rowpipe 1.0.0
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 +401 -0
- package/benchmarks/memory-bench.ts +109 -0
- package/dist/analytics/reduce.d.ts +33 -0
- package/dist/analytics/reduce.d.ts.map +1 -0
- package/dist/analytics/reduce.js +318 -0
- package/dist/analytics/reduce.js.map +1 -0
- package/dist/analytics/schema-inference.d.ts +31 -0
- package/dist/analytics/schema-inference.d.ts.map +1 -0
- package/dist/analytics/schema-inference.js +166 -0
- package/dist/analytics/schema-inference.js.map +1 -0
- package/dist/analytics/semantic-types.d.ts +6 -0
- package/dist/analytics/semantic-types.d.ts.map +1 -0
- package/dist/analytics/semantic-types.js +40 -0
- package/dist/analytics/semantic-types.js.map +1 -0
- package/dist/analytics/stats.d.ts +102 -0
- package/dist/analytics/stats.d.ts.map +1 -0
- package/dist/analytics/stats.js +307 -0
- package/dist/analytics/stats.js.map +1 -0
- package/dist/analytics/validator.d.ts +32 -0
- package/dist/analytics/validator.d.ts.map +1 -0
- package/dist/analytics/validator.js +96 -0
- package/dist/analytics/validator.js.map +1 -0
- package/dist/cli/commands/cast.d.ts +13 -0
- package/dist/cli/commands/cast.d.ts.map +1 -0
- package/dist/cli/commands/cast.js +39 -0
- package/dist/cli/commands/cast.js.map +1 -0
- package/dist/cli/commands/convert.d.ts +15 -0
- package/dist/cli/commands/convert.d.ts.map +1 -0
- package/dist/cli/commands/convert.js +89 -0
- package/dist/cli/commands/convert.js.map +1 -0
- package/dist/cli/commands/filter.d.ts +11 -0
- package/dist/cli/commands/filter.d.ts.map +1 -0
- package/dist/cli/commands/filter.js +38 -0
- package/dist/cli/commands/filter.js.map +1 -0
- package/dist/cli/commands/inspect.d.ts +10 -0
- package/dist/cli/commands/inspect.d.ts.map +1 -0
- package/dist/cli/commands/inspect.js +134 -0
- package/dist/cli/commands/inspect.js.map +1 -0
- package/dist/cli/commands/map.d.ts +11 -0
- package/dist/cli/commands/map.d.ts.map +1 -0
- package/dist/cli/commands/map.js +43 -0
- package/dist/cli/commands/map.js.map +1 -0
- package/dist/cli/commands/reduce.d.ts +12 -0
- package/dist/cli/commands/reduce.d.ts.map +1 -0
- package/dist/cli/commands/reduce.js +49 -0
- package/dist/cli/commands/reduce.js.map +1 -0
- package/dist/cli/commands/rename.d.ts +11 -0
- package/dist/cli/commands/rename.d.ts.map +1 -0
- package/dist/cli/commands/rename.js +39 -0
- package/dist/cli/commands/rename.js.map +1 -0
- package/dist/cli/commands/sample.d.ts +13 -0
- package/dist/cli/commands/sample.d.ts.map +1 -0
- package/dist/cli/commands/sample.js +40 -0
- package/dist/cli/commands/sample.js.map +1 -0
- package/dist/cli/commands/schema.d.ts +13 -0
- package/dist/cli/commands/schema.d.ts.map +1 -0
- package/dist/cli/commands/schema.js +57 -0
- package/dist/cli/commands/schema.js.map +1 -0
- package/dist/cli/commands/select.d.ts +11 -0
- package/dist/cli/commands/select.d.ts.map +1 -0
- package/dist/cli/commands/select.js +39 -0
- package/dist/cli/commands/select.js.map +1 -0
- package/dist/cli/commands/stats.d.ts +15 -0
- package/dist/cli/commands/stats.d.ts.map +1 -0
- package/dist/cli/commands/stats.js +75 -0
- package/dist/cli/commands/stats.js.map +1 -0
- package/dist/cli/commands/validate.d.ts +12 -0
- package/dist/cli/commands/validate.d.ts.map +1 -0
- package/dist/cli/commands/validate.js +86 -0
- package/dist/cli/commands/validate.js.map +1 -0
- package/dist/cli/index.d.ts +3 -0
- package/dist/cli/index.d.ts.map +1 -0
- package/dist/cli/index.js +262 -0
- package/dist/cli/index.js.map +1 -0
- package/dist/core/batch.d.ts +15 -0
- package/dist/core/batch.d.ts.map +1 -0
- package/dist/core/batch.js +67 -0
- package/dist/core/batch.js.map +1 -0
- package/dist/core/errors.d.ts +35 -0
- package/dist/core/errors.d.ts.map +1 -0
- package/dist/core/errors.js +55 -0
- package/dist/core/errors.js.map +1 -0
- package/dist/core/pipeline.d.ts +40 -0
- package/dist/core/pipeline.d.ts.map +1 -0
- package/dist/core/pipeline.js +120 -0
- package/dist/core/pipeline.js.map +1 -0
- package/dist/core/types.d.ts +73 -0
- package/dist/core/types.d.ts.map +1 -0
- package/dist/core/types.js +2 -0
- package/dist/core/types.js.map +1 -0
- package/dist/index.d.ts +22 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +28 -0
- package/dist/index.js.map +1 -0
- package/dist/readers/csv.d.ts +24 -0
- package/dist/readers/csv.d.ts.map +1 -0
- package/dist/readers/csv.js +251 -0
- package/dist/readers/csv.js.map +1 -0
- package/dist/readers/index.d.ts +17 -0
- package/dist/readers/index.d.ts.map +1 -0
- package/dist/readers/index.js +74 -0
- package/dist/readers/index.js.map +1 -0
- package/dist/readers/json.d.ts +20 -0
- package/dist/readers/json.d.ts.map +1 -0
- package/dist/readers/json.js +219 -0
- package/dist/readers/json.js.map +1 -0
- package/dist/readers/jsonl.d.ts +14 -0
- package/dist/readers/jsonl.d.ts.map +1 -0
- package/dist/readers/jsonl.js +95 -0
- package/dist/readers/jsonl.js.map +1 -0
- package/dist/readers/xlsx.d.ts +18 -0
- package/dist/readers/xlsx.d.ts.map +1 -0
- package/dist/readers/xlsx.js +138 -0
- package/dist/readers/xlsx.js.map +1 -0
- package/dist/transforms/cast.d.ts +21 -0
- package/dist/transforms/cast.d.ts.map +1 -0
- package/dist/transforms/cast.js +182 -0
- package/dist/transforms/cast.js.map +1 -0
- package/dist/transforms/expression.d.ts +67 -0
- package/dist/transforms/expression.d.ts.map +1 -0
- package/dist/transforms/expression.js +943 -0
- package/dist/transforms/expression.js.map +1 -0
- package/dist/transforms/filter.d.ts +6 -0
- package/dist/transforms/filter.d.ts.map +1 -0
- package/dist/transforms/filter.js +31 -0
- package/dist/transforms/filter.js.map +1 -0
- package/dist/transforms/map.d.ts +12 -0
- package/dist/transforms/map.d.ts.map +1 -0
- package/dist/transforms/map.js +76 -0
- package/dist/transforms/map.js.map +1 -0
- package/dist/transforms/rename.d.ts +10 -0
- package/dist/transforms/rename.d.ts.map +1 -0
- package/dist/transforms/rename.js +39 -0
- package/dist/transforms/rename.js.map +1 -0
- package/dist/transforms/sample.d.ts +10 -0
- package/dist/transforms/sample.d.ts.map +1 -0
- package/dist/transforms/sample.js +47 -0
- package/dist/transforms/sample.js.map +1 -0
- package/dist/transforms/select.d.ts +6 -0
- package/dist/transforms/select.d.ts.map +1 -0
- package/dist/transforms/select.js +63 -0
- package/dist/transforms/select.js.map +1 -0
- package/dist/utils/compression.d.ts +14 -0
- package/dist/utils/compression.d.ts.map +1 -0
- package/dist/utils/compression.js +38 -0
- package/dist/utils/compression.js.map +1 -0
- package/dist/utils/formatting.d.ts +6 -0
- package/dist/utils/formatting.d.ts.map +1 -0
- package/dist/utils/formatting.js +56 -0
- package/dist/utils/formatting.js.map +1 -0
