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 +1,95 @@
|
|
|
1
|
+
import { createReadStream } from "node:fs";
|
|
2
|
+
import { createInterface } from "node:readline";
|
|
3
|
+
import { ParseError } from "../core/errors.js";
|
|
4
|
+
export class JSONLReader {
|
|
5
|
+
input;
|
|
6
|
+
options;
|
|
7
|
+
constructor(input, options = {}) {
|
|
8
|
+
this.input = input;
|
|
9
|
+
this.options = { ...options };
|
|
10
|
+
if (typeof input === "string") {
|
|
11
|
+
this.options.filePath = input;
|
|
12
|
+
}
|
|
13
|
+
}
|
|
14
|
+
getInputStream() {
|
|
15
|
+
if (typeof this.input === "string") {
|
|
16
|
+
return createReadStream(this.input);
|
|
17
|
+
}
|
|
18
|
+
return this.input;
|
|
19
|
+
}
|
|
20
|
+
async *read(options) {
|
|
21
|
+
const mergedOptions = {
|
|
22
|
+
...this.options,
|
|
23
|
+
...options,
|
|
24
|
+
};
|
|
25
|
+
const batchSize = Math.max(1, mergedOptions.batchSize ?? 1000);
|
|
26
|
+
const stream = this.getInputStream();
|
|
27
|
+
const rl = createInterface({
|
|
28
|
+
input: stream,
|
|
29
|
+
crlfDelay: Infinity,
|
|
30
|
+
});
|
|
31
|
+
let lineNumber = 0;
|
|
32
|
+
let rowsInCurrentBatch = [];
|
|
33
|
+
let globalOffset = 0;
|
|
34
|
+
for await (const line of rl) {
|
|
35
|
+
lineNumber++;
|
|
36
|
+
const trimmed = line.trim();
|
|
37
|
+
if (!trimmed)
|
|
38
|
+
continue;
|
|
39
|
+
let parsed;
|
|
40
|
+
try {
|
|
41
|
+
parsed = JSON.parse(trimmed);
|
|
42
|
+
}
|
|
43
|
+
catch (err) {
|
|
44
|
+
throw new ParseError(`Invalid JSON on line ${lineNumber}: ${err.message}`, {
|
|
45
|
+
file: mergedOptions.filePath,
|
|
46
|
+
line: lineNumber,
|
|
47
|
+
row: lineNumber,
|
|
48
|
+
});
|
|
49
|
+
}
|
|
50
|
+
if (typeof parsed !== "object" || parsed === null || Array.isArray(parsed)) {
|
|
51
|
+
throw new ParseError(`Expected JSON object on line ${lineNumber}, got ${Array.isArray(parsed) ? "array" : typeof parsed}`, {
|
|
52
|
+
file: mergedOptions.filePath,
|
|
53
|
+
line: lineNumber,
|
|
54
|
+
row: lineNumber,
|
|
55
|
+
});
|
|
56
|
+
}
|
|
57
|
+
rowsInCurrentBatch.push(parsed);
|
|
58
|
+
if (rowsInCurrentBatch.length >= batchSize) {
|
|
59
|
+
yield {
|
|
60
|
+
rows: rowsInCurrentBatch,
|
|
61
|
+
offset: globalOffset,
|
|
62
|
+
};
|
|
63
|
+
globalOffset += rowsInCurrentBatch.length;
|
|
64
|
+
rowsInCurrentBatch = [];
|
|
65
|
+
}
|
|
66
|
+
}
|
|
67
|
+
if (rowsInCurrentBatch.length > 0) {
|
|
68
|
+
yield {
|
|
69
|
+
rows: rowsInCurrentBatch,
|
|
70
|
+
offset: globalOffset,
|
|
71
|
+
};
|
|
72
|
+
}
|
|
73
|
+
}
|
|
74
|
+
async inspect(options) {
|
|
75
|
+
let rowCount = 0;
|
|
76
|
+
let columnNames = [];
|
|
77
|
+
for await (const batch of this.read({ ...options, batchSize: 5000 })) {
|
|
78
|
+
if (columnNames.length === 0 && batch.rows.length > 0 && batch.rows[0]) {
|
|
79
|
+
columnNames = Object.keys(batch.rows[0]);
|
|
80
|
+
}
|
|
81
|
+
rowCount += batch.rows.length;
|
|
82
|
+
}
|
|
83
|
+
return {
|
|
84
|
+
format: "JSONL",
|
|
85
|
+
rowCount,
|
|
86
|
+
columnCount: columnNames.length,
|
|
87
|
+
columns: columnNames.map((name) => ({
|
|
88
|
+
name,
|
|
89
|
+
type: "string",
|
|
90
|
+
nullPercentage: 0,
|
|
91
|
+
})),
|
|
92
|
+
};
|
|
93
|
+
}
|
|
94
|
+
}
|
|
95
|
+
//# sourceMappingURL=jsonl.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"jsonl.js","sourceRoot":"","sources":["../../src/readers/jsonl.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAC;AAC3C,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AAEhD,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAc/C,MAAM,OAAO,WAAW;IACd,KAAK,CAAoB;IACzB,OAAO,CAAqB;IAEpC,YAAY,KAAwB,EAAE,UAA8B,EAAE;QACpE,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,GAAuB;YACxC,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,MAAM,GAAG,IAAI,CAAC,cAAc,EAAE,CAAC;QAErC,MAAM,EAAE,GAAG,eAAe,CAAC;YACzB,KAAK,EAAE,MAAM;YACb,SAAS,EAAE,QAAQ;SACpB,CAAC,CAAC;QAEH,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,IAAI,kBAAkB,GAAU,EAAE,CAAC;QACnC,IAAI,YAAY,GAAG,CAAC,CAAC;QAErB,IAAI,KAAK,EAAE,MAAM,IAAI,IAAI,EAAE,EAAE,CAAC;YAC5B,UAAU,EAAE,CAAC;YACb,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;YAC5B,IAAI,CAAC,OAAO;gBAAE,SAAS;YAEvB,IAAI,MAAe,CAAC;YACpB,IAAI,CAAC;gBACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;YAC/B,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,MAAM,IAAI,UAAU,CAClB,wBAAwB,UAAU,KAAM,GAAa,CAAC,OAAO,EAAE,EAC/D;oBACE,IAAI,EAAE,aAAa,CAAC,QAAQ;oBAC5B,IAAI,EAAE,UAAU;oBAChB,GAAG,EAAE,UAAU;iBAChB,CACF,CAAC;YACJ,CAAC;YAED,