rowpipe 1.0.0 → 1.2.0
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 +327 -41
- package/dist/cli/commands/diff.d.ts +30 -0
- package/dist/cli/commands/diff.d.ts.map +1 -0
- package/dist/cli/commands/diff.js +143 -0
- package/dist/cli/commands/diff.js.map +1 -0
- package/dist/cli/commands/files.d.ts +35 -0
- package/dist/cli/commands/files.d.ts.map +1 -0
- package/dist/cli/commands/files.js +93 -0
- package/dist/cli/commands/files.js.map +1 -0
- package/dist/cli/index.js +89 -5
- package/dist/cli/index.js.map +1 -1
- package/dist/core/errors.d.ts +23 -0
- package/dist/core/errors.d.ts.map +1 -1
- package/dist/core/errors.js +38 -0
- package/dist/core/errors.js.map +1 -1
- package/dist/core/types.d.ts +4 -0
- package/dist/core/types.d.ts.map +1 -1
- package/dist/diff/comparator.d.ts +17 -0
- package/dist/diff/comparator.d.ts.map +1 -0
- package/dist/diff/comparator.js +158 -0
- package/dist/diff/comparator.js.map +1 -0
- package/dist/diff/engine.d.ts +10 -0
- package/dist/diff/engine.d.ts.map +1 -0
- package/dist/diff/engine.js +333 -0
- package/dist/diff/engine.js.map +1 -0
- package/dist/diff/hash.d.ts +6 -0
- package/dist/diff/hash.d.ts.map +1 -0
- package/dist/diff/hash.js +50 -0
- package/dist/diff/hash.js.map +1 -0
- package/dist/diff/key.d.ts +21 -0
- package/dist/diff/key.d.ts.map +1 -0
- package/dist/diff/key.js +107 -0
- package/dist/diff/key.js.map +1 -0
- package/dist/diff/reporter.d.ts +18 -0
- package/dist/diff/reporter.d.ts.map +1 -0
- package/dist/diff/reporter.js +133 -0
- package/dist/diff/reporter.js.map +1 -0
- package/dist/diff/schema.d.ts +7 -0
- package/dist/diff/schema.d.ts.map +1 -0
- package/dist/diff/schema.js +65 -0
- package/dist/diff/schema.js.map +1 -0
- package/dist/diff/storage/disk-index.d.ts +28 -0
- package/dist/diff/storage/disk-index.d.ts.map +1 -0
- package/dist/diff/storage/disk-index.js +105 -0
- package/dist/diff/storage/disk-index.js.map +1 -0
- package/dist/diff/storage/memory-index.d.ts +23 -0
- package/dist/diff/storage/memory-index.d.ts.map +1 -0
- package/dist/diff/storage/memory-index.js +69 -0
- package/dist/diff/storage/memory-index.js.map +1 -0
- package/dist/diff/storage/spillable-index.d.ts +30 -0
- package/dist/diff/storage/spillable-index.d.ts.map +1 -0
- package/dist/diff/storage/spillable-index.js +108 -0
- package/dist/diff/storage/spillable-index.js.map +1 -0
- package/dist/diff/types.d.ts +127 -0
- package/dist/diff/types.d.ts.map +1 -0
- package/dist/diff/types.js +2 -0
- package/dist/diff/types.js.map +1 -0
- package/dist/files/glob.d.ts +21 -0
- package/dist/files/glob.d.ts.map +1 -0
- package/dist/files/glob.js +132 -0
- package/dist/files/glob.js.map +1 -0
- package/dist/files/hash.d.ts +10 -0
- package/dist/files/hash.d.ts.map +1 -0
- package/dist/files/hash.js +67 -0
- package/dist/files/hash.js.map +1 -0
- package/dist/files/index.d.ts +6 -0
- package/dist/files/index.d.ts.map +1 -0
- package/dist/files/index.js +6 -0
- package/dist/files/index.js.map +1 -0
- package/dist/files/mime.d.ts +9 -0
- package/dist/files/mime.d.ts.map +1 -0
- package/dist/files/mime.js +128 -0
- package/dist/files/mime.js.map +1 -0
- package/dist/files/reader.d.ts +19 -0
- package/dist/files/reader.d.ts.map +1 -0
- package/dist/files/reader.js +344 -0
- package/dist/files/reader.js.map +1 -0
- package/dist/files/types.d.ts +47 -0
- package/dist/files/types.d.ts.map +1 -0
- package/dist/files/types.js +2 -0
- package/dist/files/types.js.map +1 -0
- package/dist/index.d.ts +17 -2
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +19 -2
- package/dist/index.js.map +1 -1
- package/dist/readers/csv.d.ts.map +1 -1
- package/dist/readers/csv.js +59 -27
- package/dist/readers/csv.js.map +1 -1
- package/dist/readers/index.d.ts +2 -0
- package/dist/readers/index.d.ts.map +1 -1
- package/dist/readers/index.js +57 -3
- package/dist/readers/index.js.map +1 -1
- package/dist/readers/json.d.ts.map +1 -1
- package/dist/readers/json.js +6 -8
- package/dist/readers/json.js.map +1 -1
- package/dist/readers/jsonl.d.ts.map +1 -1
- package/dist/readers/jsonl.js +5 -7
- package/dist/readers/jsonl.js.map +1 -1
- package/dist/readers/parquet.d.ts +17 -0
- package/dist/readers/parquet.d.ts.map +1 -0
- package/dist/readers/parquet.js +146 -0
- package/dist/readers/parquet.js.map +1 -0
- package/dist/transforms/expression.d.ts.map +1 -1
- package/dist/transforms/expression.js +73 -7
- package/dist/transforms/expression.js.map +1 -1
- package/dist/utils/compression.d.ts +30 -6
- package/dist/utils/compression.d.ts.map +1 -1
- package/dist/utils/compression.js +95 -18
- package/dist/utils/compression.js.map +1 -1
- package/dist/writers/csv.d.ts.map +1 -1
- package/dist/writers/csv.js +29 -17
- package/dist/writers/csv.js.map +1 -1
- package/dist/writers/index.d.ts +2 -0
- package/dist/writers/index.d.ts.map +1 -1
- package/dist/writers/index.js +49 -4
- package/dist/writers/index.js.map +1 -1
- package/dist/writers/json.d.ts +1 -1
- package/dist/writers/json.d.ts.map +1 -1
- package/dist/writers/json.js +9 -11
- package/dist/writers/json.js.map +1 -1
- package/dist/writers/jsonl.d.ts +1 -1
- package/dist/writers/jsonl.d.ts.map +1 -1
- package/dist/writers/jsonl.js +9 -11
- package/dist/writers/jsonl.js.map +1 -1
- package/dist/writers/markdown.d.ts +21 -0
- package/dist/writers/markdown.d.ts.map +1 -0
- package/dist/writers/markdown.js +82 -0
- package/dist/writers/markdown.js.map +1 -0
- package/dist/writers/parquet.d.ts +19 -0
- package/dist/writers/parquet.d.ts.map +1 -0
- package/dist/writers/parquet.js +129 -0
- package/dist/writers/parquet.js.map +1 -0
- package/package.json +16 -3
- package/skills/rowpipe/SKILL.md +263 -0
- package/benchmarks/memory-bench.ts +0 -109
