
|
|
408
|
+
|
|
409
|
+
let rows = parsed.rows;
|
|
410
|
+
if (o.skipEmpty !== false) rows = rows.filter((r) => r.some((c) => c.trim() !== ''));
|
|
411
|
+
if (!rows.length) return { ok: true, empty: true, notes, delimiter: name };
|
|
412
|
+
|
|
413
|
+
if (parsed.quirks.stray.length) {
|
|
414
|
+
notes.push({ code: 'STRAY_QUOTE', level: 'warn', count: parsed.quirks.stray.length, lines: parsed.quirks.stray.slice(0, 6) });
|
|
415
|
+
}
|
|
416
|
+
if (parsed.quirks.newlineInCell) notes.push({ code: 'NEWLINE_IN_CELL', level: 'info', count: parsed.quirks.newlineInCell });
|
|
417
|
+
|
|
418
|
+
const width = Math.max.apply(null, rows.map((r) => r.length));
|
|
419
|
+
let headers = null;
|
|
420
|
+
let dataRows = rows;
|
|
421
|
+
if (o.header !== false) {
|
|
422
|
+
headers = rows[0].map((h) => (o.trim === false ? h : h.trim()));
|
|
423
|
+
dataRows = rows.slice(1);
|
|
424
|
+
const seen = {};
|
|
425
|
+
const dups = [];
|
|
426
|
+
let blank = 0;
|
|
427
|
+
headers = headers.map((h, idx) => {
|
|
428
|
+
let key = h;
|
|
429
|
+
if (key === '') {
|
|
430
|
+
blank += 1;
|
|
431
|
+
key = 'column' + (idx + 1);
|
|
432
|
+
}
|
|
433
|
+
if (Object.prototype.hasOwnProperty.call(seen, key)) {
|
|
434
|
+
if (dups.indexOf(key) === -1) dups.push(key);
|
|
435
|
+
seen[key] += 1;
|
|
436
|
+
key = key + '_' + seen[key];
|
|
437
|
+
} else {
|
|
438
|
+
seen[key] = 1;
|
|
439
|
+
}
|
|
440
|
+
return key;
|
|
441
|
+
});
|
|
442
|
+
if (dups.length) notes.push({ code: 'DUP_HEADER', level: 'warn', count: dups.length, items: dups.slice(0, 6) });
|
|
443
|
+
if (blank) notes.push({ code: 'EMPTY_HEADER', level: 'warn', count: blank });
|
|
444
|
+
// 見出しが1行しか無い=データが0件
|
|
445
|
+
if (!dataRows.length) notes.push({ code: 'HEADER_ONLY', level: 'info' });
|
|
446
|
+
} else {
|
|
447
|
+
notes.push({ code: 'NO_HEADER', level: 'info' });
|
|
448
|
+
}
|
|
449
|
+
|
|
450
|
+
const ragged = [];
|
|
451
|
+
const expect = headers ? headers.length : width;
|
|
452
|
+
dataRows.forEach((r, idx) => {
|
|
453
|
+
if (r.length !== expect) ragged.push(idx + (o.header !== false ? 2 : 1));
|
|
454
|
+
});
|
|
455
|
+
if (ragged.length) notes.push({ code: 'RAGGED', level: 'warn', count: ragged.length, lines: ragged.slice(0, 6), expect });
|
|
456
|
+
|
|
457
|
+
let bigNumbers = 0;
|
|
458
|
+
let keptText = 0;
|
|
459
|
+
let formulas = 0;
|
|
460
|
+
const conflicts = [];
|
|
461
|
+
const out = [];
|
|
462
|
+
|
|
463
|
+
for (const r of dataRows) {
|
|
464
|
+
let obj;
|
|
465
|
+
if (headers) {
|
|
466
|
+
obj = {};
|
|
467
|
+
for (let c = 0; c < headers.length; c += 1) {
|
|
468
|
+
const cell = r[c] === undefined ? '' : r[c];
|
|
469
|
+
if (CSV_FORMULA_RE.test(cell)) formulas += 1;
|
|
470
|
+
if (o.types !== false && cell !== '' && CSV_NUM_RE.test(cell) && String(Number(cell)) !== cell) bigNumbers += 1;
|
|
471
|
+
else if (o.types !== false && cell !== '' && !CSV_NUM_RE.test(cell) && CSV_NUMISH_RE.test(cell)) keptText += 1;
|
|
472
|
+
const value = csvInferValue(cell, o);