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,IAAI,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC3E,MAAM,IAAI,UAAU,CAClB,gCAAgC,UAAU,SAAS,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,MAAM,EAAE,EACpG;oBACE,IAAI,EAAE,aAAa,CAAC,QAAQ;oBAC5B,IAAI,EAAE,UAAU;oBAChB,GAAG,EAAE,UAAU;iBAChB,CACF,CAAC;YACJ,CAAC;YAED,kBAAkB,CAAC,IAAI,CAAC,MAAa,CAAC,CAAC;YAEvC,IAAI,kBAAkB,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;gBAC3C,MAAM;oBACJ,IAAI,EAAE,kBAAkB;oBACxB,MAAM,EAAE,YAAY;iBACrB,CAAC;gBACF,YAAY,IAAI,kBAAkB,CAAC,MAAM,CAAC;gBAC1C,kBAAkB,GAAG,EAAE,CAAC;YAC1B,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,OAAO;YACf,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 +1,18 @@
|
|
|
1
|
+
import type { Readable } from "node:stream";
|
|
2
|
+
import type { DataStream, InspectionMetadata, ReaderOptions, TabularReader } from "../core/types.js";
|
|
3
|
+
export interface XLSXReaderOptions extends ReaderOptions {
|
|
4
|
+
sheet?: string | number;
|
|
5
|
+
filePath?: string;
|
|
6
|
+
}
|
|
7
|
+
export declare class XLSXReader implements TabularReader {
|
|
8
|
+
private input;
|
|
9
|
+
private options;
|
|
10
|
+
constructor(input: Readable | string, options?: XLSXReaderOptions);
|
|
11
|
+
getParsedSheets(): Promise<Array<{
|
|
12
|
+
sheet: string;
|
|
13
|
+
data: unknown[][];
|
|
14
|
+
}>>;
|
|
15
|
+
read(options?: ReaderOptions): DataStream;
|
|
16
|
+
inspect(options?: ReaderOptions): Promise<InspectionMetadata>;
|
|
17
|
+
}
|
|
18
|
+
//# sourceMappingURL=xlsx.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"xlsx.d.ts","sourceRoot":"","sources":["../../src/readers/xlsx.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAG5C,OAAO,KAAK,EAEV,UAAU,EACV,kBAAkB,EAClB,aAAa,EAGb,aAAa,EACd,MAAM,kBAAkB,CAAC;AAE1B,MAAM,WAAW,iBAAkB,SAAQ,aAAa;IACtD,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,qBAAa,UAAW,YAAW,aAAa;IAC9C,OAAO,CAAC,KAAK,CAAoB;IACjC,OAAO,CAAC,OAAO,CAAoB;gBAEvB,KAAK,EAAE,QAAQ,GAAG,MAAM,EAAE,OAAO,GAAE,iBAAsB;IAQ/D,eAAe,IAAI,OAAO,CAAC,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,OAAO,EAAE,EAAE,CAAA;KAAE,CAAC,CAAC;IAoBtE,IAAI,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,UAAU;IA+E1C,OAAO,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,kBAAkB,CAAC;CAoDpE"}
|
|
@@ -0,0 +1,138 @@
|
|
|
1
|
+
import readXlsxFile from "read-excel-file/node";
|
|
2
|
+
import { InvalidArgumentError, ParseError } from "../core/errors.js";
|
|
3
|
+
export class XLSXReader {
|
|
4
|
+
input;
|
|
5
|
+
options;
|
|
6
|
+
constructor(input, options = {}) {
|
|
7
|
+
this.input = input;
|
|
8
|
+
this.options = { ...options };
|
|
9
|
+
if (typeof input === "string") {
|
|
10
|
+
this.options.filePath = input;
|
|
11
|
+
}
|
|
12
|
+
}
|
|
13
|
+
async getParsedSheets() {
|
|
14
|
+
try {
|
|
15
|
+
const result = await readXlsxFile(this.input, { getSheets: true });
|
|
16
|
+
if (Array.isArray(result)) {
|
|
17
|
+
if (result.length > 0 && typeof result[0] === "object" && result[0] !== null && "sheet" in result[0]) {
|
|
18
|
+
return result;
|
|
19
|
+
}
|
|
20
|
+
return [{ sheet: "Sheet1", data: result }];
|
|
21
|
+
}
|
|
22
|
+
return [];
|
|
23
|
+
}
|
|
24
|
+
catch (err) {
|
|
25
|
+
throw new ParseError(`Failed to parse XLSX: ${err.message}`, {
|
|
26
|
+
file: this.options.filePath,
|
|
27
|
+
});
|
|
28
|
+
}
|
|
29
|
+
}
|
|
30
|
+
async *read(options) {
|
|
31
|
+
const mergedOptions = {
|
|
32
|
+
...this.options,
|
|
33
|
+
...options,
|
|
34
|
+
};
|
|
35
|
+
const batchSize = Math.max(1, mergedOptions.batchSize ?? 1000);
|
|
36
|
+
const targetSheet = mergedOptions.sheet;
|
|
37
|
+
const sheets = await this.getParsedSheets();
|
|
38
|
+
if (sheets.length === 0) {
|
|
39
|
+
return;
|
|
40
|
+
}
|
|
41
|
+
let selectedSheet;
|
|
42
|
+
if (targetSheet === undefined || targetSheet === null) {
|
|
43
|
+
selectedSheet = sheets[0];
|
|
44
|
+
}
|
|
45
|
+
else if (typeof targetSheet === "number") {
|
|
46
|
+
selectedSheet = sheets[targetSheet - 1];
|
|
47
|
+
}
|
|
48
|
+
else {
|
|
49
|
+
selectedSheet = sheets.find((s) => s.sheet.toLowerCase() === String(targetSheet).toLowerCase());
|
|
50
|
+
}
|
|
51
|
+
if (!selectedSheet) {
|
|
52
|
+
throw new InvalidArgumentError(`Sheet "${targetSheet}" was not found in the workbook. Available sheets: ${sheets.map((s) => s.sheet).join(", ")}`);
|
|
53
|
+
}
|
|
54
|
+
const grid = selectedSheet.data;
|
|
55
|
+
if (grid.length === 0)
|
|
56
|
+
return;
|
|
57
|
+
let headers = null;
|
|
58
|
+
let rowsInCurrentBatch = [];
|
|
59
|
+