- package/dist/utils/progress.d.ts +13 -0
- package/dist/utils/progress.d.ts.map +1 -0
- package/dist/utils/progress.js +27 -0
- package/dist/utils/progress.js.map +1 -0
- package/dist/writers/csv.d.ts +18 -0
- package/dist/writers/csv.d.ts.map +1 -0
- package/dist/writers/csv.js +84 -0
- package/dist/writers/csv.js.map +1 -0
- package/dist/writers/index.d.ts +17 -0
- package/dist/writers/index.d.ts.map +1 -0
- package/dist/writers/index.js +73 -0
- package/dist/writers/index.js.map +1 -0
- package/dist/writers/json.d.ts +12 -0
- package/dist/writers/json.d.ts.map +1 -0
- package/dist/writers/json.js +42 -0
- package/dist/writers/json.js.map +1 -0
- package/dist/writers/jsonl.d.ts +12 -0
- package/dist/writers/jsonl.d.ts.map +1 -0
- package/dist/writers/jsonl.js +41 -0
- package/dist/writers/jsonl.js.map +1 -0
- package/dist/writers/xlsx.d.ts +13 -0
- package/dist/writers/xlsx.d.ts.map +1 -0
- package/dist/writers/xlsx.js +65 -0
- package/dist/writers/xlsx.js.map +1 -0
- package/package.json +44 -0
- package/src/analytics/reduce.ts +388 -0
- package/src/analytics/schema-inference.ts +217 -0
- package/src/analytics/semantic-types.ts +36 -0
- package/src/analytics/stats.ts +383 -0
- package/src/analytics/validator.ts +134 -0
- package/src/cli/commands/cast.ts +60 -0
- package/src/cli/commands/convert.ts +128 -0
- package/src/cli/commands/filter.ts +58 -0
- package/src/cli/commands/inspect.ts +193 -0
- package/src/cli/commands/map.ts +66 -0
- package/src/cli/commands/reduce.ts +74 -0
- package/src/cli/commands/rename.ts +59 -0
- package/src/cli/commands/sample.ts +61 -0
- package/src/cli/commands/schema.ts +86 -0
- package/src/cli/commands/select.ts +59 -0
- package/src/cli/commands/stats.ts +100 -0
- package/src/cli/commands/validate.ts +123 -0
- package/src/cli/index.ts +285 -0
- package/src/core/batch.ts +81 -0
- package/src/core/errors.ts +81 -0
- package/src/core/pipeline.ts +150 -0
- package/src/core/types.ts +110 -0
- package/src/index.ts +52 -0
- package/src/readers/csv.ts +302 -0
- package/src/readers/index.ts +88 -0
- package/src/readers/json.ts +256 -0
- package/src/readers/jsonl.ts +126 -0
- package/src/readers/xlsx.ts +182 -0
- package/src/transforms/cast.ts +212 -0
- package/src/transforms/expression.ts +1114 -0
- package/src/transforms/filter.ts +38 -0
- package/src/transforms/map.ts +99 -0
- package/src/transforms/rename.ts +47 -0
- package/src/transforms/sample.ts +57 -0
- package/src/transforms/select.ts +64 -0
- package/src/utils/compression.ts +43 -0
- package/src/utils/formatting.ts +75 -0
- package/src/utils/progress.ts +37 -0
- package/src/writers/csv.ts +106 -0
- package/src/writers/index.ts +87 -0
- package/src/writers/json.ts +52 -0
- package/src/writers/jsonl.ts +49 -0
- package/src/writers/xlsx.ts +87 -0
- package/tests/analytics.test.ts +119 -0
- package/tests/cli-integration.test.ts +187 -0
- package/tests/map-reduce.test.ts +223 -0
- package/tests/readers-writers.test.ts +189 -0
- package/tests/transforms.test.ts +199 -0
- package/tsconfig.json +21 -0
- package/vitest.config.ts +9 -0
|
@@ -0,0 +1,251 @@
|
|
|
1
|
+
import { createReadStream } from "node:fs";
|
|
2
|
+
import { ParseError } from "../core/errors.js";
|
|
3
|
+
/**
|
|
4
|
+
* Streaming CSV Reader supporting RFC 4180 quotes, multiline cells,
|
|
5
|
+
* custom & auto-detected delimiters, BOM removal, and exact error reporting.
|
|
6
|
+
*/
|
|
7
|
+
export class CSVReader {
|
|
8
|
+
input;
|
|
9
|
+
options;
|
|
10
|
+
constructor(input, options = {}) {
|
|
11
|
+
this.input = input;
|
|
12
|
+
this.options = { ...options };
|
|
13
|
+
if (typeof input === "string") {
|
|
14
|
+
this.options.filePath = input;
|
|
15
|
+
}
|
|
16
|
+
}
|
|
17
|
+
getInputStream() {
|
|
18
|
+
if (typeof this.input === "string") {
|
|
19
|
+
return createReadStream(this.input);
|
|
20
|
+
}
|
|
21
|
+
return this.input;
|
|
22
|
+
}
|
|
23
|
+
/**
|
|
24
|
+
* Auto-detects delimiter by scanning candidate delimiters in a sample string.
|
|
25
|
+
*/
|
|
26
|
+
static detectDelimiter(sample) {
|
|
27
|
+
const candidates = [",", "\t", ";", "|"];
|
|
28
|
+
const firstLine = sample.split(/\r?\n/)[0] || "";
|
|
29
|
+
let bestDelimiter = ",";
|
|
30
|
+
let maxCount = -1;
|
|
31
|
+
for (const cand of candidates) {
|
|
32
|
+
let count = 0;
|
|
33
|
+
let inQuotes = false;
|
|
34
|
+
for (let i = 0; i < firstLine.length; i++) {
|
|
35
|
+
const ch = firstLine[i];
|
|
36
|
+
if (ch === '"') {
|
|
37
|
+
inQuotes = !inQuotes;
|
|
38
|
+
}
|
|
39
|
+
else if (ch === cand && !inQuotes) {
|
|
40
|
+
count++;
|
|
41
|
+
}
|
|
42
|
+
}
|
|
43
|
+
if (count > maxCount) {
|
|
44
|
+
maxCount = count;
|
|
45
|
+
bestDelimiter = cand;
|
|
46
|
+
}
|
|
47
|
+
}
|
|
48
|
+
return maxCount > 0 ? bestDelimiter : ",";
|
|
49
|
+
}
|
|
50
|
+
async *read(options) {
|
|
51
|
+
const mergedOptions = {
|
|
52
|
+
...this.options,
|
|
53
|
+
...options,
|
|
54
|
+
};
|
|
55
|
+
const batchSize = Math.max(1, mergedOptions.batchSize ?? 1000);
|
|
56
|
+
const hasHeader = mergedOptions.header !== false;
|
|
57
|
+
let delimiter = mergedOptions.delimiter;
|
|
58
|
+
const stream = this.getInputStream();
|
|
59
|
+
let buffer = "";
|
|
60
|
+
let headers = null;
|
|
61
|
+
let isFirstChunk = true;
|
|
62
|
+
let rowsInCurrentBatch = [];
|
|
63
|
+
let globalRowOffset = 0;
|
|
64
|
+
let rowNumber = 0;
|
|
65
|
+
let byteOffset = 0;
|
|
66
|
+
let currentField = "";
|
|
67
|
+
let currentRowFields = [];
|
|
68
|
+
let inQuotes = false;
|
|
69
|
+
let cursor = 0;
|
|
70
|
+
for await (const chunk of stream) {
|
|
71
|
+
const chunkStr = typeof chunk === "string" ? chunk : chunk.toString("utf8");
|
|
72
|
+
if (isFirstChunk) {
|
|
73
|
+
isFirstChunk = false;
|
|
74
|
+
// Remove UTF-8 BOM if present
|
|
75
|
+
if (chunkStr.charCodeAt(0) === 0xfeff) {
|
|
76
|
+
buffer = chunkStr.slice(1);
|
|
77
|
+
}
|
|
78
|
+
else {
|
|
79
|
+
buffer = chunkStr;
|
|
80
|
+
}
|
|
81
|
+
// Auto-detect delimiter if not specified
|
|
82
|
+
if (!delimiter) {
|
|
83
|
+
delimiter = CSVReader.detectDelimiter(buffer);
|
|
84
|
+
}
|
|
85
|
+
}
|
|
86
|
+
else {
|
|
87
|
+
buffer += chunkStr;
|
|
88
|
+
}
|
|
89
|
+
const activeDelimiter = delimiter ?? ",";
|
|
90
|
+
const delimLen = activeDelimiter.length;
|
|
91
|
+
while (cursor < buffer.length) {
|
|
92
|
+
const char = buffer[cursor];
|
|
93
|
+
if (inQuotes) {
|
|
94
|
+
if (char === '"') {
|
|
95
|
+
if (cursor + 1 < buffer.length && buffer[cursor + 1] === '"') {
|
|
96
|
+
currentField += '"';
|
|
97
|
+
cursor += 2;
|
|
98
|
+
byteOffset += 2;
|
|
99
|
+
continue;
|
|
100
|
+
}
|
|
101
|
+
else {
|
|
102
|
+
inQuotes = false;
|
|
103
|
+
cursor++;
|
|
104
|
+
byteOffset++;
|
|
105
|
+
continue;
|
|
106
|
+
}