|
|
473
|
+
const nested = o.nest !== false && (headers[c].indexOf('.') !== -1 || /\[\d+\]/.test(headers[c]));
|
|
474
|
+
if (nested) {
|
|
475
|
+
if (!csvSetPath(obj, csvParsePath(headers[c]), value)) {
|
|
476
|
+
if (conflicts.indexOf(headers[c]) === -1) conflicts.push(headers[c]);
|
|
477
|
+
obj[headers[c]] = value;
|
|
478
|
+
}
|
|
479
|
+
} else {
|
|
480
|
+
obj[headers[c]] = value;
|
|
481
|
+
}
|
|
482
|
+
}
|
|
483
|
+
} else {
|
|
484
|
+
obj = r.map((cell) => {
|
|
485
|
+
if (CSV_FORMULA_RE.test(cell)) formulas += 1;
|
|
486
|
+
return csvInferValue(cell, o);
|
|
487
|
+
});
|
|
488
|
+
}
|
|
489
|
+
out.push(obj);
|
|
490
|
+
}
|
|
491
|
+
|
|
492
|
+
if (conflicts.length) notes.push({ code: 'NEST_CONFLICT', level: 'warn', count: conflicts.length, items: conflicts.slice(0, 6) });
|
|
493
|
+
if (bigNumbers) notes.push({ code: 'BIG_NUMBER', level: 'warn', count: bigNumbers });
|
|
494
|
+
if (keptText) notes.push({ code: 'KEPT_TEXT', level: 'info', count: keptText });
|
|
495
|
+
if (formulas) notes.push({ code: 'FORMULA', level: 'warn', count: formulas });
|
|
496
|
+
|
|
497
|
+
const indent = o.indent === 'tab' ? '\t' : Number(o.indent === undefined ? 2 : o.indent);
|
|
498
|
+
const output = JSON.stringify(out, null, indent) + '\n';
|
|
499
|
+
let depth = 0;
|
|
500
|
+
for (const obj of out) {
|
|
501
|
+
const d = csvDepth(obj);
|
|
502
|
+
if (d > depth) depth = d;
|
|
503
|
+
}
|
|
504
|
+
|
|
505
|
+
return {
|
|
506
|
+
ok: true,
|
|
507
|
+
output,
|
|
508
|
+
json: out,
|
|
509
|
+
delimiter: name,
|
|
510
|
+
columns: headers || [],
|
|
511
|
+
notes,
|
|
512
|
+
stats: { rows: out.length, columns: expect, cells: out.length * expect, depth },
|
|
513
|
+
};
|
|
514
|
+
}
|
|
515
|
+
|
|
516
|
+
/**
|
|
517
|
+
* JSON → CSV/TSV
|
|
518
|
+
* @param {string} text
|
|
519
|
+
* @param {object} o delimiter/header/nest/newline/quoteAll/bom
|
|
520
|
+
*/
|
|
521
|
+
export function csvFromJson(text, o) {
|
|
522
|
+
const notes = [];
|
|
523
|
+
const src = String(text);
|
|
524
|
+
const trimmed = src.trim();
|
|
525
|
+
if (trimmed === '') return { ok: true, empty: true, notes };
|
|
526
|
+
|
|
527
|
+
let data;
|
|
528
|
+
const jsonl = csvParseJsonLines(trimmed);
|
|
529
|
+
if (jsonl) {
|
|
530
|
+
data = jsonl;
|
|
531
|
+
notes.push({ code: 'JSONL', level: 'info', count: jsonl.length });
|
|
532
|
+
} else {
|
|
533
|
+
const parsed = csvParseJson(trimmed);
|
|
534
|
+
if (!parsed.ok) return { ok: false, error: parsed.error, notes, excerpt: csvExcerpt(src, parsed.error) };
|
|
535
|
+
data = parsed.value;
|
|
536
|
+
if (parsed.duplicates.length) {
|
|
537
|
+
notes.push({ code: 'DUP_KEY', level: 'warn', count: parsed.duplicates.length, items: parsed.duplicates.slice(0, 6) });
|
|
538
|
+
}
|
|
539
|
+
}
|
|
540
|
+
|
|
541
|
+
if (!Array.isArray(data)) {
|
|
542
|
+
// 配列でない=1件のオブジェクト、または { data: [...] } のような包み
|
|
543
|
+
if (data && typeof data === 'object') {
|
|
544
|
+
const keys = Object.keys(data);
|