- package/src/analytics/reduce.ts +0 -388
- package/src/analytics/schema-inference.ts +0 -217
- package/src/analytics/semantic-types.ts +0 -36
- package/src/analytics/stats.ts +0 -383
- package/src/analytics/validator.ts +0 -134
- package/src/cli/commands/cast.ts +0 -60
- package/src/cli/commands/convert.ts +0 -128
- package/src/cli/commands/filter.ts +0 -58
- package/src/cli/commands/inspect.ts +0 -193
- package/src/cli/commands/map.ts +0 -66
- package/src/cli/commands/reduce.ts +0 -74
- package/src/cli/commands/rename.ts +0 -59
- package/src/cli/commands/sample.ts +0 -61
- package/src/cli/commands/schema.ts +0 -86
- package/src/cli/commands/select.ts +0 -59
- package/src/cli/commands/stats.ts +0 -100
- package/src/cli/commands/validate.ts +0 -123
- package/src/cli/index.ts +0 -285
- package/src/core/batch.ts +0 -81
- package/src/core/errors.ts +0 -81
- package/src/core/pipeline.ts +0 -150
- package/src/core/types.ts +0 -110
- package/src/index.ts +0 -52
- package/src/readers/csv.ts +0 -302
- package/src/readers/index.ts +0 -88
- package/src/readers/json.ts +0 -256
- package/src/readers/jsonl.ts +0 -126
- package/src/readers/xlsx.ts +0 -182
- package/src/transforms/cast.ts +0 -212
- package/src/transforms/expression.ts +0 -1114
- package/src/transforms/filter.ts +0 -38
- package/src/transforms/map.ts +0 -99
- package/src/transforms/rename.ts +0 -47
- package/src/transforms/sample.ts +0 -57
- package/src/transforms/select.ts +0 -64
- package/src/utils/compression.ts +0 -43
- package/src/utils/formatting.ts +0 -75
- package/src/utils/progress.ts +0 -37
- package/src/writers/csv.ts +0 -106
- package/src/writers/index.ts +0 -87
- package/src/writers/json.ts +0 -52
- package/src/writers/jsonl.ts +0 -49
- package/src/writers/xlsx.ts +0 -87
- package/tests/analytics.test.ts +0 -119
- package/tests/cli-integration.test.ts +0 -187
- package/tests/map-reduce.test.ts +0 -223
- package/tests/readers-writers.test.ts +0 -189
- package/tests/transforms.test.ts +0 -199
- package/tsconfig.json +0 -21
- package/vitest.config.ts +0 -9
package/dist/writers/csv.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"csv.js","sourceRoot":"","sources":["../../src/writers/csv.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,
|
|
1
|
+
{"version":3,"file":"csv.js","sourceRoot":"","sources":["../../src/writers/csv.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,yBAAyB,EAAE,MAAM,yBAAyB,CAAC;AASzF,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,CAAC,IAAoB;QAC1C,MAAM,IAAI,GAAG,IAAI,EAAE,IAAI,IAAI,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC;QAC7C,MAAM,MAAM,GAAG,yBAAyB,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAChE,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC;QAC9B,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,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,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;YAC9B,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;QACD,IAAI,OAAO,KAAK,KAAK,SAAS,EAAE,CAAC;YAC/B,OAAO,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC;QAClC,CAAC;QACD,MAAM,GAAG,GAAG,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;QACpH,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,CAAC,aAAa,CAAC,CAAC;QACtD,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,IAAI,UAAU,GAAG,EAAE,CAAC;oBACpB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;wBACxC,IAAI,CAAC,GAAG,CAAC;4BAAE,UAAU,IAAI,SAAS,CAAC;wBACnC,UAAU,IAAI,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;oBACxD,CAAC;oBACD,UAAU,IAAI,IAAI,CAAC;oBACnB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,UAAU,CAAC,CAAC;gBAC/C,CAAC;YACH,CAAC;YAED,IAAI,SAAS,GAAG,EAAE,CAAC;YACnB,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC;YAC5B,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;gBAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC9B,IAAI,CAAC,GAAG,CAAC;wBAAE,SAAS,IAAI,SAAS,CAAC;oBAClC,SAAS,IAAI,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAE,CAAC,EAAE,SAAS,CAAC,CAAC;gBAC7D,CAAC;gBACD,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,mBAAmB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;CACF"}
|
package/dist/writers/index.d.ts
CHANGED
|
@@ -3,6 +3,8 @@ import type { TabularWriter, WriterOptions } from "../core/types.js";
|
|
|
3
3
|
export * from "./csv.js";
|
|
4
4
|
export * from "./json.js";
|
|
5
5
|
export * from "./jsonl.js";
|
|
6
|
+
export * from "./markdown.js";
|
|
7
|
+
export * from "./parquet.js";
|
|
6
8
|
export * from "./xlsx.js";
|
|
7
9
|
/**
|
|
8
10
|
* Infers format name from file path or extension.
|
|
@@ -1 +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;
|
|
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;AASpF,cAAc,UAAU,CAAC;AACzB,cAAc,WAAW,CAAC;AAC1B,cAAc,YAAY,CAAC;AAC3B,cAAc,eAAe,CAAC;AAC9B,cAAc,cAAc,CAAC;AAC7B,cAAc,WAAW,CAAC;AA+E1B;;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"}
|
package/dist/writers/index.js
CHANGED
|
@@ -1,28 +1,73 @@
|
|
|
1
1
|
import { InvalidArgumentError } from "../core/errors.js";
|
|
2
|
+
import { stripCompressionExtension } from "../utils/compression.js";
|
|
2
3
|
import { CSVWriter } from "./csv.js";
|
|
3
4
|
import { JSONWriter } from "./json.js";
|
|
4
5
|
import { JSONLWriter } from "./jsonl.js";
|
|
6
|
+
import { MarkdownWriter } from "./markdown.js";
|
|
7
|
+
import { ParquetWriter } from "./parquet.js";
|
|
5
8
|
import { XLSXWriter } from "./xlsx.js";
|
|
6
9
|
export * from "./csv.js";
|
|
7
10
|
export * from "./json.js";
|
|
8
11
|
export * from "./jsonl.js";
|
|
12
|
+
export * from "./markdown.js";
|
|
13
|
+
export * from "./parquet.js";
|
|
9
14
|
export * from "./xlsx.js";
|
|
10
15
|
const adapters = {
|
|
11
16
|
csv: {
|
|
12
17
|
name: "CSV",
|
|
13
|
-
extensions: [".csv", ".