|
|
107
|
+
}
|
|
108
|
+
else {
|
|
109
|
+
currentField += char;
|
|
110
|
+
cursor++;
|
|
111
|
+
byteOffset++;
|
|
112
|
+
continue;
|
|
113
|
+
}
|
|
114
|
+
}
|
|
115
|
+
else {
|
|
116
|
+
if (char === '"') {
|
|
117
|
+
if (currentField.length === 0) {
|
|
118
|
+
inQuotes = true;
|
|
119
|
+
cursor++;
|
|
120
|
+
byteOffset++;
|
|
121
|
+
continue;
|
|
122
|
+
}
|
|
123
|
+
else {
|
|
124
|
+
throw new ParseError(`Unexpected quote inside unquoted field at byte offset ${byteOffset}`, {
|
|
125
|
+
file: mergedOptions.filePath,
|
|
126
|
+
row: rowNumber + 1,
|
|
127
|
+
column: currentRowFields.length + 1,
|
|
128
|
+
byteOffset,
|
|
129
|
+
});
|
|
130
|
+
}
|
|
131
|
+
}
|
|
132
|
+
// Check for delimiter
|
|
133
|
+
if (buffer.startsWith(activeDelimiter, cursor)) {
|
|
134
|
+
currentRowFields.push(currentField);
|
|
135
|
+
currentField = "";
|
|
136
|
+
cursor += delimLen;
|
|
137
|
+
byteOffset += delimLen;
|
|
138
|
+
continue;
|
|
139
|
+
}
|
|
140
|
+
// Check for CRLF / LF newline
|
|
141
|
+
if (char === "\r" || char === "\n") {
|
|
142
|
+
currentRowFields.push(currentField);
|
|
143
|
+
currentField = "";
|
|
144
|
+
if (char === "\r" && cursor + 1 < buffer.length && buffer[cursor + 1] === "\n") {
|
|
145
|
+
cursor += 2;
|
|
146
|
+
byteOffset += 2;
|
|
147
|
+
}
|
|
148
|
+
else {
|
|
149
|
+
cursor++;
|
|
150
|
+
byteOffset++;
|
|
151
|
+
}
|
|
152
|
+
// Skip entirely empty row if it's trailing whitespace
|
|
153
|
+
if (currentRowFields.length === 1 &&
|
|
154
|
+
currentRowFields[0]?.trim() === "" &&
|
|
155
|
+
cursor >= buffer.length) {
|
|
156
|
+
currentRowFields = [];
|
|
157
|
+
continue;
|
|
158
|
+
}
|
|
159
|
+
if (hasHeader && headers === null) {
|
|
160
|
+
headers = currentRowFields.map((h, i) => h.trim().length > 0 ? h.trim() : `col_${i + 1}`);
|
|
161
|
+
currentRowFields = [];
|
|
162
|
+
continue;
|
|
163
|
+
}
|
|
164
|
+
if (headers === null) {
|
|
165
|
+
headers = currentRowFields.map((_, i) => `col_${i + 1}`);
|
|
166
|
+
}
|
|
167
|
+
// Construct Row object
|
|
168
|
+
const rowObj = {};
|
|
169
|
+
for (let i = 0; i < headers.length; i++) {
|
|
170
|
+
const key = headers[i] || `col_${i + 1}`;
|
|
171
|
+
const val = currentRowFields[i] ?? "";
|
|
172
|
+
rowObj[key] = val;
|
|
173
|
+
}
|
|
174
|
+
rowNumber++;
|
|
175
|
+
rowsInCurrentBatch.push(rowObj);
|
|
176
|
+
currentRowFields = [];
|
|
177
|
+
if (rowsInCurrentBatch.length >= batchSize) {
|
|
178
|
+
yield {
|
|
179
|
+
rows: rowsInCurrentBatch,
|
|
180
|
+
offset: globalRowOffset,
|
|
181
|
+
};
|
|
182
|
+
globalRowOffset += rowsInCurrentBatch.length;
|
|
183
|
+
rowsInCurrentBatch = [];
|
|
184
|
+
}
|
|
185
|
+
continue;
|
|
186
|
+
}
|
|
187
|
+
currentField += char;
|
|
188
|
+
cursor++;
|
|
189
|
+
byteOffset++;
|
|
190
|
+
}
|
|
191
|
+
}
|
|
192
|
+
// Compact processed buffer to keep memory bounded
|
|
193
|
+
buffer = buffer.slice(cursor);
|
|
194
|
+
cursor = 0;
|
|
195
|
+
}
|
|
196
|
+
// Process remainder if any
|
|
197
|
+
if (inQuotes) {
|
|
198
|
+
throw new ParseError("Unclosed quote at end of CSV stream", {
|
|
199
|
+
file: mergedOptions.filePath,
|
|
200
|
+
row: rowNumber + 1,
|
|
201
|
+
byteOffset,
|
|
202
|
+
});
|
|
203
|
+
}
|
|
204
|
+
if (currentField.length > 0 || currentRowFields.length > 0) {
|
|
205
|
+
currentRowFields.push(currentField);
|
|
206
|
+
if (hasHeader && headers === null) {
|
|
207
|
+
headers = currentRowFields.map((h, i) => h.trim().length > 0 ? h.trim() : `col_${i + 1}`);
|
|
208
|
+
}
|
|
209
|
+
else {
|
|
210
|
+
if (headers === null) {
|
|
211
|
+
headers = currentRowFields.map((_, i) => `col_${i + 1}`);
|
|
212
|
+
}
|
|
213
|
+
const rowObj = {};
|
|
214
|
+
for (let i = 0; i < headers.length; i++) {
|
|
215
|
+
const key = headers[i] || `col_${i + 1}`;
|
|
216
|
+
const val = currentRowFields[i] ?? "";
|
|
217
|
+
rowObj[key] = val;
|
|
218
|
+
}
|
|
219
|
+
rowNumber++;
|
|
220
|
+
rowsInCurrentBatch.push(rowObj);
|
|
221
|
+
}
|
|
222
|
+
}
|
|
223
|
+
if (rowsInCurrentBatch.length > 0) {
|
|
224
|
+
yield {
|
|
225
|
+
rows: rowsInCurrentBatch,
|
|
226
|
+
offset: globalRowOffset,
|
|
227
|
+
};
|
|
228
|
+
}
|
|
229
|
+
}
|
|
230
|
+
async inspect(options) {
|
|
231
|
+
let rowCount = 0;
|
|
232
|
+
let columnNames = [];
|
|
233
|
+
for await (const batch of this.read({ ...options, batchSize: 5000 })) {
|
|
234
|
+
if (columnNames.length === 0 && batch.rows.length > 0 && batch.rows[0]) {
|
|
235
|
+
columnNames = Object.keys(batch.rows[0]);
|
|
236
|
+
}
|
|
237
|
+
rowCount += batch.rows.length;
|
|
238
|
+
}
|
|
239
|
+
return {
|
|
240
|
+
format: "CSV",
|
|
241
|
+
rowCount,
|
|
242
|
+
columnCount: columnNames.length,
|
|
243
|
+
columns: columnNames.map((name) => ({
|
|
244
|
+
name,
|
|
245
|
+
type: "string",
|
|
246
|
+
nullPercentage: 0,
|
|
247
|
+
})),
|
|
248
|
+
};
|
|
249
|
+
}
|
|
250
|
+
}
|
|
251
|
+
//# sourceMappingURL=csv.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"csv.js","sourceRoot":"","sources":["../../src/readers/csv.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAC;AAE3C,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAgB/C;;;GAGG;AACH,MAAM,OAAO,SAAS;IACZ,KAAK,CAAoB;IACzB,OAAO,CAAmB;IAElC,YAAY,KAAwB,EAAE,UAA4B,EAAE;QAClE,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,OAAO,EAAE,CAAC;QAC9B,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;YAC9B,IAAI,CAAC,OAAO,CAAC,QAAQ,GAAG,KAAK,CAAC;QAChC,CAAC;IACH,CAAC;IAEO,cAAc;QACpB,IAAI,OAAO,IAAI,CAAC,KAAK,KAAK,QAAQ,EAAE,CAAC;YACnC,OAAO,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACtC,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;IAED;;OAEG;IACI,MAAM,CAAC,eAAe,CAAC,MAAc;QAC1C,MAAM,UAAU,GAAG,CAAC,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC;QACzC,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAEjD,IAAI,aAAa,GAAG,GAAG,CAAC;QACxB,IAAI,QAAQ,GAAG,CAAC,CAAC,CAAC;QAElB,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;YACd,IAAI,QAAQ,GAAG,KAAK,CAAC;YACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC1C,MAAM,EAAE,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;gBACxB,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;oBACf,QAAQ,GAAG,CAAC,QAAQ,CAAC;gBACvB,CAAC;qBAAM,IAAI,EAAE,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;oBACpC,KAAK,EAAE,CAAC;gBACV,CAAC;YACH,CAAC;YACD,IAAI,KAAK,GAAG,QAAQ,EAAE,CAAC;gBACrB,QAAQ,GAAG,KAAK,CAAC;gBACjB,aAAa,GAAG,IAAI,CAAC;YACvB,CAAC;QACH,CAAC;QAED,OAAO,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC;IAC5C,CAAC;IAED,KAAK,CAAC,CAAC,IAAI,CAAC,OAAuB;QACjC,MAAM,aAAa,GAAqB;YACtC,GAAG,IAAI,CAAC,OAAO;YACf,GAAG,OAAO;SACX,CAAC;QACF,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,aAAa,CAAC,SAAS,IAAI,IAAI,CAAC,CAAC;QAC/D,MAAM,SAAS,GAAG,aAAa,CAAC,MAAM,KAAK,KAAK,CAAC;QACjD,IAAI,SAAS,GAAG,aAAa,CAAC,SAAS,CAAC;QAExC,MAAM,MAAM,GAAG,IAAI,CAAC,cAAc,EAAE,CAAC;QAErC,IAAI,MAAM,GAAG,EAAE,CAAC;QAChB,IAAI,OAAO,GAAoB,IAAI,CAAC;QACpC,IAAI,YAAY,GAAG,IAAI,CAAC;QACxB,IAAI,kBAAkB,GAAU,EAAE,CAAC;QACnC,IAAI,eAAe,GAAG,CAAC,CAAC;QACxB,IAAI,SAAS,GAAG,CAAC,CAAC;QAClB,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,IAAI,YAAY,GAAG,EAAE,CAAC;QACtB,IAAI,gBAAgB,GAAa,EAAE,CAAC;QACpC,IAAI,QAAQ,GAAG,KAAK,CAAC;QACrB,IAAI,MAAM,GAAG,CAAC,CAAC;QAEf,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YACjC,MAAM,QAAQ,GACZ,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAE,KAAgB,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YAEzE,IAAI,YAAY,EAAE,CAAC;gBACjB,YAAY,GAAG,KAAK,CAAC;gBACrB,8BAA8B;gBAC9B,IAAI,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,KAAK,MAAM,EAAE,CAAC;oBACtC,MAAM,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;gBAC7B,CAAC;qBAAM,CAAC;oBACN,MAAM,GAAG,QAAQ,CAAC;gBACpB,CAAC;gBAED,yCAAyC;gBACzC,IAAI,CAAC,SAAS,EAAE,CAAC;oBACf,SAAS,GAAG,SAAS,CAAC,eAAe,CAAC,MAAM,CAAC,CAAC;gBAChD,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,MAAM,IAAI,QAAQ,CAAC;YACrB,CAAC;YAED,MAAM,eAAe,GAAG,SAAS,IAAI,GAAG,CAAC;YACzC,MAAM,QAAQ,GAAG,eAAe,CAAC,MAAM,CAAC;YAExC,OAAO,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;gBAC9B,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;gBAE5B,IAAI,QAAQ,EAAE,CAAC;oBACb,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACjB,IAAI,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;4BAC7D,YAAY,IAAI,GAAG,CAAC;4BACpB,MAAM,IAAI,CAAC,CAAC;4BACZ,UAAU,IAAI,CAAC,CAAC;4BAChB,SAAS;wBACX,CAAC;6BAAM,CAAC;4BACN,QAAQ,GAAG,KAAK,CAAC;4BACjB,MAAM,EAAE,CAAC;4BACT,UAAU,EAAE,CAAC;4BACb,SAAS;wBACX,CAAC;oBACH,CAAC;yBAAM,CAAC;wBACN,YAAY,IAAI,IAAI,CAAC;wBACrB,MAAM,EAAE,CAAC;wBACT,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;gBACH,CAAC;qBAAM,CAAC;oBACN,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACjB,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;4BAC9B,QAAQ,GAAG,IAAI,CAAC;4BAChB,MAAM,EAAE,CAAC;4BACT,UAAU,EAAE,CAAC;4BACb,SAAS;wBACX,CAAC;6BAAM,CAAC;4BACN,MAAM,IAAI,UAAU,CAClB,yDAAyD,UAAU,EAAE,EACrE;gCACE,IAAI,EAAE,aAAa,CAAC,QAAQ;gCAC5B,GAAG,EAAE,SAAS,GAAG,CAAC;gCAClB,MAAM,EAAE,gBAAgB,CAAC,MAAM,GAAG,CAAC;gCACnC,UAAU;6BACX,CACF,CAAC;wBACJ,CAAC;oBACH,CAAC;oBAED,sBAAsB;oBACtB,IAAI,MAAM,CAAC,UAAU,CAAC,eAAe,EAAE,MAAM,CAAC,EAAE,CAAC;wBAC/C,gBAAgB,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;wBACpC,YAAY,GAAG,EAAE,CAAC;wBAClB,MAAM,IAAI,QAAQ,CAAC;wBACnB,UAAU,IAAI,QAAQ,CAAC;wBACvB,SAAS;oBACX,CAAC;oBAED,8BAA8B;oBAC9B,IAAI,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;wBACnC,gBAAgB,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;wBACpC,YAAY,GAAG,EAAE,CAAC;wBAElB,IAAI,IAAI,KAAK,IAAI,IAAI,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;4BAC/E,MAAM,IAAI,CAAC,CAAC;4BACZ,UAAU,IAAI,CAAC,CAAC;wBAClB,CAAC;6BAAM,CAAC;4BACN,MAAM,EAAE,CAAC;4BACT,UAAU,EAAE,CAAC;wBACf,CAAC;wBAED,sDAAsD;wBACtD,IACE,gBAAgB,CAAC,MAAM,KAAK,CAAC;4BAC7B,gBAAgB,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE;4BAClC,MAAM,IAAI,MAAM,CAAC,MAAM,EACvB,CAAC;4BACD,gBAAgB,GAAG,EAAE,CAAC;4BACtB,SAAS;wBACX,CAAC;wBAED,IAAI,SAAS,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;4BAClC,OAAO,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACtC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAChD,CAAC;4BACF,gBAAgB,GAAG,EAAE,CAAC;4BACtB,SAAS;wBACX,CAAC;wBAED,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;4BACrB,OAAO,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;wBAC3D,CAAC;wBAED,uBAAuB;wBACvB,MAAM,MAAM,GAAQ,EAAE,CAAC;wBACvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;4BACxC,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;4BACzC,MAAM,GAAG,GAAG,gBAAgB,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;4BACtC,MAAM,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC;wBACpB,CAAC;wBAED,SAAS,EAAE,CAAC;wBACZ,kBAAkB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;wBAChC,gBAAgB,GAAG,EAAE,CAAC;wBAEtB,IAAI,kBAAkB,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;4BAC3C,MAAM;gCACJ,IAAI,EAAE,kBAAkB;gCACxB,MAAM,EAAE,eAAe;6BACxB,CAAC;4BACF,eAAe,IAAI,kBAAkB,CAAC,MAAM,CAAC;4BAC7C,kBAAkB,GAAG,EAAE,CAAC;wBAC1B,CAAC;wBACD,SAAS;oBACX,CAAC;oBAED,YAAY,IAAI,IAAI,CAAC;oBACrB,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;gBACf,CAAC;YACH,CAAC;YAED,kDAAkD;YAClD,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;YAC9B,MAAM,GAAG,CAAC,CAAC;QACb,CAAC;QAED,2BAA2B;QAC3B,IAAI,QAAQ,EAAE,CAAC;YACb,MAAM,IAAI,UAAU,CAAC,qCAAqC,EAAE;gBAC1D,IAAI,EAAE,aAAa,CAAC,QAAQ;gBAC5B,GAAG,EAAE,SAAS,GAAG,CAAC;gBAClB,UAAU;aACX,CAAC,CAAC;QACL,CAAC;QAED,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,IAAI,gBAAgB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC3D,gBAAgB,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;YAEpC,IAAI,SAAS,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;gBAClC,OAAO,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACtC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAChD,CAAC;YACJ,CAAC;iBAAM,CAAC;gBACN,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;oBACrB,OAAO,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;gBAC3D,CAAC;gBAED,MAAM,MAAM,GAAQ,EAAE,CAAC;gBACvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACxC,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;oBACzC,MAAM,GAAG,GAAG,gBAAgB,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;oBACtC,MAAM,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC;gBACpB,CAAC;gBAED,SAAS,EAAE,CAAC;gBACZ,kBAAkB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAClC,CAAC;QACH,CAAC;QAED,IAAI,kBAAkB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAClC,MAAM;gBACJ,IAAI,EAAE,kBAAkB;gBACxB,MAAM,EAAE,eAAe;aACxB,CAAC;QACJ,CAAC;IACH,CAAC;IAED,KAAK,CAAC,OAAO,CAAC,OAAuB;QACnC,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,IAAI,WAAW,GAAa,EAAE,CAAC;QAE/B,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;YACrE,IAAI,WAAW,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;gBACvE,WAAW,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3C,CAAC;YACD,QAAQ,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC;QAChC,CAAC;QAED,OAAO;YACL,MAAM,EAAE,KAAK;YACb,QAAQ;YACR,WAAW,EAAE,WAAW,CAAC,MAAM;YAC/B,OAAO,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBAClC,IAAI;gBACJ,IAAI,EAAE,QAAQ;gBACd,cAAc,EAAE,CAAC;aAClB,CAAC,CAAC;SACJ,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
import type { Readable } from "node:stream";
|
|
2
|
+
import type { ReaderOptions, TabularReader } from "../core/types.js";
|
|
3
|
+
export * from "./csv.js";
|
|
4
|
+
export * from "./json.js";
|
|
5
|
+
export * from "./jsonl.js";
|
|
6
|
+
export * from "./xlsx.js";
|
|
7
|
+
/**
|
|
8
|
+
* Infers format name from file path or extension.