|
545
|
+
const wrapped = keys.filter((k) => Array.isArray(data[k]));
|
|
546
|
+
if (wrapped.length === 1 && keys.length <= 3) {
|
|
547
|
+
notes.push({ code: 'UNWRAPPED', level: 'info', key: wrapped[0] });
|
|
548
|
+
data = data[wrapped[0]];
|
|
549
|
+
} else {
|
|
550
|
+
notes.push({ code: 'NOT_ARRAY', level: 'info' });
|
|
551
|
+
data = [data];
|
|
552
|
+
}
|
|
553
|
+
} else {
|
|
554
|
+
notes.push({ code: 'NOT_ARRAY', level: 'info' });
|
|
555
|
+
data = [data];
|
|
556
|
+
}
|
|
557
|
+
}
|
|
558
|
+
if (!data.length) return { ok: true, empty: true, notes };
|
|
559
|
+
|
|
560
|
+
const nest = o.nest !== false;
|
|
561
|
+
const flatNote = { stringified: 0, arrays: 0 };
|
|
562
|
+
const columns = [];
|
|
563
|
+
const rowMaps = [];
|
|
564
|
+
let scalarRows = 0;
|
|
565
|
+
|
|
566
|
+
for (const item of data) {
|
|
567
|
+
const pairs = [];
|
|
568
|
+
if (item === null || typeof item !== 'object') {
|
|
569
|
+
scalarRows += 1;
|
|
570
|
+
pairs.push(['value', item === null ? '' : String(item)]);
|
|
571
|
+
} else if (Array.isArray(item)) {
|
|
572
|
+
item.forEach((v, idx) => csvFlatten(v, String(idx + 1), pairs, nest, flatNote));
|
|
573
|
+
} else {
|
|
574
|
+
for (const key of Object.keys(item)) csvFlatten(item[key], key, pairs, nest, flatNote);
|
|
575
|
+
}
|
|
576
|
+
const map = {};
|
|
577
|
+
for (const [key, value] of pairs) {
|
|
578
|
+
map[key] = value;
|
|
579
|
+
if (columns.indexOf(key) === -1) columns.push(key);
|
|
580
|
+
}
|
|
581
|
+
rowMaps.push(map);
|
|
582
|
+
}
|
|
583
|
+
|
|
584
|
+
const missing = rowMaps.filter((m) => Object.keys(m).length !== columns.length).length;
|
|
585
|
+
if (missing) notes.push({ code: 'MIXED_KEYS', level: 'warn', count: missing, columns: columns.length });
|
|
586
|
+
if (flatNote.stringified) notes.push({ code: 'STRINGIFIED', level: 'info', count: flatNote.stringified });
|
|
587
|
+
if (scalarRows) notes.push({ code: 'SCALAR_ROWS', level: 'info', count: scalarRows });
|
|
588
|
+
|
|
589
|
+
const name = !o.delimiter || o.delimiter === 'auto' ? 'comma' : o.delimiter;
|
|
590
|
+
const delim = CSV_DELIMS[name] || ',';
|
|
591
|
+
const nl = o.newline === 'crlf' ? '\r\n' : '\n';
|
|
592
|
+
const quoteAll = Boolean(o.quoteAll);
|
|
593
|
+
const lines = [];
|
|
594
|
+
if (o.header !== false) lines.push(columns.map((c) => csvQuoteCell(c, delim, quoteAll)).join(delim));
|
|
595
|
+
let formulas = 0;
|
|
596
|
+
let multiline = 0;
|
|
597
|
+
for (const map of rowMaps) {
|
|
598
|
+
const cells = columns.map((c) => {
|
|
599
|
+
const v = map[c] === undefined ? '' : map[c];
|
|
600
|
+
if (CSV_FORMULA_RE.test(v)) formulas += 1;
|
|
601
|
+
if (/[\r\n]/.test(v)) multiline += 1;
|
|
602
|
+
return csvQuoteCell(v, delim, quoteAll);
|
|
603
|
+
});
|
|
604
|
+
lines.push(cells.join(delim));
|
|
605
|
+
}
|
|
606
|
+
if (formulas) notes.push({ code: 'FORMULA', level: 'warn', count: formulas });
|
|
607
|
+
if (multiline) notes.push({ code: 'NEWLINE_IN_CELL', level: 'info', count: multiline });