let globalOffset = 0;
|
|
60
|
+
for (let r = 0; r < grid.length; r++) {
|
|
61
|
+
const rowCells = grid[r] || [];
|
|
62
|
+
if (rowCells.length === 0 || rowCells.every((c) => c === null || c === undefined || c === "")) {
|
|
63
|
+
continue;
|
|
64
|
+
}
|
|
65
|
+
if (headers === null) {
|
|
66
|
+
headers = rowCells.map((c, i) => {
|
|
67
|
+
const str = c !== null && c !== undefined ? String(c).trim() : "";
|
|
68
|
+
return str.length > 0 ? str : `col_${i + 1}`;
|
|
69
|
+
});
|
|
70
|
+
continue;
|
|
71
|
+
}
|
|
72
|
+
const rowObj = {};
|
|
73
|
+
for (let i = 0; i < headers.length; i++) {
|
|
74
|
+
const key = headers[i] || `col_${i + 1}`;
|
|
75
|
+
const cellVal = rowCells[i];
|
|
76
|
+
rowObj[key] = cellVal !== null && cellVal !== undefined ? cellVal : "";
|
|
77
|
+
}
|
|
78
|
+
rowsInCurrentBatch.push(rowObj);
|
|
79
|
+
if (rowsInCurrentBatch.length >= batchSize) {
|
|
80
|
+
yield {
|
|
81
|
+
rows: rowsInCurrentBatch,
|
|
82
|
+
offset: globalOffset,
|
|
83
|
+
};
|
|
84
|
+
globalOffset += rowsInCurrentBatch.length;
|
|
85
|
+
rowsInCurrentBatch = [];
|
|
86
|
+
}
|
|
87
|
+
}
|
|
88
|
+
if (rowsInCurrentBatch.length > 0) {
|
|
89
|
+
yield {
|
|
90
|
+
rows: rowsInCurrentBatch,
|
|
91
|
+
offset: globalOffset,
|
|
92
|
+
};
|
|
93
|
+
}
|
|
94
|
+
}
|
|
95
|
+
async inspect(options) {
|
|
96
|
+
const sheets = await this.getParsedSheets();
|
|
97
|
+
const targetSheet = options?.sheet ?? this.options.sheet;
|
|
98
|
+
const sheetSummaries = sheets.map((s) => {
|
|
99
|
+
const nonHeaderRows = Math.max(0, s.data.length - 1);
|
|
100
|
+
const headerRow = s.data[0] || [];
|
|
101
|
+
const colHeaders = headerRow.map((c, i) => c !== null && c !== undefined ? String(c).trim() : `col_${i + 1}`);
|
|
102
|
+
return {
|
|
103
|
+
name: s.sheet,
|
|
104
|
+
rowCount: nonHeaderRows,
|
|
105
|
+
columnCount: colHeaders.length,
|
|
106
|
+
columns: colHeaders,
|
|
107
|
+
};
|
|
108
|
+
});
|
|
109
|
+
let selectedSheet = sheets[0];
|
|
110
|
+
if (targetSheet !== undefined && targetSheet !== null) {
|
|
111
|
+
selectedSheet =
|
|
112
|
+
typeof targetSheet === "number"
|
|
113
|
+
? sheets[targetSheet - 1]
|
|
114
|
+
: sheets.find((s) => s.sheet.toLowerCase() === String(targetSheet).toLowerCase());
|
|
115
|
+
}
|
|
116
|
+
let columns = [];
|
|
117
|
+
let rowCount = 0;
|
|
118
|
+
if (selectedSheet && selectedSheet.data.length > 0) {
|
|
119
|
+
const headerRow = selectedSheet.data[0] || [];
|
|
120
|
+
const colHeaders = headerRow.map((c, i) => c !== null && c !== undefined ? String(c).trim() : `col_${i + 1}`);
|
|
121
|
+
rowCount = Math.max(0, selectedSheet.data.length - 1);
|
|
122
|
+
columns = colHeaders.map((name) => ({
|
|
123
|
+
name,
|
|
124
|
+
type: "string",
|
|
125
|
+
nullPercentage: 0,
|
|
126
|
+
}));
|
|
127
|
+
}
|
|
128
|
+
return {
|
|
129
|
+
format: "XLSX",
|
|
130
|
+
rowCount,
|
|
131
|
+
columnCount: columns.length,
|
|
132
|
+
columns,
|
|
133
|
+
sheetsCount: sheets.length,
|
|
134
|
+
sheets: sheetSummaries,
|
|
135
|
+
};
|
|
136
|
+
}
|
|
137
|
+
}
|
|
138
|
+
//# sourceMappingURL=xlsx.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"xlsx.js","sourceRoot":"","sources":["../../src/readers/xlsx.ts"],"names":[],"mappings":"AACA,OAAO,YAAY,MAAM,sBAAsB,CAAC;AAChD,OAAO,EAAE,oBAAoB,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAgBrE,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;IAED,KAAK,CAAC,eAAe;QACnB,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAO,YAAmF,CACvG,IAAI,CAAC,KAAK,EACV,EAAE,SAAS,EAAE,IAAI,EAAE,CACpB,CAAC;YACF,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC1B,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,MAAM,CAAC,CAAC,CAAC,KAAK,QAAQ,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,OAAO,IAAI,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;oBACrG,OAAO,MAAqD,CAAC;gBAC/D,CAAC;gBACD,OAAO,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAqB,EAAE,CAAC,CAAC;YAC5D,CAAC;YACD,OAAO,EAAE,CAAC;QACZ,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,IAAI,UAAU,CAAC,yBAA0B,GAAa,CAAC,OAAO,EAAE,EAAE;gBACtE,IAAI,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ;aAC5B,CAAC,CAAC;QACL,CAAC;IACH,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,WAAW,GAAG,aAAa,CAAC,KAAK,CAAC;QAExC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,eAAe,EAAE,CAAC;QAC5C,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACxB,OAAO;QACT,CAAC;QAED,IAAI,aAA+D,CAAC;QAEpE,IAAI,WAAW,KAAK,SAAS,IAAI,WAAW,KAAK,IAAI,EAAE,CAAC;YACtD,aAAa,