|
|
9
|
+
*/
|
|
10
|
+
export declare function inferFormatFromPath(filePath: string): string | null;
|
|
11
|
+
/**
|
|
12
|
+
* Creates an appropriate TabularReader for input and format.
|
|
13
|
+
*/
|
|
14
|
+
export declare function createReader(input: Readable | string, options?: ReaderOptions & {
|
|
15
|
+
format?: string;
|
|
16
|
+
}): TabularReader;
|
|
17
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/readers/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,KAAK,EAAiB,aAAa,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AAMpF,cAAc,UAAU,CAAC;AACzB,cAAc,WAAW,CAAC;AAC1B,cAAc,YAAY,CAAC;AAC3B,cAAc,WAAW,CAAC;AAqC1B;;GAEG;AACH,wBAAgB,mBAAmB,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAQnE;AAED;;GAEG;AACH,wBAAgB,YAAY,CAC1B,KAAK,EAAE,QAAQ,GAAG,MAAM,EACxB,OAAO,GAAE,aAAa,GAAG;IAAE,MAAM,CAAC,EAAE,MAAM,CAAA;CAAO,GAChD,aAAa,CAoBf"}
|
|
@@ -0,0 +1,74 @@
|
|
|
1
|
+
import { InvalidArgumentError } from "../core/errors.js";
|
|
2
|
+
import { CSVReader } from "./csv.js";
|
|
3
|
+
import { JSONReader } from "./json.js";
|
|
4
|
+
import { JSONLReader } from "./jsonl.js";
|
|
5
|
+
import { XLSXReader } from "./xlsx.js";
|
|
6
|
+
export * from "./csv.js";
|
|
7
|
+
export * from "./json.js";
|
|
8
|
+
export * from "./jsonl.js";
|
|
9
|
+
export * from "./xlsx.js";
|
|
10
|
+
const adapters = {
|
|
11
|
+
csv: {
|
|
12
|
+
name: "CSV",
|
|
13
|
+
extensions: [".csv", ".tsv", ".txt"],
|
|
14
|
+
createReader: (input, options) => new CSVReader(input, options),
|
|
15
|
+
createWriter: () => {
|
|
16
|
+
throw new Error("Writer adapter called from reader");
|
|
17
|
+
},
|
|
18
|
+
},
|
|
19
|
+
json: {
|
|
20
|
+
name: "JSON",
|
|
21
|
+
extensions: [".json"],
|
|
22
|
+
createReader: (input, options) => new JSONReader(input, options),
|
|
23
|
+
createWriter: () => {
|
|
24
|
+
throw new Error("Writer adapter called from reader");
|
|
25
|
+
},
|
|
26
|
+
},
|
|
27
|
+
jsonl: {
|
|
28
|
+
name: "JSONL",
|
|
29
|
+
extensions: [".jsonl", ".ndjson", ".ldjson"],
|
|
30
|
+
createReader: (input, options) => new JSONLReader(input, options),
|
|
31
|
+
createWriter: () => {
|
|
32
|
+
throw new Error("Writer adapter called from reader");
|
|
33
|
+
},
|
|
34
|
+
},
|
|
35
|
+
xlsx: {
|
|
36
|
+
name: "XLSX",
|
|
37
|
+
extensions: [".xlsx", ".xlsm"],
|
|
38
|
+
createReader: (input, options) => new XLSXReader(input, options),
|
|
39
|
+
createWriter: () => {
|
|
40
|
+
throw new Error("Writer adapter called from reader");
|
|
41
|
+
},
|
|
42
|
+
},
|
|
43
|
+
};
|
|
44
|
+
/**
|
|
45
|
+
* Infers format name from file path or extension.
|
|
46
|
+
*/
|
|
47
|
+
export function inferFormatFromPath(filePath) {
|
|
48
|
+
const cleanPath = filePath.toLowerCase().replace(/\.gz$/, "");
|
|
49
|
+
for (const [format, adapter] of Object.entries(adapters)) {
|
|
50
|
+
if (adapter.extensions.some((ext) => cleanPath.endsWith(ext))) {
|
|
51
|
+
return format;
|
|
52
|
+
}
|
|
53
|
+
}
|
|
54
|
+
return null;
|
|
55
|
+
}
|
|
56
|
+
/**
|
|
57
|
+
* Creates an appropriate TabularReader for input and format.