|
|
608
|
+
|
|
609
|
+
const output = (o.bom ? '' : '') + lines.join(nl) + nl;
|
|
610
|
+
let depth = 0;
|
|
611
|
+
for (const item of data) {
|
|
612
|
+
const d = csvDepth(item);
|
|
613
|
+
if (d > depth) depth = d;
|
|
614
|
+
}
|
|
615
|
+
|
|
616
|
+
return {
|
|
617
|
+
ok: true,
|
|
618
|
+
output,
|
|
619
|
+
delimiter: name,
|
|
620
|
+
columns,
|
|
621
|
+
notes,
|
|
622
|
+
stats: { rows: rowMaps.length, columns: columns.length, cells: rowMaps.length * columns.length, depth },
|
|
623
|
+
};
|
|
624
|
+
}
|
|
625
|
+
|
|
626
|
+
/**
|
|
627
|
+
* CSV/TSV ⇄ JSON の相互変換。
|
|
628
|
+
*
|
|
629
|
+
* @param {object} opts
|
|
630
|
+
* @param {'csv2json'|'json2csv'} opts.direction 変換の向き
|
|
631
|
+
* @param {string} opts.text 入力
|
|
632
|
+
* @param {'auto'|'comma'|'tab'|'semicolon'|'pipe'} [opts.delimiter='auto'] 区切り文字(json2csv では auto=カンマ)
|
|
633
|
+
* @param {boolean} [opts.header=true] 1行目を見出しとして扱う/書き出す
|
|
634
|
+
* @param {boolean} [opts.nest=true] a.b 列 ⇄ 入れ子オブジェクト
|
|
635
|
+
* @param {boolean} [opts.types=true] true/false/null と数値を読み替える(csv2json)
|
|
636
|
+
* @param {boolean} [opts.trim=true] セル前後の空白を落とす(csv2json)
|
|
637
|
+
* @param {boolean} [opts.emptyNull=false] 空のセルを null にする(csv2json)
|
|
638
|
+
* @param {number|'tab'} [opts.indent=2] JSONのインデント(csv2json)
|
|
639
|
+
* @param {'lf'|'crlf'} [opts.newline='lf'] 改行コード(json2csv)
|
|
640
|
+
* @param {boolean} [opts.quoteAll=false] すべてのセルを引用符で囲む(json2csv)
|
|
641
|
+
* @param {boolean} [opts.bom=false] 先頭にBOMを付ける(json2csv・Excel向け)
|
|
642
|
+
*/
|
|
643
|
+
export function csvJsonConvert(opts) {
|
|
644
|
+
const o = opts || {};
|
|
645
|
+
const text = typeof o.text === 'string' ? o.text : '';
|
|
646
|
+
const settings = {
|
|
647
|
+
delimiter: o.delimiter || 'auto',
|
|
648
|
+
header: o.header !== false,
|
|
649
|
+
nest: o.nest !== false,
|
|
650
|
+
types: o.types !== false,
|
|
651
|
+
trim: o.trim !== false,
|
|
652
|
+
emptyNull: Boolean(o.emptyNull),
|
|
653
|
+
skipEmpty: o.skipEmpty !== false,
|
|
654
|
+
indent: o.indent === undefined ? 2 : o.indent,
|
|
655
|
+
newline: o.newline === 'crlf' ? 'crlf' : 'lf',
|
|
656
|
+
quoteAll: Boolean(o.quoteAll),
|
|
657
|
+
bom: Boolean(o.bom),
|
|
658
|
+
};
|
|
659
|
+
if (text.trim() === '') return { ok: true, empty: true, notes: [], settings, input_bytes: 0, output_bytes: 0 };
|
|
660
|
+
|
|
661
|
+
const r = o.direction === 'json2csv' ? csvFromJson(text, settings) : csvToJson(text, settings);
|
|
662
|
+
r.direction = o.direction === 'json2csv' ? 'json2csv' : 'csv2json';
|
|
663
|
+
r.settings = settings;
|
|
664
|
+
r.input_bytes = csvByteLength(text);
|
|
665
|
+
r.output_bytes = r.output ? csvByteLength(r.output) : 0;
|
|
666
|
+
if (!r.notes) r.notes = [];
|
|
667
|
+
return r;
|
|
668
|
+
}
|
|
669
|
+
|
|
670
|
+
/** UTF-8としてのバイト数 */
|
|
671
|
+
export function csvByteLength(s) {
|
|
672
|
+
let bytes = 0;
|
|
673
|
+
for (const ch of String(s)) {
|
|
674
|
+
const code = ch.codePointAt(0);
|
|
675
|
+