|
|
18
|
+
extensions: [".csv", ".txt"],
|
|
14
19
|
createReader: () => {
|
|
15
20
|
throw new Error("Reader adapter called from writer");
|
|
16
21
|
},
|
|
17
22
|
createWriter: (output, options) => new CSVWriter(output, options),
|
|
18
23
|
},
|
|
24
|
+
tsv: {
|
|
25
|
+
name: "TSV",
|
|
26
|
+
extensions: [".tsv", ".tab"],
|
|
27
|
+
createReader: () => {
|
|
28
|
+
throw new Error("Reader adapter called from writer");
|
|
29
|
+
},
|
|
30
|
+
createWriter: (output, options) => new CSVWriter(output, { ...options, delimiter: options?.delimiter ?? "\t" }),
|
|
31
|
+
},
|
|
32
|
+
psv: {
|
|
33
|
+
name: "PSV",
|
|
34
|
+
extensions: [".psv"],
|
|
35
|
+
createReader: () => {
|
|
36
|
+
throw new Error("Reader adapter called from writer");
|
|
37
|
+
},
|
|
38
|
+
createWriter: (output, options) => new CSVWriter(output, { ...options, delimiter: options?.delimiter ?? "|" }),
|
|
39
|
+
},
|
|
40
|
+
markdown: {
|
|
41
|
+
name: "Markdown",
|
|
42
|
+
extensions: [".md", ".markdown"],
|
|
43
|
+
createReader: () => {
|
|
44
|
+
throw new Error("Reader adapter called from writer");
|
|
45
|
+
},
|
|
46
|
+
createWriter: (output, options) => new MarkdownWriter(output, options),
|
|
47
|
+
},
|
|
48
|
+
md: {
|
|
49
|
+
name: "Markdown",
|
|
50
|
+
extensions: [".md"],
|
|
51
|
+
createReader: () => {
|
|
52
|
+
throw new Error("Reader adapter called from writer");
|
|
53
|
+
},
|
|
54
|
+
createWriter: (output, options) => new MarkdownWriter(output, options),
|
|
55
|
+
},
|
|
56
|
+
parquet: {
|
|
57
|
+
name: "Parquet",
|
|
58
|
+
extensions: [".parquet", ".pq"],
|
|
59
|
+
createReader: () => {
|
|
60
|
+
throw new Error("Reader adapter called from writer");
|
|
61
|
+
},
|
|
62
|
+
createWriter: (output, options) => new ParquetWriter(output, options),
|
|
63
|
+
},
|
|
19
64
|
json: {
|
|
20
65
|
name: "JSON",
|
|
21
66
|
extensions: [".json"],
|
|
22
67
|
createReader: () => {
|
|
23
68
|
throw new Error("Reader adapter called from writer");
|
|
24
69
|
},
|
|
25
|
-
createWriter: (output) => new JSONWriter(output),
|
|
70
|
+
createWriter: (output, options) => new JSONWriter(output),
|
|
26
71
|
},
|
|
27
72
|
jsonl: {
|
|
28
73
|
name: "JSONL",
|
|
@@ -30,7 +75,7 @@ const adapters = {
|
|
|
30
75
|
createReader: () => {
|
|
31
76
|
throw new Error("Reader adapter called from writer");
|
|
32
77
|
},
|
|
33
|
-
createWriter: (output) => new JSONLWriter(output),
|
|
78
|
+
createWriter: (output, options) => new JSONLWriter(output),
|
|
34
79
|
},
|
|
35
80
|
xlsx: {
|
|
36
81
|
name: "XLSX",
|
|
@@ -45,7 +90,7 @@ const adapters = {
|
|
|
45
90
|
* Infers format name from file path or extension.
|
|
46
91
|
*/
|
|
47
92
|
export function inferFormatFromPath(filePath) {
|
|
48
|
-
const cleanPath = filePath.toLowerCase()
|
|
93
|
+
const cleanPath = stripCompressionExtension(filePath.toLowerCase());
|
|
49
94
|
for (const [format, adapter] of Object.entries(adapters)) {
|
|
50
95
|
if (adapter.extensions.some((ext) => cleanPath.endsWith(ext))) {
|
|
51
96
|
return format;
|
|
@@ -1 +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;
|
|
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,yBAAyB,EAAE,MAAM,yBAAyB,CAAC;AACpE,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,cAAc,EAAE,MAAM,eAAe,CAAC;AAC/C,OAAO,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAC7C,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAEvC,cAAc,UAAU,CAAC;AACzB,cAAc,WAAW,CAAC;AAC1B,cAAc,YAAY,CAAC;AAC3B,cAAc,eAAe,CAAC;AAC9B,cAAc,cAAc,CAAC;AAC7B,cAAc,WAAW,CAAC;AAE1B,MAAM,QAAQ,GAAkC;IAC9C,GAAG,EAAE;QACH,IAAI,EAAE,KAAK;QACX,UAAU,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC;QAC5B,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,GAAG,EAAE;QACH,IAAI,EAAE,KAAK;QACX,UAAU,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC;QAC5B,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;QACD,YAAY,EAAE,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE,CAChC,IAAI,SAAS,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,IAAI,EAAE,CAAC;KAC/E;IACD,GAAG,EAAE;QACH,IAAI,EAAE,KAAK;QACX,UAAU,EAAE,CAAC,MAAM,CAAC;QACpB,YAAY,EAAE,GAAG,EAAE;YACjB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;QACD,YAAY,EAAE,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE,CAChC,IAAI,SAAS,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,GAAG,EAAE,CAAC;KAC9E;IACD,QAAQ,EAAE;QACR,IAAI,EAAE,UAAU;QAChB,UAAU,EAAE,CAAC,KAAK,EAAE,WAAW,CAAC;QAChC,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,cAAc,CAAC,MAAM,EAAE,OAAO,CAAC;KACvE;IACD,EAAE,EAAE;QACF,IAAI,EAAE,UAAU;QAChB,UAAU,EAAE,CAAC,KAAK,CAAC;QACnB,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,cAAc,CAAC,MAAM,EAAE,OAAO,CAAC;KACvE;IACD,OAAO,EAAE;QACP,IAAI,EAAE,SAAS;QACf,UAAU,EAAE,CAAC,UAAU,EAAE,KAAK,CAAC;QAC/B,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,aAAa,CAAC,MAAM,EAAE,OAAO,CAAC;KACtE;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,CAAC;KAC1D;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,OAAO,EAAE,EAAE,CAAC,IAAI,WAAW,CAAC,MAAM,CAAC;KAC3D;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,yBAAyB,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,CAAC;IACpE,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"}
|
package/dist/writers/json.d.ts
CHANGED
|
@@ -6,7 +6,7 @@ export declare class JSONWriter implements TabularWriter {
|
|
|
6
6
|
constructor(output: Writable | string);
|
|
7
7
|
private getOutputStream;
|
|
8
8
|
private writeChunk;
|
|
9
|
-
