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 +1,27 @@
|
|
|
1
|
+
import { formatNumber } from "./formatting.js";
|
|
2
|
+
export class ProgressReporter {
|
|
3
|
+
isEnabled;
|
|
4
|
+
hasReported = false;
|
|
5
|
+
constructor(options = {}) {
|
|
6
|
+
this.isEnabled =
|
|
7
|
+
!options.quiet &&
|
|
8
|
+
!options.noProgress &&
|
|
9
|
+
Boolean(process.stderr.isTTY);
|
|
10
|
+
}
|
|
11
|
+
update(info) {
|
|
12
|
+
if (!this.isEnabled)
|
|
13
|
+
return;
|
|
14
|
+
this.hasReported = true;
|
|
15
|
+
const rows = formatNumber(info.rowsProcessed);
|
|
16
|
+
const speed = formatNumber(info.rowsPerSecond);
|
|
17
|
+
const timeSec = (info.elapsedMs / 1000).toFixed(1);
|
|
18
|
+
const line = `\r\x1b[2K Processed ${rows} rows | ${timeSec}s | ${speed} rows/s`;
|
|
19
|
+
process.stderr.write(line);
|
|
20
|
+
}
|
|
21
|
+
done() {
|
|
22
|
+
if (this.isEnabled && this.hasReported) {
|
|
23
|
+
process.stderr.write("\r\x1b[2K");
|
|
24
|
+
}
|
|
25
|
+
}
|
|
26
|
+
}
|
|
27
|
+
//# sourceMappingURL=progress.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"progress.js","sourceRoot":"","sources":["../../src/utils/progress.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAO/C,MAAM,OAAO,gBAAgB;IACnB,SAAS,CAAU;IACnB,WAAW,GAAG,KAAK,CAAC;IAE5B,YAAY,UAAmC,EAAE;QAC/C,IAAI,CAAC,SAAS;YACZ,CAAC,OAAO,CAAC,KAAK;gBACd,CAAC,OAAO,CAAC,UAAU;gBACnB,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IAClC,CAAC;IAED,MAAM,CAAC,IAA0B;QAC/B,IAAI,CAAC,IAAI,CAAC,SAAS;YAAE,OAAO;QAC5B,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC;QAExB,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAC9C,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAC/C,MAAM,OAAO,GAAG,CAAC,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QAEnD,MAAM,IAAI,GAAG,wBAAwB,IAAI,WAAW,OAAO,OAAO,KAAK,SAAS,CAAC;QACjF,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC7B,CAAC;IAED,IAAI;QACF,IAAI,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;YACvC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { DataStream, TabularWriter, WriterOptions } from "../core/types.js";
|
|
3
|
+
export interface CSVWriterOptions extends WriterOptions {
|
|
4
|
+
delimiter?: string;
|
|
5
|
+
header?: boolean;
|
|
6
|
+
}
|
|
7
|
+
export declare class CSVWriter implements TabularWriter {
|
|
8
|
+
private output;
|
|
9
|
+
private options;
|
|
10
|
+
private createdStream?;
|
|
11
|
+
constructor(output: Writable | string, options?: CSVWriterOptions);
|
|
12
|
+
private getOutputStream;
|
|
13
|
+
private escapeField;
|
|
14
|
+
private writeChunk;
|
|
15
|
+
write(dataStream: DataStream, options?: WriterOptions): Promise<void>;
|
|
16
|
+
close(): Promise<void>;
|
|
17
|
+
}
|
|
18
|
+
//# sourceMappingURL=csv.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"csv.d.ts","sourceRoot":"","sources":["../../src/writers/csv.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,KAAK,EAAE,UAAU,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AAEjF,MAAM,WAAW,gBAAiB,SAAQ,aAAa;IACrD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,MAAM,CAAC,EAAE,OAAO,CAAC;CAClB;AAED,qBAAa,SAAU,YAAW,aAAa;IAC7C,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,OAAO,CAAmB;IAClC,OAAO,CAAC,aAAa,CAAC,CAAW;gBAErB,MAAM,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,gBAAqB;IAKrE,OAAO,CAAC,eAAe;IAQvB,OAAO,CAAC,WAAW;YAiBL,UAAU;IAMlB,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IA+CrE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAQ7B"}
|
|
@@ -0,0 +1,84 @@
|
|
|
1
|
+
import { createWriteStream } from "node:fs";
|
|
2
|
+
export class CSVWriter {
|
|
3
|
+
output;
|
|
4
|
+
options;
|
|
5
|
+
createdStream;
|
|
6
|
+
constructor(output, options = {}) {
|
|
7
|
+
this.output = output;
|
|
8
|
+
this.options = { ...options };
|
|
9
|
+
}
|
|
10
|
+
getOutputStream() {
|
|
11
|
+
if (typeof this.output === "string") {
|
|
12
|
+
this.createdStream = createWriteStream(this.output);