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QAC5B,CAAC;aAAM,IAAI,OAAO,WAAW,KAAK,QAAQ,EAAE,CAAC;YAC3C,aAAa,GAAG,MAAM,CAAC,WAAW,GAAG,CAAC,CAAC,CAAC;QAC1C,CAAC;aAAM,CAAC;YACN,aAAa,GAAG,MAAM,CAAC,IAAI,CACzB,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,MAAM,CAAC,WAAW,CAAC,CAAC,WAAW,EAAE,CACnE,CAAC;QACJ,CAAC;QAED,IAAI,CAAC,aAAa,EAAE,CAAC;YACnB,MAAM,IAAI,oBAAoB,CAC5B,UAAU,WAAW,sDAAsD,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACnH,CAAC;QACJ,CAAC;QAED,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC;QAChC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAE9B,IAAI,OAAO,GAAoB,IAAI,CAAC;QACpC,IAAI,kBAAkB,GAAU,EAAE,CAAC;QACnC,IAAI,YAAY,GAAG,CAAC,CAAC;QAErB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACrC,MAAM,QAAQ,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAC/B,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,IAAI,CAAC,KAAK,EAAE,CAAC,EAAE,CAAC;gBAC9F,SAAS;YACX,CAAC;YAED,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;gBACrB,OAAO,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;oBAC9B,MAAM,GAAG,GAAG,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;oBAClE,OAAO,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC/C,CAAC,CAAC,CAAC;gBACH,SAAS;YACX,CAAC;YAED,MAAM,MAAM,GAAQ,EAAE,CAAC;YACvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACxC,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;gBACzC,MAAM,OAAO,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;gBAC5B,MAAM,CAAC,GAAG,CAAC,GAAG,OAAO,KAAK,IAAI,IAAI,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;YACzE,CAAC;YAED,kBAAkB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAEhC,IAAI,kBAAkB,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;gBAC3C,MAAM;oBACJ,IAAI,EAAE,kBAAkB;oBACxB,MAAM,EAAE,YAAY;iBACrB,CAAC;gBACF,YAAY,IAAI,kBAAkB,CAAC,MAAM,CAAC;gBAC1C,kBAAkB,GAAG,EAAE,CAAC;YAC1B,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,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,eAAe,EAAE,CAAC;QAC5C,MAAM,WAAW,GAAG,OAAO,EAAE,KAAK,IAAI,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC;QAEzD,MAAM,cAAc,GAAoB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;YACvD,MAAM,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YACrD,MAAM,SAAS,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAClC,MAAM,UAAU,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACxC,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAClE,CAAC;YAEF,OAAO;gBACL,IAAI,EAAE,CAAC,CAAC,KAAK;gBACb,QAAQ,EAAE,aAAa;gBACvB,WAAW,EAAE,UAAU,CAAC,MAAM;gBAC9B,OAAO,EAAE,UAAU;aACpB,CAAC;QACJ,CAAC,CAAC,CAAC;QAEH,IAAI,aAAa,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QAC9B,IAAI,WAAW,KAAK,SAAS,IAAI,WAAW,KAAK,IAAI,EAAE,CAAC;YACtD,aAAa;gBACX,OAAO,WAAW,KAAK,QAAQ;oBAC7B,CAAC,CAAC,MAAM,CAAC,WAAW,GAAG,CAAC,CAAC;oBACzB,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,MAAM,CAAC,WAAW,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;QACxF,CAAC;QAED,IAAI,OAAO,GAA+D,EAAE,CAAC;QAC7E,IAAI,QAAQ,GAAG,CAAC,CAAC;QAEjB,IAAI,aAAa,IAAI,aAAa,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACnD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAC9C,MAAM,UAAU,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACxC,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAClE,CAAC;YACF,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,aAAa,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YACtD,OAAO,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBAClC,IAAI;gBACJ,IAAI,EAAE,QAAQ;gBACd,cAAc,EAAE,CAAC;aAClB,CAAC,CAAC,CAAC;QACN,CAAC;QAED,OAAO;YACL,MAAM,EAAE,MAAM;YACd,QAAQ;YACR,WAAW,EAAE,OAAO,CAAC,MAAM;YAC3B,OAAO;YACP,WAAW,EAAE,MAAM,CAAC,MAAM;YAC1B,MAAM,EAAE,cAAc;SACvB,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
import type { TransformFunction } from "../core/types.js";
|
|
2
|
+
export type CastOnError = "fail" | "null" | "keep" | "skip-row";
|
|
3
|
+
export interface CastOptions {
|
|
4
|
+
onError?: CastOnError;
|
|
5
|
+
}
|
|
6
|
+
/**
|
|
7
|
+
* Casts a single value to the target type.
|
|
8
|
+
*/
|
|
9
|
+
export declare function castValue(value: unknown, targetType: string, onError?: CastOnError): {
|
|
10
|
+
value: unknown;
|
|
11
|
+
success: boolean;
|
|
12
|
+
};
|
|
13
|
+
/**
|
|
14
|
+
* Creates a transform that casts specified columns to target types.
|
|
15
|
+
*/
|
|
16
|
+
export declare function castColumns(typeSpecs: Record<string, string>, options?: CastOptions): TransformFunction;
|
|
17
|
+
/**
|
|
18
|
+
* Parses command line cast specs like ["age:integer", "revenue:number"].