|
|
58
|
+
*/
|
|
59
|
+
export function createReader(input, options = {}) {
|
|
60
|
+
let format = options.format?.toLowerCase();
|
|
61
|
+
if (!format && typeof input === "string" && input !== "-") {
|
|
62
|
+
format = inferFormatFromPath(input) ?? undefined;
|
|
63
|
+
}
|
|
64
|
+
if (!format) {
|
|
65
|
+
// Default to CSV if cannot infer
|
|
66
|
+
format = "csv";
|
|
67
|
+
}
|
|
68
|
+
const adapter = adapters[format];
|
|
69
|
+
if (!adapter) {
|
|
70
|
+
throw new InvalidArgumentError(`Unsupported input format: "${format}". Supported formats are: ${Object.keys(adapters).join(", ")}`);
|
|
71
|
+
}
|
|
72
|
+
return adapter.createReader(input, options);
|
|
73
|
+
}
|
|
74
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/readers/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,oBAAoB,EAAE,MAAM,mBAAmB,CAAC;AAEzD,OAAO,EAAE,SAAS,EAAE,MAAM,UAAU,CAAC;AACrC,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AACzC,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAEvC,cAAc,UAAU,CAAC;AACzB,cAAc,WAAW,CAAC;AAC1B,cAAc,YAAY,CAAC;AAC3B,cAAc,WAAW,CAAC;AAE1B,MAAM,QAAQ,GAAkC;IAC9C,GAAG,EAAE;QACH,IAAI,EAAE,KAAK;QACX,UAAU,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC;QACpC,YAAY,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,SAAS,CAAC,KAAK,EAAE,OAAO,CAAC;QAC/D,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;KACF;IACD,IAAI,EAAE;QACJ,IAAI,EAAE,MAAM;QACZ,UAAU,EAAE,CAAC,OAAO,CAAC;QACrB,YAAY,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC;QAChE,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;KACF;IACD,KAAK,EAAE;QACL,IAAI,EAAE,OAAO;QACb,UAAU,EAAE,CAAC,QAAQ,EAAE,SAAS,EAAE,SAAS,CAAC;QAC5C,YAAY,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC;QACjE,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;KACF;IACD,IAAI,EAAE;QACJ,IAAI,EAAE,MAAM;QACZ,UAAU,EAAE,CAAC,OAAO,EAAE,OAAO,CAAC;QAC9B,YAAY,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC;QAChE,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;KACF;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,UAAU,mBAAmB,CAAC,QAAgB;IAClD,MAAM,SAAS,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IAC9D,KAAK,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;QACzD,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YAC9D,OAAO,MAAM,CAAC;QAChB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,YAAY,CAC1B,KAAwB,EACxB,UAA+C,EAAE;IAEjD,IAAI,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,WAAW,EAAE,CAAC;IAE3C,IAAI,CAAC,MAAM,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,GAAG,EAAE,CAAC;QAC1D,MAAM,GAAG,mBAAmB,CAAC,KAAK,CAAC,IAAI,SAAS,CAAC;IACnD,CAAC;IAED,IAAI,CAAC,MAAM,EAAE,CAAC;QACZ,iCAAiC;QACjC,MAAM,GAAG,KAAK,CAAC;IACjB,CAAC;IAED,MAAM,OAAO,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC;IACjC,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,MAAM,IAAI,oBAAoB,CAC5B,8BAA8B,MAAM,6BAA6B,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACpG,CAAC;IACJ,CAAC;IAED,OAAO,OAAO,CAAC,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;AAC9C,CAAC"}
|
|
@@ -0,0 +1,20 @@
|
|
|
1
|
+
import type { Readable } from "node:stream";
|
|
2
|
+
import type { DataStream, InspectionMetadata, ReaderOptions, TabularReader } from "../core/types.js";
|
|
3
|
+
export interface JSONReaderOptions extends ReaderOptions {
|
|
4
|
+
path?: string;
|
|
5
|
+
filePath?: string;
|
|
6
|
+
}
|
|
7
|
+
/**
|
|
8
|
+
* Incremental streaming JSON parser for top-level arrays [ {...}, ... ]
|
|
9
|
+
* and nested array paths (e.g. data.results).
|
|
10
|
+
* Never loads the full JSON file into memory.
|
|
11
|
+
*/
|
|
12
|
+
export declare class JSONReader implements TabularReader {
|
|
13
|
+
private input;
|
|
14
|
+
private options;
|
|
15
|
+
constructor(input: Readable | string, options?: JSONReaderOptions);
|
|
16
|
+
private getInputStream;
|
|
17
|
+
read(options?: ReaderOptions): DataStream;
|
|
18
|
+
inspect(options?: ReaderOptions): Promise<InspectionMetadata>;
|
|
19
|
+
}
|
|
20
|
+
//# sourceMappingURL=json.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.d.ts","sourceRoot":"","sources":["../../src/readers/json.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,KAAK,EAEV,UAAU,EACV,kBAAkB,EAClB,aAAa,EAEb,aAAa,EACd,MAAM,kBAAkB,CAAC;AAE1B,MAAM,WAAW,iBAAkB,SAAQ,aAAa;IACtD,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED;;;;GAIG;AACH,qBAAa,UAAW,YAAW,aAAa;IAC9C,OAAO,CAAC,KAAK,CAAoB;IACjC,OAAO,CAAC,OAAO,CAAoB;gBAEvB,KAAK,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,iBAAsB;IAQrE,OAAO,CAAC,cAAc;IAOf,IAAI,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,UAAU;IAgM1C,OAAO,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,kBAAkB,CAAC;CAsBpE"}
|
|
@@ -0,0 +1,219 @@
|
|
|
1
|
+
import { createReadStream } from "node:fs";
|
|
2
|
+
import { ParseError } from "../core/errors.js";
|
|
3
|
+
/**
|
|
4
|
+
* Incremental streaming JSON parser for top-level arrays [ {...}, ... ]
|
|
5
|
+
* and nested array paths (e.g. data.results).
|
|
6
|
+
* Never loads the full JSON file into memory.
|
|
7
|
+
*/
|
|
8
|
+
export class JSONReader {
|
|
9
|
+
input;
|
|
10
|
+
options;
|
|
11
|
+
constructor(input, options = {}) {
|
|
12
|
+
this.input = input;
|
|
13
|
+
this.options = { ...options };
|
|
14
|
+
if (typeof input === "string") {
|
|
15
|
+
this.options.filePath = input;
|
|
16
|
+
}
|
|
17
|
+
}
|
|
18
|
+
getInputStream() {
|
|
19
|
+
if (typeof this.input === "string") {
|
|
20
|
+
return createReadStream(this.input);
|
|
21
|
+
}
|
|
22
|
+
return this.input;
|
|
23
|
+
}
|
|
24
|
+
async *read(options) {
|
|
25
|
+
const mergedOptions = {
|
|
26
|
+
...this.options,
|
|
27
|
+
...options,
|
|
28
|
+
};
|
|
29
|
+
const batchSize = Math.max(1, mergedOptions.batchSize ?? 1000);
|
|
30
|
+
const targetPath = mergedOptions.path; // e.g. "data.results"
|
|
31
|
+
const stream = this.getInputStream();
|
|
32
|
+
let buffer = "";
|
|
33
|
+
let cursor = 0;
|
|
34
|
+
let byteOffset = 0;
|
|
35
|
+
let foundTargetArray = !targetPath; // if no path, we look for first '['
|
|
36
|
+
let currentPath = [];
|
|
37
|
+
// State machine for array item scanning
|
|
38
|
+
let inArray = false;
|
|
39
|
+
let inString = false;
|
|
40
|
+
let isEscaped = false;
|
|
41
|
+
let objectDepth = 0;
|
|
42
|
+
let objectStartIndex = -1;
|
|
43
|
+
let rowsInCurrentBatch = [];
|
|
44
|
+
let globalOffset = 0;
|
|
45
|
+
let rowIndex = 0;
|
|
46
|
+
// Helper for path navigation if targetPath is set
|
|
47
|
+
const targetPathParts = targetPath ? targetPath.split(".") : [];
|
|
48
|
+
for await (const chunk of stream) {
|
|
49
|
+
const chunkStr = typeof chunk === "string" ? chunk : chunk.toString("utf8");
|
|
50
|
+
buffer += chunkStr;
|
|
51
|
+
while (cursor < buffer.length) {
|
|
52
|
+
const char = buffer[cursor];
|
|
53
|
+
if (isEscaped) {
|
|
54
|
+
isEscaped = false;
|
|
55
|
+
cursor++;
|
|
56
|
+
byteOffset++;
|
|
57
|
+
continue;
|
|
58
|
+
}
|
|
59
|
+
if (char === "\\") {
|
|
60
|
+
isEscaped = true;
|
|
61
|
+
cursor++;
|
|
62
|
+
byteOffset++;
|
|
63
|
+
continue;
|
|
64
|
+
}
|
|
65
|
+
if (char === '"') {
|
|
66
|
+
inString = !inString;
|
|
67
|
+
cursor++;
|
|
68
|
+
byteOffset++;
|
|
69
|
+
continue;
|
|
70
|
+
}
|
|
71
|
+
if (inString) {
|
|
72
|
+
cursor++;
|
|
73
|
+
byteOffset++;
|
|
74
|
+
continue;
|
|
75
|
+
}
|
|
76
|
+
// We are outside of any string literal
|
|
77
|
+
if (!foundTargetArray) {
|
|
78
|
+
// If we need to find target path (e.g. data.results)
|
|
79
|
+
// Look for keys and match path parts
|
|
80
|
+
if (char === "{") {
|
|
81
|
+
cursor++;
|
|
82
|
+
byteOffset++;
|
|
83
|
+
continue;
|
|
84
|
+
}
|
|
85
|
+
else if (char === "[") {
|
|
86
|
+
// Check if current path matches targetPath
|
|
87
|
+
// If path matched or we hit array at root
|
|
88
|
+
foundTargetArray = true;
|
|
89
|
+
inArray = true;
|
|
90
|
+
cursor++;
|
|
91
|
+
byteOffset++;
|
|
92
|
+
continue;
|
|
93
|
+
}
|
|
94
|
+
else if (char === ":") {
|
|
95
|
+
// Scan backwards from ':' to find previous string key
|
|
96
|
+
let keyEnd = cursor - 1;
|
|
97
|
+
while (keyEnd >= 0 && /\s/.test(buffer[keyEnd]))
|
|
98
|
+
keyEnd--;
|
|
99
|
+
if (keyEnd >= 0 && buffer[keyEnd] === '"') {
|
|
100
|
+
let keyStart = keyEnd - 1;
|
|
101
|
+
while (keyStart >= 0 && buffer[keyStart] !== '"')
|
|
102
|
+
keyStart--;
|
|
103
|
+
if (keyStart >= 0) {
|
|
104
|
+
const keyName = buffer.slice(keyStart + 1, keyEnd);
|
|
105
|
+
currentPath.push(keyName);
|
|
106
|
+
if (currentPath.join(".") === targetPath ||
|
|
107
|
+
keyName === targetPathParts[targetPathParts.length - 1]) {
|
|
108
|
+
// Next non-whitespace should be '['
|
|
109
|
+
// Let it continue to find '['
|
|
110
|
+
}
|
|
111
|
+
}
|
|
112
|
+
}
|
|
113
|
+
cursor++;
|
|
114
|
+
byteOffset++;
|
|
115
|
+
continue;
|
|
116
|
+
}
|
|
117
|
+
cursor++;
|
|
118
|
+
byteOffset++;
|
|
119
|
+
continue;
|
|
120
|
+
}
|
|
121
|
+
// Now we are streaming inside the target array
|
|
122
|
+
if (!inArray) {
|
|
123
|
+
if (char === "[") {
|
|
124
|
+
inArray = true;
|
|
125
|
+
cursor++;
|
|
126
|
+
byteOffset++;
|
|
127
|
+
continue;
|
|
128
|
+
}
|
|
129
|
+
}
|
|
130
|
+
else {
|
|
131
|
+
// Inside array: looking for objects '{ ... }'
|
|
132
|
+
if (char === "{") {
|
|
133
|
+
if (objectDepth === 0) {
|
|
134
|
+
objectStartIndex = cursor;
|
|
135
|
+
}
|
|
136
|
+
objectDepth++;
|
|
137
|
+
}
|
|
138
|
+
else if (char === "}") {
|
|
139
|
+
objectDepth--;
|
|
140
|
+
if (objectDepth === 0 && objectStartIndex !== -1) {
|
|
141
|
+
// We have captured a complete JSON object!