bytes += code < 0x80 ? 1 : code < 0x800 ? 2 : code < 0x10000 ? 3 : 4;
|
|
676
|
+
}
|
|
677
|
+
return bytes;
|
|
678
|
+
}
|
|
679
|
+
|
|
680
|
+
/* ==================== ここまで変換コア ==================== */
|
|
681
|
+
|
|
682
|
+
/* ---- ここから下はMCP側だけの層(入出力・文言)。site側の app.js には UI 用の同じ文言表がある ---- */
|
|
683
|
+
|
|
684
|
+
export class CsvConvertError extends Error {}
|
|
685
|
+
|
|
686
|
+
const DELIM_JA = { comma: 'カンマ', tab: 'タブ', semicolon: 'セミコロン', pipe: 'パイプ' };
|
|
687
|
+
const DELIM_EN = { comma: 'comma', tab: 'tab', semicolon: 'semicolon', pipe: 'pipe' };
|
|
688
|
+
|
|
689
|
+
const ERR_JA = {
|
|
690
|
+
UNCLOSED_QUOTE: () => '引用符 " が閉じられていません。ここから後ろがすべて1つのセルとして読まれています。セルの中の " は "" と2つ重ねて書きます。',
|
|
691
|
+
UNEXPECTED_END: () => 'JSONが途中で終わっています。閉じ括弧か引用符が足りません。',
|
|
692
|
+
NEWLINE_IN_STRING: () => '文字列の途中で改行しています。JSONの文字列に生の改行は書けないので \\n と書きます。',
|
|
693
|
+
BAD_ESCAPE: (e) => `使えないエスケープです(${e.found})。JSONで書けるのは \\" \\\\ \\/ \\b \\f \\n \\r \\t と \\u に続く16進4桁だけです。`,
|
|
694
|
+
EXPECTED_KEY: (e) => `オブジェクトのキーが要ります(${e.found ? '見つかったのは ' + e.found : '入力がここで終わっています'})。キーは必ず " で囲みます。`,
|
|
695
|
+
EXPECTED_COLON: () => 'キーの後ろにコロン : が要ります。',
|
|
696
|
+
EXPECTED_COMMA: (e) => `区切りのカンマか閉じ括弧が要ります(${e.found ? '見つかったのは ' + e.found : '入力がここで終わっています'})。最後の要素の後ろにカンマは書けません。`,
|
|
697
|
+
UNEXPECTED_CHAR: (e) => `値として読めない文字です(${e.found})。文字列は " で囲み、true / false / null は小文字で書きます。`,
|
|
698
|
+
TRAILING_TEXT: (e) => `JSONの終わりの後ろに文字が残っています(${e.found})。1行1件のJSON Lines として渡すこともできます。`,
|
|
699
|
+
};
|
|
700
|
+
|
|
701
|
+
const ERR_EN = {
|
|
702
|
+
UNCLOSED_QUOTE: () => 'A quotation mark is never closed, so everything after it is being read as one cell. Write a literal " as "" inside a quoted cell.',
|
|
703
|
+
UNEXPECTED_END: () => 'The JSON ends early — a closing bracket or quote is missing.',
|
|
704
|
+
NEWLINE_IN_STRING: () => 'A string contains a raw line break. JSON strings must write it as \\n.',
|
|
705
|
+
BAD_ESCAPE: (e) => `Unsupported escape (${e.found}). JSON allows \\" \\\\ \\/ \\b \\f \\n \\r \\t and \\u followed by four hex digits.`,
|
|
706
|
+
EXPECTED_KEY: (e) => `An object key is expected here${e.found ? ` (found ${e.found})` : ' (the input ends here)'}. Keys must be quoted with ".`,
|
|
707
|
+
EXPECTED_COLON: () => 'A colon : is expected after the key.',
|
|
708
|
+
EXPECTED_COMMA: (e) => `A comma or a closing bracket is expected${e.found ? ` (found ${e.found})` : ' (the input ends here)'}. A trailing comma after the last element is not valid JSON.`,
|
|
709
|
+
UNEXPECTED_CHAR: (e) => `This character cannot start a value (${e.found}). Strings need double quotes, and true / false / null are lower case.`,
|
|
710
|
+
TRAILING_TEXT: (e) => `There is text after the end of the JSON value (${e.found}). You can also paste JSON Lines — one object per line.`,
|
|
711
|
+
};
|
|
712
|
+
|
|
713
|
+
const NOTE_JA = {
|
|
714
|
+
BOM: () => '先頭にBOMが付いていました(Excelが書いたCSVによくあります)。読み取り時に取り除いています。',
|
|
715
|
+
CRLF: () => '改行コードはCRLF(Windows)です。JSONへの変換では影響しません。',