|
|
13
|
+
return this.createdStream;
|
|
14
|
+
}
|
|
15
|
+
return this.output;
|
|
16
|
+
}
|
|
17
|
+
escapeField(value, delimiter) {
|
|
18
|
+
if (value === null || value === undefined) {
|
|
19
|
+
return "";
|
|
20
|
+
}
|
|
21
|
+
const str = String(value);
|
|
22
|
+
const mustQuote = str.includes(delimiter) ||
|
|
23
|
+
str.includes('"') ||
|
|
24
|
+
str.includes("\n") ||
|
|
25
|
+
str.includes("\r");
|
|
26
|
+
if (mustQuote) {
|
|
27
|
+
return `"${str.replace(/"/g, '""')}"`;
|
|
28
|
+
}
|
|
29
|
+
return str;
|
|
30
|
+
}
|
|
31
|
+
async writeChunk(stream, data) {
|
|
32
|
+
if (!stream.write(data)) {
|
|
33
|
+
await new Promise((resolve) => stream.once("drain", resolve));
|
|
34
|
+
}
|
|
35
|
+
}
|
|
36
|
+
async write(dataStream, options) {
|
|
37
|
+
const mergedOptions = {
|
|
38
|
+
...this.options,
|
|
39
|
+
...options,
|
|
40
|
+
};
|
|
41
|
+
const delimiter = mergedOptions.delimiter ?? ",";
|
|
42
|
+
const writeHeader = mergedOptions.header !== false;
|
|
43
|
+
const outStream = this.getOutputStream();
|
|
44
|
+
let headers = null;
|
|
45
|
+
for await (const batch of dataStream) {
|
|
46
|
+
if (batch.rows.length === 0)
|
|
47
|
+
continue;
|
|
48
|
+
if (headers === null) {
|
|
49
|
+
// Collect all distinct keys from first batch to form header
|
|
50
|
+
const keySet = new Set();
|
|
51
|
+
for (const row of batch.rows) {
|
|
52
|
+
for (const key of Object.keys(row)) {
|
|
53
|
+
keySet.add(key);
|
|
54
|
+
}
|
|
55
|
+
}
|
|
56
|
+
headers = Array.from(keySet);
|
|
57
|
+
if (writeHeader && headers.length > 0) {
|
|
58
|
+
const headerLine = headers.map((h) => this.escapeField(h, delimiter)).join(delimiter) +
|
|
59
|
+
"\n";
|
|
60
|
+
await this.writeChunk(outStream, headerLine);
|
|
61
|
+
}
|
|
62
|
+
}
|
|
63
|
+
let chunkText = "";
|
|
64
|
+
for (const row of batch.rows) {
|
|
65
|
+
const line = headers
|
|
66
|
+
.map((h) => this.escapeField(row[h], delimiter))
|
|
67
|
+
.join(delimiter) + "\n";
|
|
68
|
+
chunkText += line;
|
|
69
|
+
}
|
|
70
|
+
if (chunkText.length > 0) {
|
|
71
|
+
await this.writeChunk(outStream, chunkText);
|
|
72
|
+
}
|
|
73
|
+
}
|
|
74
|
+
}
|
|
75
|
+
async close() {
|
|
76
|
+
if (this.createdStream) {
|
|
77
|
+
await new Promise((resolve, reject) => {
|
|
78
|
+
this.createdStream.end(() => resolve());
|
|
79
|
+
this.createdStream.on("error", reject);
|
|
80
|
+
});
|
|
81
|
+
}
|
|
82
|
+
}
|
|
83
|
+
}
|
|
84
|
+
//# sourceMappingURL=csv.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"csv.js","sourceRoot":"","sources":["../../src/writers/csv.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,MAAM,SAAS,CAAC;AAS5C,MAAM,OAAO,SAAS;IACZ,MAAM,CAAoB;IAC1B,OAAO,CAAmB;IAC1B,aAAa,CAAY;IAEjC,YAAY,MAAyB,EAAE,UAA4B,EAAE;QACnE,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,OAAO,EAAE,CAAC;IAChC,CAAC;IAEO,eAAe;QACrB,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpD,OAAO,IAAI,CAAC,aAAa,CAAC;QAC5B,CAAC;QACD,OAAO,IAAI,CAAC,MAAM,CAAC;IACrB,CAAC;IAEO,WAAW,CAAC,KAAc,EAAE,SAAiB;QACnD,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YAC1C,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;QAC1B,MAAM,SAAS,GACb,GAAG,CAAC,QAAQ,CAAC,SAAS,CAAC;YACvB,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC;YACjB,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC;YAClB,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAErB,IAAI,SAAS,EAAE,CAAC;YACd,OAAO,IAAI,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC;QACxC,CAAC;QACD,OAAO,GAAG,CAAC;IACb,CAAC;IAEO,KAAK,CAAC,UAAU,CAAC,MAAgB,EAAE,IAAY;QACrD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;YACxB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC;QACtE,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,UAAsB,EAAE,OAAuB;QACzD,MAAM,aAAa,GAAqB;YACtC,GAAG,IAAI,CAAC,OAAO;YACf,GAAG,OAAO;SACX,CAAC;QACF,MAAM,SAAS,GAAG,aAAa,CAAC,SAAS,IAAI,GAAG,CAAC;QACjD,MAAM,WAAW,GAAG,aAAa,CAAC,MAAM,KAAK,KAAK,CAAC;QAEnD,MAAM,SAAS,GAAG,IAAI,CAAC,eAAe,EAAE,CAAC;QACzC,IAAI,OAAO,GAAoB,IAAI,CAAC;QAEpC,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;YACrC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAEtC,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;gBACrB,4DAA4D;gBAC5D,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;gBACjC,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;oBAC7B,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;wBACnC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;oBAClB,CAAC;gBACH,CAAC;gBACD,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBAE7B,IAAI,WAAW,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACtC,MAAM,UAAU,GACd,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC;wBAClE,IAAI,CAAC;oBACP,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,UAAU,CAAC,CAAC;gBAC/C,CAAC;YACH,CAAC;YAED,IAAI,SAAS,GAAG,EAAE,CAAC;YACnB,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;gBAC7B,MAAM,IAAI,GACR,OAAO;qBACJ,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;qBAC/C,IAAI,CAAC,SAAS,CAAC,GAAG,IAAI,CAAC;gBAC5B,SAAS,IAAI,IAAI,CAAC;YACpB,CAAC;YAED,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACzB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC;YAC9C,CAAC;QACH,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK;QACT,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;YACvB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;gBAC1C,IAAI,CAAC,aAAc,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC;gBACzC,IAAI,CAAC,aAAc,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC1C,CAAC,CAAC,CAAC;QACL,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { TabularWriter, WriterOptions } 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 TabularWriter for output and format.
|
|
13
|
+
*/
|
|
14
|
+
export declare function createWriter(output: Writable | string, options?: WriterOptions & {
|
|
15
|
+
format?: string;
|
|
16
|
+
}): TabularWriter;
|
|
17
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/writers/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,MAAM,EAAE,QAAQ,GAAG,MAAM,EACzB,OAAO,GAAE,aAAa,GAAG;IAAE,MAAM,CAAC,EAAE,MAAM,CAAA;CAAO,GAChD,aAAa,CAmBf"}
|
|
@@ -0,0 +1,73 @@
|
|
|
1
|
+
import { InvalidArgumentError } from "../core/errors.js";
|
|
2
|
+
import { CSVWriter } from "./csv.js";
|
|
3
|
+
import { JSONWriter } from "./json.js";
|
|
4
|
+
import { JSONLWriter } from "./jsonl.js";
|
|
5
|
+
import { XLSXWriter } 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: () => {
|
|
15
|
+
throw new Error("Reader adapter called from writer");
|
|
16
|
+
},
|
|
17
|
+
createWriter: (output, options) => new CSVWriter(output, options),
|
|
18
|
+
},
|
|
19
|
+
json: {
|
|
20
|
+
name: "JSON",
|
|
21
|
+
extensions: [".json"],
|
|
22
|
+
createReader: () => {
|
|
23
|
+
throw new Error("Reader adapter called from writer");
|
|
24
|
+
},
|
|
25
|
+
createWriter: (output) => new JSONWriter(output),
|
|
26
|
+
},
|
|
27
|
+
jsonl: {
|
|
28
|
+
name: "JSONL",
|
|
29
|
+
extensions: [".jsonl", ".ndjson", ".ldjson"],
|
|
30
|
+
createReader: () => {
|
|
31
|
+
throw new Error("Reader adapter called from writer");
|
|
32
|
+
},
|
|
33
|
+
createWriter: (output) => new JSONLWriter(output),
|
|
34
|
+
},
|
|
35
|
+
xlsx: {
|
|
36
|
+
name: "XLSX",
|
|
37
|
+
extensions: [".xlsx"],
|
|
38
|
+
createReader: () => {
|
|
39
|
+
throw new Error("Reader adapter called from writer");
|
|
40
|
+
},
|
|
41
|
+
createWriter: (output, options) => new XLSXWriter(output, options),
|
|
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$/, "");
|
|
49
|
+
for (const [format, adapter] of Object.entries(adapters)) {
|
|
50
|
+
if (adapter.extensions.some((ext) => cleanPath.endsWith(ext))) {
|
|
51
|
+
return format;
|
|
52
|
+
}
|
|
53
|
+
}
|
|
54
|
+
return null;
|
|
55
|
+
}
|
|
56
|
+
/**
|
|
57
|
+
* Creates an appropriate TabularWriter for output and format.