write(dataStream: DataStream,
|
|
9
|
+
write(dataStream: DataStream, options?: WriterOptions): Promise<void>;
|
|
10
10
|
close(): Promise<void>;
|
|
11
11
|
}
|
|
12
12
|
//# sourceMappingURL=json.d.ts.map
|
|
@@ -1 +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;
|
|
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;YAST,UAAU;IAMlB,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAiBrE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAK7B"}
|
package/dist/writers/json.js
CHANGED
|
@@ -1,24 +1,25 @@
|
|
|
1
|
-
import {
|
|
1
|
+
import { closeWritableStream, openCompressedWriteStream } from "../utils/compression.js";
|
|
2
2
|
export class JSONWriter {
|
|
3
3
|
output;
|
|
4
4
|
createdStream;
|
|
5
5
|
constructor(output) {
|
|
6
6
|
this.output = output;
|
|
7
7
|
}
|
|
8
|
-
getOutputStream() {
|
|
8
|
+
getOutputStream(opts) {
|
|
9
|
+
const gzip = opts?.gzip;
|
|
10
|
+
const stream = openCompressedWriteStream(this.output, { gzip });
|
|
9
11
|
if (typeof this.output === "string") {
|
|
10
|
-
this.createdStream =
|
|
11
|
-
return this.createdStream;
|
|
12
|
+
this.createdStream = stream;
|
|
12
13
|
}
|
|
13
|
-
return
|
|
14
|
+
return stream;
|
|
14
15
|
}
|
|
15
16
|
async writeChunk(stream, data) {
|
|
16
17
|
if (!stream.write(data)) {
|
|
17
18
|
await new Promise((resolve) => stream.once("drain", resolve));
|
|
18
19
|
}
|
|
19
20
|
}
|
|
20
|
-
async write(dataStream,
|
|
21
|
-
const outStream = this.getOutputStream();
|
|
21
|
+
async write(dataStream, options) {
|
|
22
|
+
const outStream = this.getOutputStream(options);
|
|
22
23
|
let isFirstRow = true;
|
|
23
24
|
await this.writeChunk(outStream, "[\n");
|
|
24
25
|
for await (const batch of dataStream) {
|
|
@@ -32,10 +33,7 @@ export class JSONWriter {
|
|
|
32
33
|
}
|
|
33
34
|
async close() {
|
|
34
35
|
if (this.createdStream) {
|
|
35
|
-
await
|
|
36
|
-
this.createdStream.end(() => resolve());
|
|
37
|
-
this.createdStream.on("error", reject);
|
|
38
|
-
});
|
|
36
|
+
await closeWritableStream(this.createdStream);
|
|
39
37
|
}
|
|
40
38
|
}
|
|
41
39
|
}
|
package/dist/writers/json.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"json.js","sourceRoot":"","sources":["../../src/writers/json.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,
|
|
1
|
+
{"version":3,"file":"json.js","sourceRoot":"","sources":["../../src/writers/json.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,yBAAyB,EAAE,MAAM,yBAAyB,CAAC;AAIzF,MAAM,OAAO,UAAU;IACb,MAAM,CAAoB;IAC1B,aAAa,CAAY;IAEjC,YAAY,MAAyB;QACnC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAEO,eAAe,CAAC,IAAoB;QAC1C,MAAM,IAAI,GAAG,IAAI,EAAE,IAAI,CAAC;QACxB,MAAM,MAAM,GAAG,yBAAyB,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAChE,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC;QAC9B,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,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,SAAS,GAAG,IAAI,CAAC,eAAe,CAAC,OAAO,CAAC,CAAC;QAChD,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,mBAAmB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;CACF"}
|
package/dist/writers/jsonl.d.ts
CHANGED
|
@@ -6,7 +6,7 @@ export declare class JSONLWriter implements TabularWriter {
|
|
|
6
6
|
constructor(output: Writable | string);
|
|
7
7
|
private getOutputStream;
|
|
8
8
|
private writeChunk;
|
|
9
|
-
write(dataStream: DataStream,
|
|
9
|
+
write(dataStream: DataStream, options?: WriterOptions): Promise<void>;
|
|
10
10
|
close(): Promise<void>;
|
|
11
11
|
}
|
|
12
12
|
//# sourceMappingURL=jsonl.d.ts.map
|
|
@@ -1 +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;
|
|
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;YAST,UAAU;IAMlB,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAcrE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAK7B"}
|
package/dist/writers/jsonl.js
CHANGED
|
@@ -1,24 +1,25 @@
|
|
|
1
|
-
import {
|
|
1
|
+
import { closeWritableStream, openCompressedWriteStream } from "../utils/compression.js";
|
|
2
2
|
export class JSONLWriter {
|
|
3
3
|
output;
|
|
4
4
|
createdStream;
|
|
5
5
|
constructor(output) {
|
|
6
6
|
this.output = output;
|
|
7
7
|
}
|
|
8
|
-
getOutputStream() {
|
|
8
|
+
getOutputStream(opts) {
|
|
9
|
+
const gzip = opts?.gzip;
|
|
10
|
+
const stream = openCompressedWriteStream(this.output, { gzip });
|
|
9
11
|
if (typeof this.output === "string") {
|
|
10
|
-
this.createdStream =
|
|
11
|
-
return this.createdStream;
|
|
12
|
+
this.createdStream = stream;
|
|
12
13
|
}
|
|
13
|
-
return
|
|
14
|
+
return stream;
|
|
14
15
|
}
|
|
15
16
|
async writeChunk(stream, data) {
|
|
16
17
|
if (!stream.write(data)) {
|
|
17
18
|
await new Promise((resolve) => stream.once("drain", resolve));
|
|
18
19
|
}
|
|
19
20
|
}
|
|
20
|
-
async write(dataStream,
|
|
21
|
-
const outStream = this.getOutputStream();
|
|
21
|
+
async write(dataStream, options) {
|
|
22
|
+
const outStream = this.getOutputStream(options);
|
|
22
23
|
for await (const batch of dataStream) {
|
|
23
24
|
let chunkText = "";
|
|
24
25
|
for (const row of batch.rows) {
|
|
@@ -31,10 +32,7 @@ export class JSONLWriter {
|
|
|
31
32
|
}
|
|
32
33
|
async close() {
|
|
33
34
|
if (this.createdStream) {
|
|
34
|
-
await
|
|
35
|
-
this.createdStream.end(() => resolve());
|
|
36
|
-
this.createdStream.on("error", reject);
|
|
37
|
-
});
|
|
35
|
+
await closeWritableStream(this.createdStream);
|
|
38
36
|
}
|
|
39
37
|
}
|
|
40
38