|
|
716
|
+
DELIM_AUTO: (n) => `区切り文字は${DELIM_JA[n.delimiter]}と判定しました。ずれている場合は delimiter で固定できます。`,
|
|
717
|
+
STRAY_QUOTE: (n) => `セルの途中に閉じられていない引用符があります(${n.count}行)。文字として読み進めましたが、書き出した側が壊れている可能性があります。`,
|
|
718
|
+
NEWLINE_IN_CELL: (n) => `セルの中に改行を含む値が ${n.count} 件あります。引用符で囲んで保持しています。`,
|
|
719
|
+
DUP_HEADER: (n) => `見出しに同じ列名が ${n.count} 組あります: ${n.items.join(' / ')} — 後ろの列は ${n.items[0]}_2 のように連番を付けて区別しました。`,
|
|
720
|
+
EMPTY_HEADER: (n) => `見出しが空の列が ${n.count} 列あります。column1 のような仮の名前を付けています。`,
|
|
721
|
+
HEADER_ONLY: () => '見出し行だけでデータ行がありません。結果は空の配列になります。',
|
|
722
|
+
NO_HEADER: () => '見出し行なしとして読んだので、各行は配列(値の並び)になります。',
|
|
723
|
+
RAGGED: (n) => `列数が見出しと合わない行が ${n.count} 行あります(見出しは ${n.expect} 列)。足りない分は空、多い分は捨てています。`,
|
|
724
|
+
NEST_CONFLICT: (n) => `入れ子にできない列名が ${n.count} 件あります: ${n.items.join(' / ')} — 同じ名前がすでに値で埋まっているため、その列は平らなキーのままにしました。`,
|
|
725
|
+
BIG_NUMBER: (n) => `数値として正確に表せない値が ${n.count} 件あります(2の53乗を超える整数など)。桁落ちを避けるため文字列のまま残しました。`,
|
|
726
|
+
KEPT_TEXT: (n) => `数字に見えるが表記が変わってしまう値が ${n.count} 件あります(0123・+1・090-1234-5678 など)。文字列のまま残しました。`,
|
|
727
|
+
FORMULA: (n) => `=・+・-・@ で始まるセルが ${n.count} 件あります。表計算ソフトが数式として実行することがあるため、外部から受け取ったCSVでは中身を確認してください。`,
|
|
728
|
+
JSONL: (n) => `1行1JSON(JSON Lines)として ${n.count} 件読み取りました。`,
|
|
729
|
+
DUP_KEY: (n) => `同じキーが重複しているオブジェクトがあります: ${n.items.join(' / ')} — 後に書いた値が勝ちます。`,
|
|
730
|
+
UNWRAPPED: (n) => `配列を包んでいるキー「${n.key}」の中身を表にしました。`,
|
|
731
|
+
NOT_ARRAY: () => '配列ではなかったので、1件のデータとして1行に書き出しました。',
|
|
732
|
+
MIXED_KEYS: (n) => `キーの並びが他と違うオブジェクトが ${n.count} 件あります。全体で ${n.columns} 列に揃え、無い項目は空にしました。`,
|
|
733
|
+
STRINGIFIED: (n) => `入れ子の値 ${n.count} 件を、JSONの文字列としてセルに入れました。nest=true にすると列へ割れます。`,
|
|
734
|
+
SCALAR_ROWS: (n) => `オブジェクトでない要素が ${n.count} 件あります。value 列に入れました。`,
|
|
735
|
+
};
|
|
736
|
+
|
|
737
|
+
const NOTE_EN = {
|
|
738
|
+
BOM: () => 'The input started with a BOM (common in CSV written by Excel). It was removed while reading.',
|
|
739
|
+
CRLF: () => 'Line endings are CRLF (Windows). This does not affect the JSON output.',
|
|
740
|
+
DELIM_AUTO: (n) => `Detected the ${DELIM_EN[n.delimiter]} as the delimiter. Set delimiter explicitly if that is wrong.`,
|
|
741
|
+
STRAY_QUOTE: (n) => `A quotation mark appears in the middle of a cell on ${n.count} line(s). It was read as a literal character, but whatever produced this file may be broken.`,
|
|
742
|
+
NEWLINE_IN_CELL: (n) => `${n.count} value(s) contain a line break. They are kept, wrapped in quotes.`,
|
|
743
|
+
DUP_HEADER: (n) => `${n.count} header name(s) appear more than once: ${n.items.join(' / ')} — later columns were numbered (${n.items[0]}_2) to keep them apart.`,
|
|
744
|
+
EMPTY_HEADER: (n) => `${n.count} column(s) have an empty header. They were named column1, column2 and so on.`,
|
|
745
|
+
HEADER_ONLY: () => 'There is a header row but no data rows, so the result is an empty array.',
|
|
746
|
+
NO_HEADER: () => 'Read without a header row, so every row becomes an array of values.',
|
|
747
|
+