|
|
58
|
+
*/
|
|
59
|
+
export function createWriter(output, options = {}) {
|
|
60
|
+
let format = options.format?.toLowerCase();
|
|
61
|
+
if (!format && typeof output === "string" && output !== "-") {
|
|
62
|
+
format = inferFormatFromPath(output) ?? undefined;
|
|
63
|
+
}
|
|
64
|
+
if (!format) {
|
|
65
|
+
format = "csv";
|
|
66
|
+
}
|
|
67
|
+
const adapter = adapters[format];
|
|
68
|
+
if (!adapter) {
|
|
69
|
+
throw new InvalidArgumentError(`Unsupported output format: "${format}". Supported formats are: ${Object.keys(adapters).join(", ")}`);
|
|
70
|
+
}
|
|
71
|
+
return adapter.createWriter(output, options);
|
|
72
|
+
}
|
|
73
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/writers/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,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;QACD,YAAY,EAAE,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,SAAS,CAAC,MAAM,EAAE,OAAO,CAAC;KAClE;IACD,IAAI,EAAE;QACJ,IAAI,EAAE,MAAM;QACZ,UAAU,EAAE,CAAC,OAAO,CAAC;QACrB,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;QACD,YAAY,EAAE,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,UAAU,CAAC,MAAM,CAAC;KACjD;IACD,KAAK,EAAE;QACL,IAAI,EAAE,OAAO;QACb,UAAU,EAAE,CAAC,QAAQ,EAAE,SAAS,EAAE,SAAS,CAAC;QAC5C,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;QACD,YAAY,EAAE,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,WAAW,CAAC,MAAM,CAAC;KAClD;IACD,IAAI,EAAE;QACJ,IAAI,EAAE,MAAM;QACZ,UAAU,EAAE,CAAC,OAAO,CAAC;QACrB,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;QACD,YAAY,EAAE,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,UAAU,CAAC,MAAM,EAAE,OAAO,CAAC;KACnE;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,MAAyB,EACzB,UAA+C,EAAE;IAEjD,IAAI,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,WAAW,EAAE,CAAC;IAE3C,IAAI,CAAC,MAAM,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,GAAG,EAAE,CAAC;QAC5D,MAAM,GAAG,mBAAmB,CAAC,MAAM,CAAC,IAAI,SAAS,CAAC;IACpD,CAAC;IAED,IAAI,CAAC,MAAM,EAAE,CAAC;QACZ,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,+BAA+B,MAAM,6BAA6B,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACrG,CAAC;IACJ,CAAC;IAED,OAAO,OAAO,CAAC,YAAY,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AAC/C,CAAC"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { DataStream, TabularWriter, WriterOptions } from "../core/types.js";
|
|
3
|
+
export declare class JSONWriter implements TabularWriter {
|
|
4
|
+
private output;
|
|
5
|
+
private createdStream?;
|
|
6
|
+
constructor(output: Writable | string);
|
|
7
|
+
private getOutputStream;
|
|
8
|
+
private writeChunk;
|
|
9
|
+
write(dataStream: DataStream, _options?: WriterOptions): Promise<void>;
|
|
10
|
+
close(): Promise<void>;
|
|
11
|
+
}
|
|
12
|
+
//# sourceMappingURL=json.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.d.ts","sourceRoot":"","sources":["../../src/writers/json.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,KAAK,EAAE,UAAU,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AAEjF,qBAAa,UAAW,YAAW,aAAa;IAC9C,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,aAAa,CAAC,CAAW;gBAErB,MAAM,EAAE,QAAQ,GAAG,MAAM;IAIrC,OAAO,CAAC,eAAe;YAQT,UAAU;IAMlB,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,QAAQ,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAiBtE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAQ7B"}
|
|
@@ -0,0 +1,42 @@
|
|
|
1
|
+
import { createWriteStream } from "node:fs";
|
|
2
|
+
export class JSONWriter {
|
|
3
|
+
output;
|
|
4
|
+
createdStream;
|
|
5
|
+
constructor(output) {
|
|
6
|