|
|
142
|
+
const objectStr = buffer.slice(objectStartIndex, cursor + 1);
|
|
143
|
+
let parsedRow;
|
|
144
|
+
try {
|
|
145
|
+
parsedRow = JSON.parse(objectStr);
|
|
146
|
+
}
|
|
147
|
+
catch (err) {
|
|
148
|
+
throw new ParseError(`Malformed JSON object in array: ${err.message}`, {
|
|
149
|
+
file: mergedOptions.filePath,
|
|
150
|
+
row: rowIndex + 1,
|
|
151
|
+
byteOffset,
|
|
152
|
+
});
|
|
153
|
+
}
|
|
154
|
+
if (typeof parsedRow === "object" &&
|
|
155
|
+
parsedRow !== null &&
|
|
156
|
+
!Array.isArray(parsedRow)) {
|
|
157
|
+
rowsInCurrentBatch.push(parsedRow);
|
|
158
|
+
rowIndex++;
|
|
159
|
+
if (rowsInCurrentBatch.length >= batchSize) {
|
|
160
|
+
yield {
|
|
161
|
+
rows: rowsInCurrentBatch,
|
|
162
|
+
offset: globalOffset,
|
|
163
|
+
};
|
|
164
|
+
globalOffset += rowsInCurrentBatch.length;
|
|
165
|
+
rowsInCurrentBatch = [];
|
|
166
|
+
}
|
|
167
|
+
}
|
|
168
|
+
objectStartIndex = -1;
|
|
169
|
+
}
|
|
170
|
+
}
|
|
171
|
+
else if (char === "]" && objectDepth === 0) {
|
|
172
|
+
// End of array reached
|
|
173
|
+
inArray = false;
|
|
174
|
+
break;
|
|
175
|
+
}
|
|
176
|
+
cursor++;
|
|
177
|
+
byteOffset++;
|
|
178
|
+
}
|
|
179
|
+
}
|
|
180
|
+
// Compact buffer while preserving active object if in the middle of parsing
|
|
181
|
+
if (objectStartIndex !== -1) {
|
|
182
|
+
buffer = buffer.slice(objectStartIndex);
|
|
183
|
+
cursor = cursor - objectStartIndex;
|
|
184
|
+
objectStartIndex = 0;
|
|
185
|
+
}
|
|
186
|
+
else {
|
|
187
|
+
buffer = buffer.slice(cursor);
|
|
188
|
+
cursor = 0;
|
|
189
|
+
}
|
|
190
|
+
}
|
|
191
|
+
if (rowsInCurrentBatch.length > 0) {
|
|
192
|
+
yield {
|
|
193
|
+
rows: rowsInCurrentBatch,
|
|
194
|
+
offset: globalOffset,
|
|
195
|
+
};
|
|
196
|
+
}
|
|
197
|
+
}
|
|
198
|
+
async inspect(options) {
|
|
199
|
+
let rowCount = 0;
|
|
200
|
+
let columnNames = [];
|
|
201
|
+
for await (const batch of this.read({ ...options, batchSize: 5000 })) {
|
|
202
|
+
if (columnNames.length === 0 && batch.rows.length > 0 && batch.rows[0]) {
|
|
203
|
+
columnNames = Object.keys(batch.rows[0]);
|
|
204
|
+
}
|
|
205
|
+
rowCount += batch.rows.length;
|
|
206
|
+
}
|
|
207
|
+
return {
|
|
208
|
+
format: "JSON",
|
|
209
|
+
rowCount,
|
|
210
|
+
columnCount: columnNames.length,
|
|
211
|
+
columns: columnNames.map((name) => ({
|
|
212
|
+
name,
|
|
213
|
+
type: "string",
|
|
214
|
+
nullPercentage: 0,
|
|
215
|
+
})),
|
|
216
|
+
};
|
|
217
|
+
}
|
|
218
|
+
}
|
|
219
|
+
//# sourceMappingURL=json.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.js","sourceRoot":"","sources":["../../src/readers/json.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAC;AAE3C,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAe/C;;;;GAIG;AACH,MAAM,OAAO,UAAU;IACb,KAAK,CAAoB;IACzB,OAAO,CAAoB;IAEnC,YAAY,KAAwB,EAAE,UAA6B,EAAE;QACnE,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,OAAO,EAAE,CAAC;QAC9B,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;YAC9B,IAAI,CAAC,OAAO,CAAC,QAAQ,GAAG,KAAK,CAAC;QAChC,CAAC;IACH,CAAC;IAEO,cAAc;QACpB,IAAI,OAAO,IAAI,CAAC,KAAK,KAAK,QAAQ,EAAE,CAAC;YACnC,OAAO,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACtC,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;IAED,KAAK,CAAC,CAAC,IAAI,CAAC,OAAuB;QACjC,MAAM,aAAa,GAAsB;YACvC,GAAG,IAAI,CAAC,OAAO;YACf,GAAG,OAAO;SACX,CAAC;QACF,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,aAAa,CAAC,SAAS,IAAI,IAAI,CAAC,CAAC;QAC/D,MAAM,UAAU,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,sBAAsB;QAC7D,MAAM,MAAM,GAAG,IAAI,CAAC,cAAc,EAAE,CAAC;QAErC,IAAI,MAAM,GAAG,EAAE,CAAC;QAChB,IAAI,MAAM,GAAG,CAAC,CAAC;QACf,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,IAAI,gBAAgB,GAAG,CAAC,UAAU,CAAC,CAAC,oCAAoC;QACxE,IAAI,WAAW,GAAa,EAAE,CAAC;QAE/B,wCAAwC;QACxC,IAAI,OAAO,GAAG,KAAK,CAAC;QACpB,IAAI,QAAQ,GAAG,KAAK,CAAC;QACrB,IAAI,SAAS,GAAG,KAAK,CAAC;QACtB,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,IAAI,gBAAgB,GAAG,CAAC,CAAC,CAAC;QAE1B,IAAI,kBAAkB,GAAU,EAAE,CAAC;QACnC,IAAI,YAAY,GAAG,CAAC,CAAC;QACrB,IAAI,QAAQ,GAAG,CAAC,CAAC;QAEjB,kDAAkD;QAClD,MAAM,eAAe,GAAG,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAEhE,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YACjC,MAAM,QAAQ,GACZ,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAE,KAAgB,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YACzE,MAAM,IAAI,QAAQ,CAAC;YAEnB,OAAO,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;gBAC9B,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;gBAE5B,IAAI,SAAS,EAAE,CAAC;oBACd,SAAS,GAAG,KAAK,CAAC;oBAClB,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;oBACb,SAAS;gBACX,CAAC;gBAED,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;oBAClB,SAAS,GAAG,IAAI,CAAC;oBACjB,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;oBACb,SAAS;gBACX,CAAC;gBAED,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;oBACjB,QAAQ,GAAG,CAAC,QAAQ,CAAC;oBACrB,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;oBACb,SAAS;gBACX,CAAC;gBAED,IAAI,QAAQ,EAAE,CAAC;oBACb,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;oBACb,SAAS;gBACX,CAAC;gBAED,uCAAuC;gBACvC,IAAI,CAAC,gBAAgB,EAAE,CAAC;oBACtB,qDAAqD;oBACrD,qCAAqC;oBACrC,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACjB,MAAM,EAAE,CAAC;wBACT,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;yBAAM,