|
|
19
|
+
*/
|
|
20
|
+
export declare function parseCastSpecs(specs: string[]): Record<string, string>;
|
|
21
|
+
//# sourceMappingURL=cast.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cast.d.ts","sourceRoot":"","sources":["../../src/transforms/cast.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAmB,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAE3E,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,MAAM,GAAG,MAAM,GAAG,UAAU,CAAC;AAEhE,MAAM,WAAW,WAAW;IAC1B,OAAO,CAAC,EAAE,WAAW,CAAC;CACvB;AAED;;GAEG;AACH,wBAAgB,SAAS,CACvB,KAAK,EAAE,OAAO,EACd,UAAU,EAAE,MAAM,EAClB,OAAO,GAAE,WAAoB,GAC5B;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,OAAO,EAAE,OAAO,CAAA;CAAE,CAmFtC;AAED;;GAEG;AACH,wBAAgB,WAAW,CACzB,SAAS,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,EACjC,OAAO,GAAE,WAAgB,GACxB,iBAAiB,CA0FnB;AAED;;GAEG;AACH,wBAAgB,cAAc,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAStE"}
|
|
@@ -0,0 +1,182 @@
|
|
|
1
|
+
import { InvalidArgumentError } from "../core/errors.js";
|
|
2
|
+
/**
|
|
3
|
+
* Casts a single value to the target type.
|
|
4
|
+
*/
|
|
5
|
+
export function castValue(value, targetType, onError = "null") {
|
|
6
|
+
if (value === null || value === undefined || value === "") {
|
|
7
|
+
return { value: null, success: true };
|
|
8
|
+
}
|
|
9
|
+
const str = String(value).trim();
|
|
10
|
+
const lowerType = targetType.toLowerCase();
|
|
11
|
+
try {
|
|
12
|
+
switch (lowerType) {
|
|
13
|
+
case "string":
|
|
14
|
+
return { value: String(value), success: true };
|
|
15
|
+
case "integer":
|
|
16
|
+
case "int": {
|
|
17
|
+
if (!/^-?\d+$/.test(str)) {
|
|
18
|
+
throw new Error(`Cannot cast "${str}" to integer`);
|
|
19
|
+
}
|
|
20
|
+
const parsedInt = Number.parseInt(str, 10);
|
|
21
|
+
if (Number.isNaN(parsedInt)) {
|
|
22
|
+
throw new Error(`Cannot cast "${str}" to integer`);
|
|
23
|
+
}
|
|
24
|
+
return { value: parsedInt, success: true };
|
|
25
|
+
}
|
|
26
|
+
case "number":
|
|
27
|
+
case "float":
|
|
28
|
+
case "double": {
|
|
29
|
+
const parsedNum = Number.parseFloat(str);
|
|
30
|
+
if (Number.isNaN(parsedNum)) {
|
|
31
|
+
throw new Error(`Cannot cast "${str}" to number`);
|
|
32
|
+
}
|
|
33
|
+
return { value: parsedNum, success: true };
|
|
34
|
+
}
|
|
35
|
+
case "boolean":
|
|
36
|
+
case "bool": {
|
|
37
|
+
const lower = str.toLowerCase();
|
|
38
|
+
if (["true", "1", "yes", "t", "y"].includes(lower)) {
|
|
39
|
+
return { value: true, success: true };
|
|
40
|
+
}
|
|
41
|
+
if (["false", "0", "no", "f", "n"].includes(lower)) {
|
|
42
|
+
return { value: false, success: true };
|
|
43
|
+
}
|
|
44
|
+
throw new Error(`Cannot cast "${str}" to boolean`);
|
|
45
|
+
}
|
|
46
|
+
case "date": {
|
|
47
|
+
const d = new Date(str);
|
|
48
|
+
if (Number.isNaN(d.getTime())) {
|
|
49
|
+
throw new Error(`Cannot cast "${str}" to date`);
|
|
50
|
+
}
|
|
51
|
+
return { value: d.toISOString().split("T")[0], success: true };
|
|
52
|
+
}
|
|
53
|
+
case "datetime": {
|
|
54
|
+
const d = new Date(str);
|
|
55
|
+
if (Number.isNaN(d.getTime())) {
|
|
56
|
+
throw new Error(`Cannot cast "${str}" to datetime`);
|
|
57
|
+
}
|
|
58
|
+
return { value: d.toISOString(), success: true };
|
|
59
|
+
}
|
|
60
|
+
case "json": {
|
|
61
|
+
const parsed = JSON.parse(str);
|
|
62
|
+
return { value: parsed, success: true };
|
|
63
|
+
}
|
|
64
|
+
default:
|
|
65
|
+
throw new Error(`Unknown target type "${targetType}"`);
|
|
66
|
+
}
|
|
67
|
+
}
|
|
68
|
+
catch (err) {
|
|
69
|
+
if (onError === "fail") {
|
|
70
|
+
throw new InvalidArgumentError(`Failed to cast value "${str}" to type "${targetType}": ${err.message}`);
|
|
71
|
+
}
|
|
72
|
+
if (onError === "keep") {
|
|
73
|
+
return { value, success: false };
|
|
74
|
+
}
|
|
75
|
+
// "null" or "skip-row"
|
|
76
|
+
return { value: null, success: false };
|
|
77
|
+
}
|
|
78
|
+
}
|
|
79
|
+
/**
|
|
80
|
+
* Creates a transform that casts specified columns to target types.