+
this.output = output;
|
|
7
|
+
}
|
|
8
|
+
getOutputStream() {
|
|
9
|
+
if (typeof this.output === "string") {
|
|
10
|
+
this.createdStream = createWriteStream(this.output);
|
|
11
|
+
return this.createdStream;
|
|
12
|
+
}
|
|
13
|
+
return this.output;
|
|
14
|
+
}
|
|
15
|
+
async writeChunk(stream, data) {
|
|
16
|
+
if (!stream.write(data)) {
|
|
17
|
+
await new Promise((resolve) => stream.once("drain", resolve));
|
|
18
|
+
}
|
|
19
|
+
}
|
|
20
|
+
async write(dataStream, _options) {
|
|
21
|
+
const outStream = this.getOutputStream();
|
|
22
|
+
let isFirstRow = true;
|
|
23
|
+
await this.writeChunk(outStream, "[\n");
|
|
24
|
+
for await (const batch of dataStream) {
|
|
25
|
+
for (const row of batch.rows) {
|
|
26
|
+
const prefix = isFirstRow ? " " : ",\n ";
|
|
27
|
+
isFirstRow = false;
|
|
28
|
+
await this.writeChunk(outStream, `${prefix}${JSON.stringify(row)}`);
|
|
29
|
+
}
|
|
30
|
+
}
|
|
31
|
+
await this.writeChunk(outStream, "\n]\n");
|
|
32
|
+
}
|
|
33
|
+
async close() {
|
|
34
|
+
if (this.createdStream) {
|
|
35
|
+
await new Promise((resolve, reject) => {
|
|
36
|
+
this.createdStream.end(() => resolve());
|
|
37
|
+
this.createdStream.on("error", reject);
|
|
38
|
+
});
|
|
39
|
+
}
|
|
40
|
+
}
|
|
41
|
+
}
|
|
42
|
+
//# sourceMappingURL=json.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.js","sourceRoot":"","sources":["../../src/writers/json.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,MAAM,SAAS,CAAC;AAI5C,MAAM,OAAO,UAAU;IACb,MAAM,CAAoB;IAC1B,aAAa,CAAY;IAEjC,YAAY,MAAyB;QACnC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAEO,eAAe;QACrB,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpD,OAAO,IAAI,CAAC,aAAa,CAAC;QAC5B,CAAC;QACD,OAAO,IAAI,CAAC,MAAM,CAAC;IACrB,CAAC;IAEO,KAAK,CAAC,UAAU,CAAC,MAAgB,EAAE,IAAY;QACrD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;YACxB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC;QACtE,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,UAAsB,EAAE,QAAwB;QAC1D,MAAM,SAAS,GAAG,IAAI,CAAC,eAAe,EAAE,CAAC;QACzC,IAAI,UAAU,GAAG,IAAI,CAAC;QAEtB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC;QAExC,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;YACrC,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;gBAC7B,MAAM,MAAM,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC;gBAC3C,UAAU,GAAG,KAAK,CAAC;gBACnB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,GAAG,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACtE,CAAC;QACH,CAAC;QAED,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IAC5C,CAAC;IAED,KAAK,CAAC,KAAK;QACT,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;YACvB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;gBAC1C,IAAI,CAAC,aAAc,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC;gBACzC,IAAI,CAAC,aAAc,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC1C,CAAC,CAAC,CAAC;QACL,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { DataStream, TabularWriter, WriterOptions } from "../core/types.js";
|
|
3
|
+
export declare class JSONLWriter implements TabularWriter {
|
|
4
|
+
private output;
|
|
5
|
+
private createdStream?;
|
|
6
|
+
constructor(output: Writable | string);
|
|
7
|
+
private getOutputStream;
|
|
8
|
+
private writeChunk;
|
|
9
|
+
write(dataStream: DataStream, _options?: WriterOptions): Promise<void>;
|
|
10
|
+
close(): Promise<void>;
|
|
11
|
+
}
|
|
12
|
+