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACxB,2CAA2C;wBAC3C,0CAA0C;wBAC1C,gBAAgB,GAAG,IAAI,CAAC;wBACxB,OAAO,GAAG,IAAI,CAAC;wBACf,MAAM,EAAE,CAAC;wBACT,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;yBAAM,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACxB,sDAAsD;wBACtD,IAAI,MAAM,GAAG,MAAM,GAAG,CAAC,CAAC;wBACxB,OAAO,MAAM,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAE,CAAC;4BAAE,MAAM,EAAE,CAAC;wBAC3D,IAAI,MAAM,IAAI,CAAC,IAAI,MAAM,CAAC,MAAM,CAAC,KAAK,GAAG,EAAE,CAAC;4BAC1C,IAAI,QAAQ,GAAG,MAAM,GAAG,CAAC,CAAC;4BAC1B,OAAO,QAAQ,IAAI,CAAC,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,GAAG;gCAAE,QAAQ,EAAE,CAAC;4BAC7D,IAAI,QAAQ,IAAI,CAAC,EAAE,CAAC;gCAClB,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,QAAQ,GAAG,CAAC,EAAE,MAAM,CAAC,CAAC;gCACnD,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;gCAC1B,IACE,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,UAAU;oCACpC,OAAO,KAAK,eAAe,CAAC,eAAe,CAAC,MAAM,GAAG,CAAC,CAAC,EACvD,CAAC;oCACD,oCAAoC;oCACpC,8BAA8B;gCAChC,CAAC;4BACH,CAAC;wBACH,CAAC;wBACD,MAAM,EAAE,CAAC;wBACT,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;oBACD,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;oBACb,SAAS;gBACX,CAAC;gBAED,+CAA+C;gBAC/C,IAAI,CAAC,OAAO,EAAE,CAAC;oBACb,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACjB,OAAO,GAAG,IAAI,CAAC;wBACf,MAAM,EAAE,CAAC;wBACT,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;gBACH,CAAC;qBAAM,CAAC;oBACN,8CAA8C;oBAC9C,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACjB,IAAI,WAAW,KAAK,CAAC,EAAE,CAAC;4BACtB,gBAAgB,GAAG,MAAM,CAAC;wBAC5B,CAAC;wBACD,WAAW,EAAE,CAAC;oBAChB,CAAC;yBAAM,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;wBACxB,WAAW,EAAE,CAAC;wBACd,IAAI,WAAW,KAAK,CAAC,IAAI,gBAAgB,KAAK,CAAC,CAAC,EAAE,CAAC;4BACjD,2CAA2C;4BAC3C,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC,gBAAgB,EAAE,MAAM,GAAG,CAAC,CAAC,CAAC;4BAC7D,IAAI,SAAkB,CAAC;4BACvB,IAAI,CAAC;gCACH,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;4BACpC,CAAC;4BAAC,OAAO,GAAG,EAAE,CAAC;gCACb,MAAM,IAAI,UAAU,CAClB,mCAAoC,GAAa,CAAC,OAAO,EAAE,EAC3D;oCACE,IAAI,EAAE,aAAa,CAAC,QAAQ;oCAC5B,GAAG,EAAE,QAAQ,GAAG,CAAC;oCACjB,UAAU;iCACX,CACF,CAAC;4BACJ,CAAC;4BAED,IACE,OAAO,SAAS,KAAK,QAAQ;gCAC7B,SAAS,KAAK,IAAI;gCAClB,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,EACzB,CAAC;gCACD,kBAAkB,CAAC,IAAI,CAAC,SAAgB,CAAC,CAAC;gCAC1C,QAAQ,EAAE,CAAC;gCAEX,IAAI,kBAAkB,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;oCAC3C,MAAM;wCACJ,IAAI,EAAE,kBAAkB;wCACxB,MAAM,EAAE,YAAY;qCACrB,CAAC;oCACF,YAAY,IAAI,kBAAkB,CAAC,MAAM,CAAC;oCAC1C,kBAAkB,GAAG,EAAE,CAAC;gCAC1B,CAAC;4BACH,CAAC;4BAED,gBAAgB,GAAG,CAAC,CAAC,CAAC;wBACxB,CAAC;oBACH,CAAC;yBAAM,IAAI,IAAI,KAAK,GAAG,IAAI,WAAW,KAAK,CAAC,EAAE,CAAC;wBAC7C,uBAAuB;wBACvB,OAAO,GAAG,KAAK,CAAC;wBAChB,MAAM;oBACR,CAAC;oBAED,MAAM,EAAE,CAAC;oBACT,UAAU,EAAE,CAAC;gBACf,CAAC;YACH,CAAC;YAED,4EAA4E;YAC5E,IAAI,gBAAgB,KAAK,CAAC,CAAC,EAAE,CAAC;gBAC5B,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,gBAAgB,CAAC,CAAC;gBACxC,MAAM,GAAG,MAAM,GAAG,gBAAgB,CAAC;gBACnC,gBAAgB,GAAG,CAAC,CAAC;YACvB,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;gBAC9B,MAAM,GAAG,CAAC,CAAC;YACb,CAAC;QACH,CAAC;QAED,IAAI,kBAAkB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAClC,MAAM;gBACJ,IAAI,EAAE,kBAAkB;gBACxB,MAAM,EAAE,YAAY;aACrB,CAAC;QACJ,CAAC;IACH,CAAC;IAED,KAAK,CAAC,OAAO,CAAC,OAAuB;QACnC,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,IAAI,WAAW,GAAa,EAAE,CAAC;QAE/B,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;YACrE,IAAI,WAAW,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;gBACvE,WAAW,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3C,CAAC;YACD,QAAQ,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC;QAChC,CAAC;QAED,OAAO;YACL,MAAM,EAAE,MAAM;YACd,QAAQ;YACR,WAAW,EAAE,WAAW,CAAC,MAAM;YAC/B,OAAO,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBAClC,IAAI;gBACJ,IAAI,EAAE,QAAQ;gBACd,cAAc,EAAE,CAAC;aAClB,CAAC,CAAC;SACJ,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
import type { Readable } from "node:stream";
|
|
2
|
+
import type { DataStream, InspectionMetadata, ReaderOptions, TabularReader } from "../core/types.js";
|
|
3
|
+
export interface JSONLReaderOptions extends ReaderOptions {
|
|
4
|
+
filePath?: string;
|
|
5
|
+
}
|
|
6
|
+
export declare class JSONLReader implements TabularReader {
|
|
7
|
+
private input;
|
|
8
|
+
private options;
|
|
9
|
+
constructor(input: Readable | string, options?: JSONLReaderOptions);
|
|
10
|
+
private getInputStream;
|
|
11
|
+
read(options?: ReaderOptions): DataStream;
|
|
12
|
+
inspect(options?: ReaderOptions): Promise<InspectionMetadata>;
|
|
13
|
+
}
|
|
14
|
+
//# sourceMappingURL=jsonl.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"jsonl.d.ts","sourceRoot":"","sources":["../../src/readers/jsonl.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,KAAK,EAEV,UAAU,EACV,kBAAkB,EAClB,aAAa,EAEb,aAAa,EACd,MAAM,kBAAkB,CAAC;AAE1B,MAAM,WAAW,kBAAmB,SAAQ,aAAa;IACvD,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,qBAAa,WAAY,YAAW,aAAa;IAC/C,OAAO,CAAC,KAAK,CAAoB;IACjC,OAAO,CAAC,OAAO,CAAqB;gBAExB,KAAK,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,kBAAuB;IAQtE,OAAO,CAAC,cAAc;IAOf,IAAI,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,UAAU;IAmE1C,OAAO,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,kBAAkB,CAAC;CAsBpE"}
|