|
|
81
|
+
*/
|
|
82
|
+
export function castColumns(typeSpecs, options = {}) {
|
|
83
|
+
const onError = options.onError || "null";
|
|
84
|
+
const compiled = Object.entries(typeSpecs).map(([col, targetType]) => {
|
|
85
|
+
const lower = targetType.toLowerCase();
|
|
86
|
+
// Fast-path specialized caster functions
|
|
87
|
+
if (lower === "number" || lower === "float" || lower === "double") {
|
|
88
|
+
return {
|
|
89
|
+
col,
|
|
90
|
+
cast: (val) => {
|
|
91
|
+
if (val === null || val === undefined || val === "")
|
|
92
|
+
return { value: null, success: true };
|
|
93
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
94
|
+
if (!Number.isNaN(num))
|
|
95
|
+
return { value: num, success: true };
|
|
96
|
+
if (onError === "fail")
|
|
97
|
+
throw new InvalidArgumentError(`Cannot cast "${String(val)}" to number`);
|
|
98
|
+
if (onError === "keep")
|
|
99
|
+
return { value: val, success: false };
|
|
100
|
+
return { value: null, success: false };
|
|
101
|
+
},
|
|
102
|
+
};
|
|
103
|
+
}
|
|
104
|
+
if (lower === "integer" || lower === "int") {
|
|
105
|
+
return {
|
|
106
|
+
col,
|
|
107
|
+
cast: (val) => {
|
|
108
|
+
if (val === null || val === undefined || val === "")
|
|
109
|
+
return { value: null, success: true };
|
|
110
|
+
const str = String(val).trim();
|
|
111
|
+
if (!/^-?\d+$/.test(str)) {
|
|
112
|
+
if (onError === "fail")
|
|
113
|
+
throw new InvalidArgumentError(`Cannot cast "${str}" to integer`);
|
|
114
|
+
if (onError === "keep")
|
|
115
|
+
return { value: val, success: false };
|
|
116
|
+
return { value: null, success: false };
|
|
117
|
+
}
|
|
118
|
+
const num = Number.parseInt(str, 10);
|
|
119
|
+
return { value: num, success: true };
|
|
120
|
+
},
|
|
121
|
+
};
|
|
122
|
+
}
|
|
123
|
+
if (lower === "string") {
|
|
124
|
+
return {
|
|
125
|
+
col,
|
|
126
|
+
cast: (val) => ({ value: val !== null && val !== undefined ? String(val) : "", success: true }),
|
|
127
|
+
};
|
|
128
|
+
}
|
|
129
|
+
return {
|
|
130
|
+
col,
|
|
131
|
+
cast: (val) => castValue(val, targetType, onError),
|
|
132
|
+
};
|
|
133
|
+
});
|
|
134
|
+
const compiledLen = compiled.length;
|
|
135
|
+
return function (stream) {
|
|
136
|
+
return (async function* () {
|
|
137
|
+
for await (const batch of stream) {
|
|
138
|
+
const rows = batch.rows;
|
|
139
|
+
const rowCount = rows.length;
|
|
140
|
+
const transformedRows = [];
|
|
141
|
+
for (let i = 0; i < rowCount; i++) {
|
|
142
|
+
const row = rows[i];
|
|
143
|
+
let skipRow = false;
|
|
144
|
+
const newRow = { ...row };
|
|
145
|
+
for (let j = 0; j < compiledLen; j++) {
|
|
146
|
+
const entry = compiled[j];
|
|
147
|
+
const col = entry.col;
|
|
148
|
+
if (col in newRow) {
|
|
149
|
+
const res = entry.cast(newRow[col]);
|
|
150
|
+
if (!res.success && onError === "skip-row") {
|
|
151
|
+
skipRow = true;
|
|
152
|
+
break;
|
|
153
|
+
}
|
|
154
|
+
newRow[col] = res.value;
|
|
155
|
+
}
|
|
156
|
+
}
|
|
157
|
+
if (!skipRow) {
|
|
158
|
+
transformedRows.push(newRow);
|
|
159
|
+
}
|
|
160
|
+
}
|
|
161
|
+
yield {
|
|
162
|
+
rows: transformedRows,
|
|
163
|
+
offset: batch.offset,
|
|
164
|
+
};
|
|
165
|
+
}
|
|
166
|
+
})();
|
|
167
|
+
};
|
|
168
|
+
}
|
|
169
|
+
/**
|
|
170
|
+
* Parses command line cast specs like ["age:integer", "revenue:number"].
|
|
171
|
+
*/
|
|
172
|
+
export function parseCastSpecs(specs) {
|
|
173
|
+
const mapping = {};
|
|
174
|
+
for (const spec of specs) {
|
|
175
|
+
const parts = spec.split(":");
|
|
176
|
+
if (parts.length === 2 && parts[0] && parts[1]) {
|
|
177
|
+
mapping[parts[0].trim()] = parts[1].trim();
|
|
178
|
+
}
|
|
179
|
+
}
|
|
180
|
+
return mapping;
|
|
181
|
+
}
|
|
182
|
+
//# sourceMappingURL=cast.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cast.js","sourceRoot":"","sources":["../../src/transforms/cast.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,mBAAmB,CAAC;AASzD;;GAEG;AACH,MAAM,UAAU,SAAS,CACvB,KAAc,EACd,UAAkB,EAClB,UAAuB,MAAM;IAE7B,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,EAAE,EAAE,CAAC;QAC1D,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;IACxC,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;IACjC,MAAM,SAAS,GAAG,UAAU,CAAC,WAAW,EAAE,CAAC;IAE3C,IAAI,CAAC;QACH,QAAQ,SAAS,EAAE,CAAC;YAClB,KAAK,QAAQ;gBACX,OAAO,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;YAEjD,KAAK,SAAS,CAAC;YACf,KAAK,KAAK,CAAC,CAAC,CAAC;gBACX,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;oBACzB,MAAM,IAAI,KAAK,CAAC,gBAAgB,GAAG,cAAc,CAAC,CAAC;gBACrD,CAAC;gBACD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;gBAC3C,IAAI,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;oBAC5B,MAAM,IAAI,KAAK,CAAC,gBAAgB,GAAG,cAAc,CAAC,CAAC;gBACrD,CAAC;gBACD,OAAO,EAAE,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;YAC7C,CAAC;YAED,KAAK,QAAQ,CAAC;YACd,KAAK,OAAO,CAAC;YACb,KAAK,QAAQ,CAAC,CAAC,CAAC;gBACd,MAAM,SAAS,GAAG,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;gBACzC,IAAI,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;oBAC5B,MAAM,IAAI,KAAK,CAAC,gBAAgB,GAAG,aAAa,CAAC,CAAC;gBACpD,CAAC;gBACD,OAAO,EAAE,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;YAC7C,CAAC;YAED,KAAK,SAAS,CAAC;YACf,KAAK,MAAM,CAAC,CAAC,CAAC;gBACZ,MAAM,KAAK,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC;gBAChC,IAAI,CAAC,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;oBACnD,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;gBACxC,CAAC;gBACD,IAAI,CAAC,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;oBACnD,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;gBACzC,CAAC;gBACD,MAAM,IAAI,KAAK,CAAC,gBAAgB,GAAG,cAAc,CAAC,CAAC;YACrD,CAAC;YAED,KAAK,MAAM,CAAC,CAAC,CAAC;gBACZ,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;gBACxB,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,EAAE,CAAC;oBAC9B,MAAM,IAAI,KAAK,CAAC,gBAAgB,GAAG,WAAW,CAAC,CAAC;gBAClD,CAAC;gBACD,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;YACjE,CAAC;YAED,KAAK,UAAU,CAAC,CAAC,CAAC;gBAChB,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;gBACxB,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,EAAE,CAAC;oBAC9B,MAAM,IAAI,KAAK,CAAC,gBAAgB,GAAG,eAAe,CAAC,CAAC;gBACtD,CAAC;gBACD,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC,WAAW,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;YACnD,CAAC;YAED,KAAK,MAAM,CAAC,CAAC,CAAC;gBACZ,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;gBAC/B,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;YAC1C,CAAC;YAED;gBACE,MAAM,IAAI,KAAK,CAAC,wBAAwB,UAAU,GAAG,CAAC,CAAC;QAC3D,CAAC;IACH,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,MAAM,IAAI,oBAAoB,CAC5B,yBAAyB,GAAG,cAAc,UAAU,MAAO,GAAa,CAAC,OAAO,EAAE,CACnF,CAAC;QACJ,CAAC;QACD,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;QACnC,CAAC;QACD,uBAAuB;QACvB,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;IACzC,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,WAAW,CACzB,SAAiC,EACjC,UAAuB,EAAE;IAEzB,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,MAAM,CAAC;IAE1C,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,UAAU,CAAC,EAAE,EAAE;QACnE,MAAM,KAAK,GAAG,UAAU,CAAC,WAAW,EAAE,CAAC;QAEvC,yCAAyC;QACzC,IAAI,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,OAAO,IAAI,KAAK,KAAK,QAAQ,EAAE,CAAC;YAClE,OAAO;gBACL,GAAG;gBACH,IAAI,EAAE,CAAC,GAAY,EAAE,EAAE;oBACrB,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;wBAAE,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;oBAC3F,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;oBACxD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC;wBAAE,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;oBAC7D,IAAI,OAAO,KAAK,MAAM;wBAAE,MAAM,IAAI,oBAAoB,CAAC,gBAAgB,MAAM,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;oBACjG,IAAI,OAAO,KAAK,MAAM;wBAAE,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;oBAC9D,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;gBACzC,CAAC;aACF,CAAC;QACJ,CAAC;QAED,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,KAAK,EAAE,CAAC;YAC3C,OAAO;gBACL,GAAG;gBACH,IAAI,EAAE,CAAC,GAAY,EAAE,EAAE;oBACrB,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;wBAAE,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;oBAC3F,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;oBAC/B,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;wBACzB,IAAI,OAAO,KAAK,MAAM;4BAAE,MAAM,IAAI,oBAAoB,CAAC,gBAAgB,GAAG,cAAc,CAAC,CAAC;wBAC1F,IAAI,OAAO,KAAK,MAAM;4BAAE,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;wBAC9D,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;oBACzC,CAAC;oBACD,MAAM,GAAG,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;oBACrC,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;gBACvC,CAAC;aACF,CAAC;QACJ,CAAC;QAED,IAAI,KAAK,KAAK,QAAQ,EAAE,CAAC;YACvB,OAAO;gBACL,GAAG;gBACH,IAAI,EAAE,CAAC,GAAY,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;aACzG,CAAC;QACJ,CAAC;QAED,OAAO;YACL,GAAG;YACH,IAAI,EAAE,CAAC,GAAY,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,EAAE,UAAU,EAAE,OAAO,CAAC;SAC5D,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,CAAC;IAEpC,OAAO,UAAU,MAAkB;QACjC,OAAO,CAAC,KAAK,SAAS,CAAC;YACrB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBACjC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;gBACxB,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAAC;gBAC7B,MAAM,eAAe,GAAU,EAAE,CAAC;gBAElC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;oBAClC,MAAM,GAAG,GAAG,IAAI,CAAC,CAAC,CAAE,CAAC;oBACrB,IAAI,OAAO,GAAG,KAAK,CAAC;oBACpB,MAAM,MAAM,GAAQ,EAAE,GAAG,GAAG,EAAE,CAAC;oBAE/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC;wBACrC,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC;wBAC3B,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;wBACtB,IAAI,GAAG,IAAI,MAAM,EAAE,CAAC;4BAClB,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;4BACpC,IAAI,CAAC,GAAG,CAAC,OAAO,IAAI,OAAO,KAAK,UAAU,EAAE,CAAC;gCAC3C,OAAO,GAAG,IAAI,CAAC;gCACf,MAAM;4BACR,CAAC;4BACD,MAAM,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC;wBAC1B,CAAC;oBACH,CAAC;oBAED,IAAI,CAAC,OAAO,EAAE,CAAC;wBACb,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;oBAC/B,CAAC;gBACH,CAAC;gBAED,MAAM;oBACJ,IAAI,EAAE,eAAe;oBACrB,MAAM,EAAE,KAAK,CAAC,MAAM;iBACrB,CAAC;YACJ,CAAC;QACH,CAAC,CAAC,EAAE,CAAC;IACP,CAAC,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAAC,KAAe;IAC5C,MAAM,OAAO,GAA2B,EAAE,CAAC;IAC3C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC9B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;YAC/C,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC7C,CAAC;IACH,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC"}
|
|
@@ -0,0 +1,67 @@
|
|
|
1
|
+
import type { Row } from "../core/types.js";
|
|
2
|
+
type TokenType = "NUMBER" | "STRING" | "BOOLEAN" | "NULL" | "IDENTIFIER" | "OP" | "PIPE" | "LPAREN" | "RPAREN" | "COMMA" | "EOF";
|
|
3
|
+
interface Token {
|
|
4
|
+
type: TokenType;
|
|
5
|
+
value: string | number | boolean | null;
|
|
6
|
+
pos: number;
|
|
7
|
+
}
|
|
8
|
+
export declare class Lexer {
|
|
9
|
+
private input;
|
|
10
|
+
private cursor;
|
|
11
|
+
constructor(input: string);
|
|
12
|
+
private isWhitespace;
|
|
13
|
+
private isDigit;
|
|
14
|
+
private isAlpha;
|
|
15
|
+
tokenize(): Token[];
|
|
16
|
+
}
|
|
17
|
+
export type ASTNode = {
|
|
18
|
+
type: "Literal";
|
|
19
|
+
value: unknown;
|
|
20
|
+
} | {
|
|
21
|
+
type: "Identifier";
|
|
22
|
+
name: string;
|
|
23
|
+
} | {
|
|
24
|
+
type: "UnaryOp";
|
|
25
|
+
operator: string;
|
|
26
|
+
argument: ASTNode;
|
|
27
|
+
} | {
|
|
28
|
+
type: "BinaryOp";
|
|
29
|
+
operator: string;
|
|
30
|
+
left: ASTNode;
|
|
31
|
+
right: ASTNode;
|
|
32
|
+
} | {
|
|
33
|
+
type: "CallExpr";
|
|
34
|
+
callee: string;
|
|
35
|
+
args: ASTNode[];
|
|
36
|
+
};
|
|
37
|
+
export declare class Parser {
|
|
38
|
+
private tokens;
|
|
39
|
+
private current;
|
|
40
|
+
constructor(tokens: Token[]);
|
|
41
|
+
private peek;
|
|
42
|
+
private advance;
|
|
43
|
+
private match;
|
|
44
|
+
parse(): ASTNode;
|
|
45
|
+
private parseLogicalOr;
|
|
46
|
+
private parseLogicalAnd;
|
|
47
|
+
private parseEquality;
|
|
48
|
+
private parseRelational;
|
|
49
|
+
private parseAdditive;
|
|
50
|
+
private parseMultiplicative;
|
|
51
|
+
/**
|
|
52
|
+
* Parses pipe expressions e.g. email | trim | lower | startsWith("admin")
|
|
53
|
+
*/
|
|
54
|
+
private parsePipe;
|
|
55
|
+
private parseUnary;
|
|
56
|
+
private parsePrimary;
|
|
57
|
+
}
|
|
58
|
+
/**
|
|
59
|
+
* Compiles a filter expression string into an optimized predicate function.
|
|
60
|
+
*/
|
|
61
|
+
export declare function compileExpression(expressionStr: string): (row: Row) => boolean;
|
|
62
|
+
/**
|
|
63
|
+
* Compiles an expression string into an optimized value evaluator returning computed result (number, string, etc.).
|
|
64
|
+
*/
|
|
65
|
+
export declare function compileValueExpression(expressionStr: string): (row: Row) => unknown;
|
|
66
|
+
export {};
|
|
67
|
+
//# sourceMappingURL=expression.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"expression.d.ts","sourceRoot":"","sources":["../../src/transforms/expression.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,GAAG,EAAE,MAAM,kBAAkB,CAAC;AAI5C,KAAK,SAAS,GACV,QAAQ,GACR,QAAQ,GACR,SAAS,GACT,MAAM,GACN,YAAY,GACZ,IAAI,GACJ,MAAM,GACN,QAAQ,GACR,QAAQ,GACR,OAAO,GACP,KAAK,CAAC;AAEV,UAAU,KAAK;IACb,IAAI,EAAE,SAAS,CAAC;IAChB,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,OAAO,GAAG,IAAI,CAAC;IACxC,GAAG,EAAE,MAAM,CAAC;CACb;AAGD,qBAAa,KAAK;IAChB,OAAO,CAAC,KAAK,CAAS;IACtB,OAAO,CAAC,MAAM,CAAK;gBAEP,KAAK,EAAE,MAAM;IAIzB,OAAO,CAAC,YAAY;IAIpB,OAAO,CAAC,OAAO;IAIf,OAAO,CAAC,OAAO;IAUf,QAAQ,IAAI,KAAK,EAAE;CAyJpB;AAGD,MAAM,MAAM,OAAO,GACf;IAAE,IAAI,EAAE,SAAS,CAAC;IAAC,KAAK,EAAE,OAAO,CAAA;CAAE,GACnC;IAAE,IAAI,EAAE,YAAY,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GACpC;IAAE,IAAI,EAAE,SAAS,CAAC;IAAC,QAAQ,EAAE,MAAM,CAAC;IAAC,QAAQ,EAAE,OAAO,CAAA;CAAE,GACxD;IAAE,IAAI,EAAE,UAAU,CAAC;IAAC,QAAQ,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,OAAO,CAAA;CAAE,GACrE;IAAE,IAAI,EAAE,UAAU,CAAC;IAAC,MAAM,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,OAAO,EAAE,CAAA;CAAE,CAAC;AAE1D,qBAAa,MAAM;IACjB,OAAO,CAAC,MAAM,CAAU;IACxB,OAAO,CAAC,OAAO,CAAK;gBAER,MAAM,EAAE,KAAK,EAAE;IAI3B,OAAO,CAAC,IAAI;IAIZ,OAAO,CAAC,OAAO;IAMf,OAAO,CAAC,KAAK;IASb,KAAK,IAAI,OAAO;IAUhB,OAAO,CAAC,cAAc;IAWtB,OAAO,CAAC,eAAe;IAWvB,OAAO,CAAC,aAAa;IAYrB,OAAO,CAAC,eAAe;IAiBvB,OAAO,CAAC,aAAa;IAYrB,OAAO,CAAC,mBAAmB;IAgB3B;;OAEG;IACH,OAAO,CAAC,SAAS;IAwCjB,OAAO,CAAC,UAAU;IAUlB,OAAO,CAAC,YAAY;CAqDrB;AAmoBD;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,aAAa,EAAE,MAAM,GAAG,CAAC,GAAG,EAAE,GAAG,KAAK,OAAO,CAc9E;AAED;;GAEG;AACH,wBAAgB,sBAAsB,CAAC,aAAa,EAAE,MAAM,GAAG,CAAC,GAAG,EAAE,GAAG,KAAK,OAAO,CAWnF"}
|