//# sourceMappingURL=jsonl.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"jsonl.d.ts","sourceRoot":"","sources":["../../src/writers/jsonl.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,KAAK,EAAE,UAAU,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AAEjF,qBAAa,WAAY,YAAW,aAAa;IAC/C,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,aAAa,CAAC,CAAW;gBAErB,MAAM,EAAE,QAAQ,GAAG,MAAM;IAIrC,OAAO,CAAC,eAAe;YAQT,UAAU;IAMlB,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,QAAQ,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IActE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAQ7B"}
|
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
import { createWriteStream } from "node:fs";
|
|
2
|
+
export class JSONLWriter {
|
|
3
|
+
output;
|
|
4
|
+
createdStream;
|
|
5
|
+
constructor(output) {
|
|
6
|
+
this.output = output;
|
|
7
|
+
}
|
|
8
|
+
getOutputStream() {
|
|
9
|
+
if (typeof this.output === "string") {
|
|
10
|
+
this.createdStream = createWriteStream(this.output);
|
|
11
|
+
return this.createdStream;
|
|
12
|
+
}
|
|
13
|
+
return this.output;
|
|
14
|
+
}
|
|
15
|
+
async writeChunk(stream, data) {
|
|
16
|
+
if (!stream.write(data)) {
|
|
17
|
+
await new Promise((resolve) => stream.once("drain", resolve));
|
|
18
|
+
}
|
|
19
|
+
}
|
|
20
|
+
async write(dataStream, _options) {
|
|
21
|
+
const outStream = this.getOutputStream();
|
|
22
|
+
for await (const batch of dataStream) {
|
|
23
|
+
let chunkText = "";
|
|
24
|
+
for (const row of batch.rows) {
|
|
25
|
+
chunkText += JSON.stringify(row) + "\n";
|
|
26
|
+
}
|
|
27
|
+
if (chunkText.length > 0) {
|
|
28
|
+
await this.writeChunk(outStream, chunkText);
|
|
29
|
+
}
|
|
30
|
+
}
|
|
31
|
+
}
|
|
32
|
+
async close() {
|
|
33
|
+
if (this.createdStream) {
|
|
34
|
+
await new Promise((resolve, reject) => {
|
|
35
|
+
this.createdStream.end(() => resolve());
|
|
36
|
+
this.createdStream.on("error", reject);
|
|
37
|
+
});
|
|
38
|
+
}
|
|
39
|
+
}
|
|
40
|
+
}
|
|
41
|
+
//# sourceMappingURL=jsonl.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"jsonl.js","sourceRoot":"","sources":["../../src/writers/jsonl.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,MAAM,SAAS,CAAC;AAI5C,MAAM,OAAO,WAAW;IACd,MAAM,CAAoB;IAC1B,aAAa,CAAY;IAEjC,YAAY,MAAyB;QACnC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAEO,eAAe;QACrB,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpD,OAAO,IAAI,CAAC,aAAa,CAAC;QAC5B,CAAC;QACD,OAAO,IAAI,CAAC,MAAM,CAAC;IACrB,CAAC;IAEO,KAAK,CAAC,UAAU,CAAC,MAAgB,EAAE,IAAY;QACrD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;YACxB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC;QACtE,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,UAAsB,EAAE,QAAwB;QAC1D,MAAM,SAAS,GAAG,IAAI,CAAC,eAAe,EAAE,CAAC;QAEzC,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;YACrC,IAAI,SAAS,GAAG,EAAE,CAAC;YACnB,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;gBAC7B,SAAS,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC;YAC1C,CAAC;YACD,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACzB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC;YAC9C,CAAC;QACH,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK;QACT,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;YACvB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;gBAC1C,IAAI,CAAC,aAAc,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC;gBACzC,IAAI,CAAC,aAAc,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC1C,CAAC,CAAC,CAAC;QACL,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { DataStream, TabularWriter, WriterOptions } from "../core/types.js";
|
|
3
|
+
export interface XLSXWriterOptions extends WriterOptions {