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"jsonl.js","sourceRoot":"","sources":["../../src/writers/jsonl.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,
|
|
1
|
+
{"version":3,"file":"jsonl.js","sourceRoot":"","sources":["../../src/writers/jsonl.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,yBAAyB,EAAE,MAAM,yBAAyB,CAAC;AAIzF,MAAM,OAAO,WAAW;IACd,MAAM,CAAoB;IAC1B,aAAa,CAAY;IAEjC,YAAY,MAAyB;QACnC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAEO,eAAe,CAAC,IAAoB;QAC1C,MAAM,IAAI,GAAG,IAAI,EAAE,IAAI,CAAC;QACxB,MAAM,MAAM,GAAG,yBAAyB,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAChE,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC;QAC9B,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,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,SAAS,GAAG,IAAI,CAAC,eAAe,CAAC,OAAO,CAAC,CAAC;QAEhD,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,mBAAmB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { DataStream, TabularWriter, WriterOptions } from "../core/types.js";
|
|
3
|
+
export interface MarkdownWriterOptions extends WriterOptions {
|
|
4
|
+
headers?: string[];
|
|
5
|
+
}
|
|
6
|
+
/**
|
|
7
|
+
* Streaming GitHub-Flavored Markdown Table Writer.
|
|
8
|
+
* Formats rows into clean markdown tables with proper cell escaping and backpressure.
|
|
9
|
+
*/
|
|
10
|
+
export declare class MarkdownWriter implements TabularWriter {
|
|
11
|
+
private output;
|
|
12
|
+
private options;
|
|
13
|
+
private createdStream?;
|
|
14
|
+
constructor(output: Writable | string, options?: MarkdownWriterOptions);
|
|
15
|
+
private getOutputStream;
|
|
16
|
+
private escapeCell;
|
|
17
|
+
private writeChunk;
|
|
18
|
+
write(dataStream: DataStream, options?: WriterOptions): Promise<void>;
|
|
19
|
+
close(): Promise<void>;
|
|
20
|
+
}
|
|
21
|
+
//# sourceMappingURL=markdown.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"markdown.d.ts","sourceRoot":"","sources":["../../src/writers/markdown.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,qBAAsB,SAAQ,aAAa;IAC1D,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;CACpB;AAED;;;GAGG;AACH,qBAAa,cAAe,YAAW,aAAa;IAClD,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,OAAO,CAAwB;IACvC,OAAO,CAAC,aAAa,CAAC,CAAW;gBAErB,MAAM,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,qBAA0B;IAK1E,OAAO,CAAC,eAAe;IASvB,OAAO,CAAC,UAAU;YAUJ,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;CAK7B"}
|
|
@@ -0,0 +1,82 @@
|
|
|
1
|
+
import { closeWritableStream, openCompressedWriteStream } from "../utils/compression.js";
|
|
2
|
+
/**
|
|
3
|
+
* Streaming GitHub-Flavored Markdown Table Writer.
|
|
4
|
+
* Formats rows into clean markdown tables with proper cell escaping and backpressure.
|
|
5
|
+
*/
|
|
6
|
+
export class MarkdownWriter {
|
|
7
|
+
output;
|
|
8
|
+
options;
|
|
9
|
+
createdStream;
|
|
10
|
+
constructor(output, options = {}) {
|
|
11
|
+
this.output = output;
|
|
12
|
+
this.options = { ...options };
|
|
13
|
+
}
|
|
14
|
+
getOutputStream(opts) {
|
|
15
|
+
const gzip = opts?.gzip ?? this.options.gzip;
|
|
16
|
+
const stream = openCompressedWriteStream(this.output, { gzip });
|
|
17
|
+
if (typeof this.output === "string") {
|
|
18
|
+
this.createdStream = stream;
|
|
19
|
+
}
|
|
20
|
+
return stream;
|
|
21
|
+
}
|
|
22
|
+
escapeCell(value) {
|
|
23
|
+
if (value === null || value === undefined) {
|
|
24
|
+
return "";
|
|
25
|
+
}
|
|
26
|
+
const str = typeof value === "object" ? JSON.stringify(value) : String(value);
|
|
27
|
+
return str
|
|
28
|
+
.replace(/\|/g, "\\|")
|
|
29
|
+
.replace(/\r?\n/g, "<br>");
|
|
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 outStream = this.getOutputStream(mergedOptions);
|
|
42
|
+
let headers = mergedOptions.headers ?? null;
|
|
43
|
+
for await (const batch of dataStream) {
|
|
44
|
+
if (batch.rows.length === 0)
|
|
45
|
+
continue;
|
|
46
|
+
if (headers === null) {
|
|
47
|
+
headers = Object.keys(batch.rows[0] || {});
|
|
48
|
+
if (headers.length === 0) {
|
|
49
|
+
headers = ["value"];
|
|
50
|
+
}
|
|
51
|
+
// Write header row: | Col1 | Col2 |
|
|
52
|
+
const headerLine = `| ${headers.join(" | ")} |\n`;
|
|
53
|
+
// Write separator row: |---|---|
|
|
54
|
+
const separatorLine = `| ${headers.map(() => "---").join(" | ")} |\n`;
|
|
55
|
+
await this.writeChunk(outStream, headerLine + separatorLine);
|
|
56
|
+
}
|
|
57
|
+
let chunk = "";
|
|
58
|
+
for (let r = 0; r < batch.rows.length; r++) {
|
|
59
|
+
const row = batch.rows[r];
|
|
60
|
+
const cells = new Array(headers.length);
|
|
61
|
+
for (let i = 0; i < headers.length; i++) {
|
|
62
|
+
const colName = headers[i];
|
|
63
|
+
cells[i] = this.escapeCell(row[colName]);
|
|
64
|
+
}
|
|
65
|
+
chunk += `| ${cells.join(" | ")} |\n`;
|
|
66
|
+
}
|
|
67
|
+
if (chunk.length > 0) {
|
|
68
|
+
await this.writeChunk(outStream, chunk);
|
|
69
|
+
}
|
|
70
|
+
}
|
|
71
|
+
if (headers === null) {
|
|
72
|
+
// Empty dataset
|
|
73
|
+
await this.writeChunk(outStream, "| (empty) |\n|---|\n");
|
|
74
|
+
}
|
|
75
|
+
}
|
|
76
|
+
async close() {
|
|
77
|
+
if (this.createdStream) {
|
|
78
|
+
await closeWritableStream(this.createdStream);
|
|
79
|
+
}
|
|
80
|
+
}
|
|
81
|
+
}
|
|
82
|
+