RAGGED: (n) => `${n.count} row(s) do not have the same number of columns as the header (${n.expect}). Missing cells are empty; extra cells are dropped.`,
|
|
748
|
+
NEST_CONFLICT: (n) => `${n.count} column(s) could not be nested: ${n.items.join(' / ')} — the same name already holds a value, so they were kept as flat keys.`,
|
|
749
|
+
BIG_NUMBER: (n) => `${n.count} value(s) cannot be represented exactly as a number (integers beyond 2^53, for example). They were kept as strings so no digits are lost.`,
|
|
750
|
+
KEPT_TEXT: (n) => `${n.count} value(s) look numeric but would change if converted (0123, +1, 090-1234-5678). They were kept as strings.`,
|
|
751
|
+
FORMULA: (n) => `${n.count} cell(s) start with =, +, - or @. Spreadsheets may execute those as formulas, so check them when the CSV came from outside.`,
|
|
752
|
+
JSONL: (n) => `Read ${n.count} records as JSON Lines (one object per line).`,
|
|
753
|
+
DUP_KEY: (n) => `Some objects repeat the same key: ${n.items.join(' / ')} — the last one wins.`,
|
|
754
|
+
UNWRAPPED: (n) => `Used the array inside the wrapper key “${n.key}” as the table.`,
|
|
755
|
+
NOT_ARRAY: () => 'The value was not an array, so it was written as a single row.',
|
|
756
|
+
MIXED_KEYS: (n) => `${n.count} object(s) have a different set of keys. The table was squared off to ${n.columns} columns and missing values are empty.`,
|
|
757
|
+
STRINGIFIED: (n) => `${n.count} nested value(s) were written into a single cell as JSON text. Set nest=true to split them into columns.`,
|
|
758
|
+
SCALAR_ROWS: (n) => `${n.count} element(s) are not objects. They were placed in a “value” column.`,
|
|
759
|
+
};
|
|
760
|
+
|
|
761
|
+
/** 向きを指定されなかったときの推定。JSONは [ か { で始まる */
|
|
762
|
+
export function csvGuessDirection(text) {
|
|
763
|
+
const head = String(text).replace(/^/, '').trim();
|
|
764
|
+
return head.startsWith('[') || head.startsWith('{') ? 'json2csv' : 'csv2json';
|
|
765
|
+
}
|
|
766
|
+
|
|
767
|
+
/**
|
|
768
|
+
* MCPツール csv_convert の本体。text / path を受け取り、変換結果と指摘事項を返す。
|
|
769
|
+
* 変換そのものは csvJsonConvert(site版と同一のコア)に任せ、ここは入出力と文言だけを持つ。
|
|
770
|
+
*/
|
|
771
|
+
export async function csvConvertTool(opts = {}) {
|
|
772
|
+
const o = opts || {};
|
|
773
|
+
const lang = o.lang === undefined ? 'ja' : o.lang;
|
|
774
|
+
if (lang !== 'ja' && lang !== 'en') throw new CsvConvertError(`lang は ja か en: ${o.lang}`);
|
|
775
|
+
|
|
776
|
+
const hasText = typeof o.text === 'string';
|
|
777
|
+
if (hasText === Boolean(o.path)) throw new CsvConvertError('text か path のどちらか一方を渡してください');
|
|
778
|
+
|
|
779
|
+
if (o.delimiter !== undefined && o.delimiter !== 'auto' && !CSV_DELIMS[o.delimiter]) {
|
|
780
|
+
throw new CsvConvertError(`delimiter は auto / ${Object.keys(CSV_DELIMS).join(' / ')} のいずれか: ${o.delimiter}`);
|
|
781
|
+
}
|
|
782
|
+
if (o.indent !== undefined && o.indent !== 'tab' && !(Number.isInteger(o.indent) && o.indent >= 0 && o.indent <= 8)) {
|
|
783
|
+
throw new CsvConvertError(`indent は 0-8 の整数か "tab": ${o.indent}`);