|
|
4
|
+
sheet?: string;
|
|
5
|
+
}
|
|
6
|
+
export declare class XLSXWriter implements TabularWriter {
|
|
7
|
+
private output;
|
|
8
|
+
private options;
|
|
9
|
+
constructor(output: Writable | string, options?: XLSXWriterOptions);
|
|
10
|
+
write(dataStream: DataStream, options?: WriterOptions): Promise<void>;
|
|
11
|
+
close(): Promise<void>;
|
|
12
|
+
}
|
|
13
|
+
//# sourceMappingURL=xlsx.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"xlsx.d.ts","sourceRoot":"","sources":["../../src/writers/xlsx.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,KAAK,EAAE,UAAU,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AAEjF,MAAM,WAAW,iBAAkB,SAAQ,aAAa;IACtD,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,qBAAa,UAAW,YAAW,aAAa;IAC9C,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,OAAO,CAAoB;gBAEvB,MAAM,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,iBAAsB;IAKhE,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAkErE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAG7B"}
|
|
@@ -0,0 +1,65 @@
|
|
|
1
|
+
import writeXlsxFile from "write-excel-file/node";
|
|
2
|
+
export class XLSXWriter {
|
|
3
|
+
output;
|
|
4
|
+
options;
|
|
5
|
+
constructor(output, options = {}) {
|
|
6
|
+
this.output = output;
|
|
7
|
+
this.options = { ...options };
|
|
8
|
+
}
|
|
9
|
+
async write(dataStream, options) {
|
|
10
|
+
const mergedOptions = {
|
|
11
|
+
...this.options,
|
|
12
|
+
...options,
|
|
13
|
+
};
|
|
14
|
+
const sheetName = mergedOptions.sheet || "Sheet1";
|
|
15
|
+
let headers = null;
|
|
16
|
+
const grid = [];
|
|
17
|
+
for await (const batch of dataStream) {
|
|
18
|
+
if (batch.rows.length === 0)
|
|
19
|
+
continue;
|
|
20
|
+
if (headers === null) {
|
|
21
|
+
const keySet = new Set();
|
|
22
|
+
for (const row of batch.rows) {
|
|
23
|
+
for (const key of Object.keys(row)) {
|
|
24
|
+
keySet.add(key);
|
|
25
|
+
}
|
|
26
|
+
}
|
|
27
|
+
headers = Array.from(keySet);
|
|
28
|
+
// Header row
|
|
29
|
+
grid.push(headers.map((h) => ({
|
|
30
|
+
value: h,
|
|
31
|
+
fontWeight: "bold",
|
|
32
|
+
})));
|
|
33
|
+
}
|
|
34
|
+
for (const row of batch.rows) {
|
|
35
|
+
grid.push(headers.map((h) => {
|
|
36
|
+
const val = row[h];
|
|
37
|
+
if (val === null || val === undefined || val === "") {
|
|
38
|
+
return { value: undefined };
|
|
39
|
+
}
|
|
40
|
+
if (typeof val === "number" || typeof val === "boolean" || typeof val === "string") {
|
|
41
|
+
return { value: val };
|
|
42
|
+
}
|
|
43
|
+
if (val instanceof Date) {
|
|
44
|
+
return { value: val };
|
|
45
|
+
}
|
|
46
|
+
return { value: String(val) };
|
|
47
|
+
}));
|
|
48
|
+
}
|
|
49
|
+
}
|
|
50
|
+
if (grid.length === 0) {
|
|
51
|
+
grid.push([{ value: "" }]);
|
|
52
|
+
}
|
|
53
|
+
const writerInstance = writeXlsxFile(grid, { sheet: sheetName });
|
|
54
|
+
if (typeof this.output === "string") {
|
|
55
|
+
await writerInstance.toFile(this.output);
|
|
56
|
+
}
|
|
57
|
+
else {
|
|
58
|
+
await writerInstance.toStream(this.output);
|
|
59
|
+
}
|
|
60
|
+
}
|
|
61
|
+
async close() {
|
|
62
|
+
// Handled on write
|
|
63
|
+
}
|
|
64
|
+
}
|
|
65
|
+