//# sourceMappingURL=markdown.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"markdown.js","sourceRoot":"","sources":["../../src/writers/markdown.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,yBAAyB,EAAE,MAAM,yBAAyB,CAAC;AAQzF;;;GAGG;AACH,MAAM,OAAO,cAAc;IACjB,MAAM,CAAoB;IAC1B,OAAO,CAAwB;IAC/B,aAAa,CAAY;IAEjC,YAAY,MAAyB,EAAE,UAAiC,EAAE;QACxE,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,OAAO,EAAE,CAAC;IAChC,CAAC;IAEO,eAAe,CAAC,IAAoB;QAC1C,MAAM,IAAI,GAAG,IAAI,EAAE,IAAI,IAAI,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC;QAC7C,MAAM,MAAM,GAAG,yBAAyB,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAChE,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YACpC,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC;QAC9B,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC;IAEO,UAAU,CAAC,KAAc;QAC/B,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YAC1C,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,GAAG,GAAG,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAC9E,OAAO,GAAG;aACP,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC;aACrB,OAAO,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;IAC/B,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,GAA0B;YAC3C,GAAG,IAAI,CAAC,OAAO;YACf,GAAG,OAAO;SACX,CAAC;QAEF,MAAM,SAAS,GAAG,IAAI,CAAC,eAAe,CAAC,aAAa,CAAC,CAAC;QACtD,IAAI,OAAO,GAAoB,aAAa,CAAC,OAAO,IAAI,IAAI,CAAC;QAE7D,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,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;gBAC3C,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;oBACzB,OAAO,GAAG,CAAC,OAAO,CAAC,CAAC;gBACtB,CAAC;gBAED,oCAAoC;gBACpC,MAAM,UAAU,GAAG,KAAK,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;gBAClD,iCAAiC;gBACjC,MAAM,aAAa,GAAG,KAAK,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;gBACtE,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,UAAU,GAAG,aAAa,CAAC,CAAC;YAC/D,CAAC;YAED,IAAI,KAAK,GAAG,EAAE,CAAC;YACf,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;gBAC3B,MAAM,KAAK,GAAa,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;gBAClD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACxC,MAAM,OAAO,GAAG,OAAO,CAAC,CAAC,CAAE,CAAC;oBAC5B,KAAK,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;gBAC3C,CAAC;gBACD,KAAK,IAAI,KAAK,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;YACxC,CAAC;YAED,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC;YAC1C,CAAC;QACH,CAAC;QAED,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;YACrB,gBAAgB;YAChB,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,sBAAsB,CAAC,CAAC;QAC3D,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAK;QACT,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;YACvB,MAAM,mBAAmB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,19 @@
|
|
|
1
|
+
import type { Writable } from "node:stream";
|
|
2
|
+
import type { DataStream, TabularWriter, WriterOptions } from "../core/types.js";
|
|
3
|
+
export interface ParquetWriterOptions extends WriterOptions {
|
|
4
|
+
schema?: Record<string, any>;
|
|
5
|
+
compression?: "UNCOMPRESSED" | "SNAPPY" | "GZIP";
|
|
6
|
+
}
|
|
7
|
+
/**
|
|
8
|
+
* Streaming Apache Parquet Writer.
|
|
9
|
+
* Batches rows and writes binary columnar data with Snappy compression.
|
|
10
|
+
*/
|
|
11
|
+
export declare class ParquetWriter implements TabularWriter {
|
|
12
|
+
private output;
|
|
13
|
+
private options;
|
|
14
|
+
private writer?;
|
|
15
|
+
constructor(output: Writable | string, options?: ParquetWriterOptions);
|
|
16
|
+
write(dataStream: DataStream, options?: WriterOptions): Promise<void>;
|
|
17
|
+
close(): Promise<void>;
|
|
18
|
+
}
|
|
19
|
+
//# sourceMappingURL=parquet.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parquet.d.ts","sourceRoot":"","sources":["../../src/writers/parquet.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAG5C,OAAO,KAAK,EAAE,UAAU,EAAO,aAAa,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AAEtF,MAAM,WAAW,oBAAqB,SAAQ,aAAa;IACzD,MAAM,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAC7B,WAAW,CAAC,EAAE,cAAc,GAAG,QAAQ,GAAG,MAAM,CAAC;CAClD;AAiED;;;GAGG;AACH,qBAAa,aAAc,YAAW,aAAa;IACjD,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,OAAO,CAAuB;IACtC,OAAO,CAAC,MAAM,CAAC,CAAM;gBAET,MAAM,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,oBAAyB;IAKnE,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAiDrE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAS7B"}
|
|
@@ -0,0 +1,129 @@
|
|
|
1
|
+
// @ts-ignore
|
|
2
|
+
import parquet from "parquetjs-lite";
|
|
3
|
+
/**
|
|
4
|
+
* Infers a Parquet schema from sample rows.
|
|
5
|
+
*/
|
|
6
|
+
function inferParquetSchema(sampleRows) {
|
|
7
|
+
const schemaDef = {};
|
|
8
|
+
for (const row of sampleRows) {
|
|
9
|
+
for (const [key, val] of Object.entries(row)) {
|
|
10
|
+
if (schemaDef[key] && schemaDef[key].type === "DOUBLE") {
|
|
11
|
+
continue;
|
|
12
|
+
}
|
|
13
|
+
if (typeof val === "number") {
|
|
14
|
+
if (!schemaDef[key] || schemaDef[key].type === "INT64") {
|
|
15
|
+
schemaDef[key] = {
|
|
16
|
+
type: Number.isInteger(val) ? "INT64" : "DOUBLE",
|
|
17
|
+
optional: true,
|
|
18
|
+
};
|
|
19
|
+
}
|
|
20
|
+
}
|
|
21
|
+
else if (typeof val === "boolean") {
|
|
22
|
+
schemaDef[key] = { type: "BOOLEAN", optional: true };
|
|
23
|
+
}
|
|
24
|
+
else {
|
|
25
|
+
schemaDef[key] = { type: "UTF8", optional: true };
|
|
26
|
+
}
|
|
27
|
+
}
|
|
28
|
+
}
|
|
29
|
+
// Ensure at least one field
|
|
30
|
+
if (Object.keys(schemaDef).length === 0) {
|
|
31
|
+
schemaDef["value"] = { type: "UTF8", optional: true };
|
|
32
|
+
}
|
|
33
|
+
return new parquet.ParquetSchema(schemaDef);
|
|
34
|
+
}
|
|
35
|
+
/**
|
|
36
|
+
* Sanitizes a row according to schema definitions before passing to ParquetWriter.