|
|
784
|
+
}
|
|
785
|
+
if (o.newline !== undefined && o.newline !== 'lf' && o.newline !== 'crlf') {
|
|
786
|
+
throw new CsvConvertError(`newline は lf か crlf: ${o.newline}`);
|
|
787
|
+
}
|
|
788
|
+
if (o.direction !== undefined && ['auto', 'csv2json', 'json2csv'].indexOf(o.direction) === -1) {
|
|
789
|
+
throw new CsvConvertError(`direction は auto / csv2json / json2csv のいずれか: ${o.direction}`);
|
|
790
|
+
}
|
|
791
|
+
|
|
792
|
+
const input = hasText ? o.text : await readFile(o.path, 'utf8');
|
|
793
|
+
const guessed = !o.direction || o.direction === 'auto';
|
|
794
|
+
const direction = guessed ? csvGuessDirection(input) : o.direction;
|
|
795
|
+
|
|
796
|
+
const r = csvJsonConvert({
|
|
797
|
+
direction,
|
|
798
|
+
text: input,
|
|
799
|
+
delimiter: o.delimiter,
|
|
800
|
+
header: o.header,
|
|
801
|
+
nest: o.nest,
|
|
802
|
+
types: o.types,
|
|
803
|
+
trim: o.trim,
|
|
804
|
+
emptyNull: o.emptyNull,
|
|
805
|
+
indent: o.indent,
|
|
806
|
+
newline: o.newline,
|
|
807
|
+
quoteAll: o.quoteAll,
|
|
808
|
+
bom: o.bom,
|
|
809
|
+
});
|
|
810
|
+
|
|
811
|
+
const source = hasText ? { type: 'text' } : { path: o.path, name: basename(o.path) };
|
|
812
|
+
|
|
813
|
+
if (r.ok === false) {
|
|
814
|
+
const table = lang === 'en' ? ERR_EN : ERR_JA;
|
|
815
|
+
const e = r.error;
|
|
816
|
+
const message = table[e.code] ? table[e.code](e) : e.code;
|
|
817
|
+
// 入力が壊れているのは「呼び出し方の誤り」ではないので、例外ではなく結果として返す。
|
|
818
|
+
// 行・桁・前後の抜粋がそのまま直しどころになる。
|
|
819
|
+
return {
|
|
820
|
+
ok: false,
|
|
821
|
+
direction,
|
|
822
|
+
direction_guessed: guessed,
|
|
823
|
+
source,
|
|
824
|
+
error: { code: e.code, message, line: e.line, column: e.col, found: e.found },
|
|
825
|
+
excerpt: r.excerpt,
|
|
826
|
+
notes: renderNotes(r.notes, lang),
|
|
827
|
+
};
|
|
828
|
+
}
|
|
829
|
+
|
|
830
|
+
const result = {
|
|
831
|
+
ok: true,
|
|
832
|
+
direction,
|
|
833
|
+
direction_guessed: guessed,
|
|
834
|
+
source,
|
|
835
|
+
settings: r.settings,
|
|
836
|
+
notes: renderNotes(r.notes, lang),
|
|
837
|
+
};
|
|
838
|
+
if (r.empty) {
|
|
839
|
+
result.empty = true;
|
|
840
|
+
return result;
|
|
841
|
+
}
|
|
842
|
+
result.delimiter = r.delimiter;
|
|
843
|
+
result.columns = r.columns;
|
|
844
|
+
result.stats = r.stats;
|
|
845
|
+
result.input_bytes = r.input_bytes;
|
|
846
|
+
result.output_bytes = r.output_bytes;
|
|
847
|
+
|
|
848
|
+
if (o.outputPath) {
|
|
849
|
+
await writeFile(o.outputPath, r.output, 'utf8');
|
|
850
|
+
result.output_path = o.outputPath;
|
|
851
|
+
// ファイルへ書けたなら本文は重複した重い情報でしかない
|
|
852
|
+
} else {
|
|
853
|
+
result.output = r.output;
|
|
854
|
+
}
|
|
855
|
+
return result;
|
|
856
|
+
}
|
|
857
|
+
|
|
858
|
+
function renderNotes(notes, lang) {
|
|
859
|
+
const table = lang === 'en' ? NOTE_EN : NOTE_JA;
|
|
860
|
+
return (notes || []).map((n) => ({
|
|
861
|
+
code: n.code,
|
|
862
|
+
level: n.level || 'info',
|
|
863
|
+
message: table[n.code] ? table[n.code](n) : n.code,
|
|
864
|
+
}));
|
|
865
|
+
}
|