//# sourceMappingURL=xlsx.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"xlsx.js","sourceRoot":"","sources":["../../src/writers/xlsx.ts"],"names":[],"mappings":"AACA,OAAO,aAAa,MAAM,uBAAuB,CAAC;AAOlD,MAAM,OAAO,UAAU;IACb,MAAM,CAAoB;IAC1B,OAAO,CAAoB;IAEnC,YAAY,MAAyB,EAAE,UAA6B,EAAE;QACpE,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,OAAO,EAAE,CAAC;IAChC,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,UAAsB,EAAE,OAAuB;QACzD,MAAM,aAAa,GAAsB;YACvC,GAAG,IAAI,CAAC,OAAO;YACf,GAAG,OAAO;SACX,CAAC;QACF,MAAM,SAAS,GAAG,aAAa,CAAC,KAAK,IAAI,QAAQ,CAAC;QAElD,IAAI,OAAO,GAAoB,IAAI,CAAC;QACpC,MAAM,IAAI,GAAoF,EAAE,CAAC;QAEjG,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;YACrC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAEtC,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;gBACrB,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;gBACjC,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;oBAC7B,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;wBACnC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;oBAClB,CAAC;gBACH,CAAC;gBACD,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBAE7B,aAAa;gBACb,IAAI,CAAC,IAAI,CACP,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;oBAClB,KAAK,EAAE,CAAC;oBACR,UAAU,EAAE,MAAM;iBACnB,CAAC,CAAC,CACJ,CAAC;YACJ,CAAC;YAED,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;gBAC7B,IAAI,CAAC,IAAI,CACP,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;oBAChB,MAAM,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;oBACnB,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;wBACpD,OAAO,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;oBAC9B,CAAC;oBACD,IAAI,OAAO,GAAG,KAAK,QAAQ,IAAI,OAAO,GAAG,KAAK,SAAS,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;wBACnF,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC;oBACxB,CAAC;oBACD,IAAI,GAAG,YAAY,IAAI,EAAE,CAAC;wBACxB,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC;oBACxB,CAAC;oBACD,OAAO,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;gBAChC,CAAC,CAAC,CACH,CAAC;YACJ,CAAC;QACH,CAAC;QAED,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACtB,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC;QAC7B,CAAC;QAED,MAAM,cAAc,GAAI,aAGtB,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,CAAC;QAE/B,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,MAAM,cAAc,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC3C,CAAC;aAAM,CAAC;YACN,MAAM,cAAc,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC7C,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK;QACT,mBAAmB;IACrB,CAAC;CACF"}
|
package/package.json
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
1
|
+
{
|
|
2
|
+
"name": "rowpipe",
|
|
3
|
+
"version": "1.0.0",
|
|
4
|
+
"description": "A stream-first tabular data toolkit for CSV, JSON, JSONL, and XLSX",
|
|
5
|
+
"main": "dist/index.js",
|
|
6
|
+
"types": "dist/index.d.ts",
|
|
7
|
+
"bin": {
|
|
8
|
+
"rowpipe": "dist/cli/index.js"
|
|
9
|
+
},
|
|
10
|
+
"type": "module",
|
|
11
|
+
"scripts": {
|
|
12
|
+
"build": "tsc",
|
|
13
|
+
"dev": "tsx src/cli/index.ts",
|
|
14
|
+
"test": "vitest run",
|
|
15
|
+
"test:watch": "vitest",
|
|
16
|
+
"bench": "tsx benchmarks/memory-bench.ts"
|
|
17
|
+
},
|
|
18
|
+
"keywords": [
|
|
19
|
+
"csv",
|
|
20
|
+
"json",
|
|
21
|
+
"jsonl",
|
|
22
|
+
"ndjson",
|
|
23
|
+
"xlsx",
|
|
24
|
+
"streaming",
|
|
25
|
+
"tabular",
|
|
26
|
+
"data",
|
|
27
|
+
"cli",
|
|
28
|
+
"pipeline",
|
|
29
|
+
"analytics"
|
|
30
|
+
],
|
|
31
|
+
"author": "",
|
|
32
|
+
"license": "MIT",
|
|
33
|
+
"dependencies": {
|
|
34
|
+
"commander": "^13.1.0",
|
|
35
|
+
"read-excel-file": "^9.3.10",
|
|
36
|
+
"write-excel-file": "^4.1.1"
|
|
37
|
+
},
|
|
38
|
+
"devDependencies": {
|
|
39
|
+
"@types/node": "^22.13.9",
|
|
40
|
+
"tsx": "^4.19.3",
|
|
41
|
+
"typescript": "^5.8.2",
|
|
42
|
+
"vitest": "^3.0.7"
|
|
43
|
+
}
|
|
44
|
+
}
|