|
|
37
|
+
*/
|
|
38
|
+
function sanitizeRowForParquet(row, schemaDef) {
|
|
39
|
+
const sanitized = {};
|
|
40
|
+
for (const [key, fieldDef] of Object.entries(schemaDef)) {
|
|
41
|
+
const rawVal = row[key];
|
|
42
|
+
if (rawVal === null || rawVal === undefined) {
|
|
43
|
+
continue;
|
|
44
|
+
}
|
|
45
|
+
if (fieldDef.type === "INT64") {
|
|
46
|
+
const num = Number(rawVal);
|
|
47
|
+
sanitized[key] = isNaN(num) ? 0 : Math.trunc(num);
|
|
48
|
+
}
|
|
49
|
+
else if (fieldDef.type === "DOUBLE" || fieldDef.type === "FLOAT") {
|
|
50
|
+
const num = Number(rawVal);
|
|
51
|
+
sanitized[key] = isNaN(num) ? 0 : num;
|
|
52
|
+
}
|
|
53
|
+
else if (fieldDef.type === "BOOLEAN") {
|
|
54
|
+
sanitized[key] = Boolean(rawVal);
|
|
55
|
+
}
|
|
56
|
+
else {
|
|
57
|
+
sanitized[key] = typeof rawVal === "object" ? JSON.stringify(rawVal) : String(rawVal);
|
|
58
|
+
}
|
|
59
|
+
}
|
|
60
|
+
return sanitized;
|
|
61
|
+
}
|
|
62
|
+
/**
|
|
63
|
+
* Streaming Apache Parquet Writer.
|
|
64
|
+
* Batches rows and writes binary columnar data with Snappy compression.
|
|
65
|
+
*/
|
|
66
|
+
export class ParquetWriter {
|
|
67
|
+
output;
|
|
68
|
+
options;
|
|
69
|
+
writer;
|
|
70
|
+
constructor(output, options = {}) {
|
|
71
|
+
this.output = output;
|
|
72
|
+
this.options = { ...options };
|
|
73
|
+
}
|
|
74
|
+
async write(dataStream, options) {
|
|
75
|
+
const mergedOptions = {
|
|
76
|
+
...this.options,
|
|
77
|
+
...options,
|
|
78
|
+
};
|
|
79
|
+
let parquetSchema = null;
|
|
80
|
+
let schemaFields = {};
|
|
81
|
+
for await (const batch of dataStream) {
|
|
82
|
+
if (batch.rows.length === 0)
|
|
83
|
+
continue;
|
|
84
|
+
if (!this.writer) {
|
|
85
|
+
parquetSchema = inferParquetSchema(batch.rows);
|
|
86
|
+
schemaFields = parquetSchema.schema;
|
|
87
|
+
if (typeof this.output === "string") {
|
|
88
|
+
this.writer = await parquet.ParquetWriter.openFile(parquetSchema, this.output, {
|
|
89
|
+
useDataPageV2: false,
|
|
90
|
+
});
|
|
91
|
+
}
|
|
92
|
+
else {
|
|
93
|
+
this.writer = await parquet.ParquetWriter.openStream(parquetSchema, this.output, {
|
|
94
|
+
useDataPageV2: false,
|
|
95
|
+
});
|
|
96
|
+
}
|
|
97
|
+
}
|
|
98
|
+
for (let i = 0; i < batch.rows.length; i++) {
|
|
99
|
+
const row = batch.rows[i];
|
|
100
|
+
const sanitized = sanitizeRowForParquet(row, schemaFields);
|
|
101
|
+
await this.writer.appendRow(sanitized);
|
|
102
|
+
}
|
|
103
|
+
}
|
|
104
|
+
if (!this.writer) {
|
|
105
|
+
// Empty stream fallback
|
|
106
|
+
parquetSchema = new parquet.ParquetSchema({
|
|
107
|
+
empty: { type: "UTF8", optional: true },
|
|
108
|
+
});
|
|
109
|
+
if (typeof this.output === "string") {
|
|
110
|
+
this.writer = await parquet.ParquetWriter.openFile(parquetSchema, this.output);
|
|
111
|
+
}
|
|
112
|
+
else {
|
|
113
|
+
this.writer = await parquet.ParquetWriter.openStream(parquetSchema, this.output);
|
|
114
|
+
}
|
|
115
|
+
}
|
|
116
|
+
await this.writer.close();
|
|
117
|
+
}
|
|
118
|
+
async close() {
|
|
119
|
+
if (this.writer) {
|
|
120
|
+
try {
|
|
121
|
+
await this.writer.close();
|
|
122
|
+
}
|
|
123
|
+
catch {
|
|
124
|
+
// Writer already closed
|
|
125
|
+
}
|
|
126
|
+
}
|
|
127
|
+
}
|
|
128
|
+
}
|
|
129
|
+
//# sourceMappingURL=parquet.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parquet.js","sourceRoot":"","sources":["../../src/writers/parquet.ts"],"names":[],"mappings":"AAEA,aAAa;AACb,OAAO,OAAO,MAAM,gBAAgB,CAAC;AAQrC;;GAEG;AACH,SAAS,kBAAkB,CAAC,UAAiB;IAC3C,MAAM,SAAS,GAAwD,EAAE,CAAC;IAE1E,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,KAAK,MAAM,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;YAC7C,IAAI,SAAS,CAAC,GAAG,CAAC,IAAI,SAAS,CAAC,GAAG,CAAE,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBACxD,SAAS;YACX,CAAC;YAED,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;gBAC5B,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,SAAS,CAAC,GAAG,CAAE,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;oBACxD,SAAS,CAAC,GAAG,CAAC,GAAG;wBACf,IAAI,EAAE,MAAM,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ;wBAChD,QAAQ,EAAE,IAAI;qBACf,CAAC;gBACJ,CAAC;YACH,CAAC;iBAAM,IAAI,OAAO,GAAG,KAAK,SAAS,EAAE,CAAC;gBACpC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;YACvD,CAAC;iBAAM,CAAC;gBACN,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;YACpD,CAAC;QACH,CAAC;IACH,CAAC;IAED,4BAA4B;IAC5B,IAAI,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxC,SAAS,CAAC,OAAO,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;IACxD,CAAC;IAED,OAAO,IAAI,OAAO,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC;AAC9C,CAAC;AAED;;GAEG;AACH,SAAS,qBAAqB,CAAC,GAAQ,EAAE,SAA8B;IACrE,MAAM,SAAS,GAAwB,EAAE,CAAC;IAE1C,KAAK,MAAM,CAAC,GAAG,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;QACxD,MAAM,MAAM,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC;QACxB,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YAC5C,SAAS;QACX,CAAC;QAED,IAAI,QAAQ,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC9B,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;YAC3B,SAAS,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACpD,CAAC;aAAM,IAAI,QAAQ,CAAC,IAAI,KAAK,QAAQ,IAAI,QAAQ,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YACnE,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;YAC3B,SAAS,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;QACxC,CAAC;aAAM,IAAI,QAAQ,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YACvC,SAAS,CAAC,GAAG,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACnC,CAAC;aAAM,CAAC;YACN,SAAS,CAAC,GAAG,CAAC,GAAG,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;QACxF,CAAC;IACH,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,aAAa;IAChB,MAAM,CAAoB;IAC1B,OAAO,CAAuB;IAC9B,MAAM,CAAO;IAErB,YAAY,MAAyB,EAAE,UAAgC,EAAE;QACvE,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,GAAyB;YAC1C,GAAG,IAAI,CAAC,OAAO;YACf,GAAG,OAAO;SACX,CAAC;QAEF,IAAI,aAAa,GAAQ,IAAI,CAAC;QAC9B,IAAI,YAAY,GAAwB,EAAE,CAAC;QAE3C,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;YACrC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAEtC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;gBACjB,aAAa,GAAG,kBAAkB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;gBAC/C,YAAY,GAAG,aAAa,CAAC,MAAM,CAAC;gBAEpC,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;oBACpC,IAAI,CAAC,MAAM,GAAG,MAAM,OAAO,CAAC,aAAa,CAAC,QAAQ,CAAC,aAAa,EAAE,IAAI,CAAC,MAAM,EAAE;wBAC7E,aAAa,EAAE,KAAK;qBACrB,CAAC,CAAC;gBACL,CAAC;qBAAM,CAAC;oBACN,IAAI,CAAC,MAAM,GAAG,MAAM,OAAO,CAAC,aAAa,CAAC,UAAU,CAAC,aAAa,EAAE,IAAI,CAAC,MAAM,EAAE;wBAC/E,aAAa,EAAE,KAAK;qBACrB,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;gBAC3B,MAAM,SAAS,GAAG,qBAAqB,CAAC,GAAG,EAAE,YAAY,CAAC,CAAC;gBAC3D,MAAM,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,SAAS,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;QAED,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACjB,wBAAwB;YACxB,aAAa,GAAG,IAAI,OAAO,CAAC,aAAa,CAAC;gBACxC,KAAK,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,IAAI,EAAE;aACxC,CAAC,CAAC;YACH,IAAI,OAAO,IAAI,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;gBACpC,IAAI,CAAC,MAAM,GAAG,MAAM,OAAO,CAAC,aAAa,CAAC,QAAQ,CAAC,aAAa,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;YACjF,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,MAAM,GAAG,MAAM,OAAO,CAAC,aAAa,CAAC,UAAU,CAAC,aAAa,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;YACnF,CAAC;QACH,CAAC;QAED,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;IAC5B,CAAC;IAED,KAAK,CAAC,KAAK;QACT,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,IAAI,CAAC;gBACH,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;YAC5B,CAAC;YAAC,MAAM,CAAC;gBACP,wBAAwB;YAC1B,CAAC;QACH,CAAC;IACH,CAAC;CACF"}
|
package/package.json
CHANGED
|
@@ -1,26 +1,36 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "rowpipe",
|
|
3
|
-
"version": "1.
|
|
4
|
-
"description": "
|
|
3
|
+
"version": "1.2.0",
|
|
4
|
+
"description": "Stream-first tabular data toolkit for CSV, TSV, JSON, JSONL, XLSX, Apache Parquet, and Markdown",
|
|
5
5
|
"main": "dist/index.js",
|
|
6
6
|
"types": "dist/index.d.ts",
|
|
7
7
|
"bin": {
|
|
8
8
|
"rowpipe": "dist/cli/index.js"
|
|
9
9
|
},
|
|
10
|
+
"files": [
|
|
11
|
+
"dist",
|
|
12
|
+
"skills",
|
|
13
|
+
"README.md"
|
|
14
|
+
],
|
|
10
15
|
"type": "module",
|
|
11
16
|
"scripts": {
|
|
12
17
|
"build": "tsc",
|
|
13
18
|
"dev": "tsx src/cli/index.ts",
|
|
14
19
|
"test": "vitest run",
|
|
15
20
|
"test:watch": "vitest",
|
|
16
|
-
"bench": "tsx benchmarks/memory-bench.ts"
|
|
21
|
+
"bench": "tsx benchmarks/memory-bench.ts",
|
|
22
|
+
"benchmark:diff": "tsx benchmarks/diff-bench.ts",
|
|
23
|
+
"benchmark:files": "tsx benchmarks/files-bench.ts"
|
|
17
24
|
},
|
|
18
25
|
"keywords": [
|
|
19
26
|
"csv",
|
|
27
|
+
"tsv",
|
|
20
28
|
"json",
|
|
21
29
|
"jsonl",
|
|
22
30
|
"ndjson",
|
|
31
|
+
"parquet",
|
|
23
32
|
"xlsx",
|
|
33
|
+
"markdown",
|
|
24
34
|
"streaming",
|
|
25
35
|
"tabular",
|
|
26
36
|
"data",
|
|
@@ -32,6 +42,9 @@
|
|
|
32
42
|
"license": "MIT",
|
|
33
43
|
"dependencies": {
|
|
34
44
|
"commander": "^13.1.0",
|
|
45
|
+
"hyparquet": "^1.30.1",
|
|
46
|
+
"hyparquet-compressors": "^1.1.1",
|
|
47
|
+
"parquetjs-lite": "^0.8.7",
|
|
35
48
|
"read-excel-file": "^9.3.10",
|
|
36
49
|
"write-excel-file": "^4.1.1"
|
|
37
50
|
},
|