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 +1,102 @@
|
|
|
1
|
+
import type { Aggregator, Row } from "../core/types.js";
|
|
2
|
+
/**
|
|
3
|
+
* HyperLogLog (HLL) distinct count estimator with bounded O(1) memory.
|
|
4
|
+
* Uses m = 1024 registers (p = 10), standard error ~3.25%.
|
|
5
|
+
* Hardware Math.clz32 for instantaneous zero counting.
|
|
6
|
+
*/
|
|
7
|
+
export declare class HyperLogLog {
|
|
8
|
+
private p;
|
|
9
|
+
private m;
|
|
10
|
+
private registers;
|
|
11
|
+
private alphaMM;
|
|
12
|
+
constructor(p?: number);
|
|
13
|
+
add(value: unknown): void;
|
|
14
|
+
merge(other: HyperLogLog): void;
|
|
15
|
+
count(): number;
|
|
16
|
+
}
|
|
17
|
+
export interface NumericColumnStats {
|
|
18
|
+
count: number;
|
|
19
|
+
nullCount: number;
|
|
20
|
+
min: number | null;
|
|
21
|
+
max: number | null;
|
|
22
|
+
sum: number;
|
|
23
|
+
mean: number | null;
|
|
24
|
+
variance: number | null;
|
|
25
|
+
stddev: number | null;
|
|
26
|
+
approxDistinct: number;
|
|
27
|
+
}
|
|
28
|
+
export interface StringColumnStats {
|
|
29
|
+
count: number;
|
|
30
|
+
nullCount: number;
|
|
31
|
+
emptyCount: number;
|
|
32
|
+
minLength: number | null;
|
|
33
|
+
maxLength: number | null;
|
|
34
|
+
avgLength: number | null;
|
|
35
|
+
emptyCountRatio?: number;
|
|
36
|
+
approxDistinct: number;
|
|
37
|
+
topValues: Array<{
|
|
38
|
+
value: string;
|
|
39
|
+
count: number;
|
|
40
|
+
}>;
|
|
41
|
+
}
|
|
42
|
+
export interface ColumnStatsResult {
|
|
43
|
+
column: string;
|
|
44
|
+
type: "numeric" | "string";
|
|
45
|
+
numeric?: NumericColumnStats;
|
|
46
|
+
string?: StringColumnStats;
|
|
47
|
+
}
|
|
48
|
+
export interface DatasetStatsResult {
|
|
49
|
+
totalRows: number;
|
|
50
|
+
columns: Record<string, ColumnStatsResult>;
|
|
51
|
+
}
|
|
52
|
+
/**
|
|
53
|
+
* Online Welford statistics collector for a numeric column.
|
|
54
|
+
*/
|
|
55
|
+
export declare class NumericStatsCollector {
|
|
56
|
+
count: number;
|
|
57
|
+
nullCount: number;
|
|
58
|
+
min: number | null;
|
|
59
|
+
max: number | null;
|
|
60
|
+
sum: number;
|
|
61
|
+
mean: number;
|
|
62
|
+
M2: number;
|
|
63
|
+
hll: HyperLogLog;
|
|
64
|
+
add(val: unknown): void;
|
|
65
|
+
merge(other: NumericStatsCollector): void;
|
|
66
|
+
result(): NumericColumnStats;
|
|
67
|
+
}
|
|
68
|
+
/**
|
|
69
|
+
* Online statistics collector for a string column.
|
|
70
|
+
*/
|
|
71
|
+
export declare class StringStatsCollector {
|
|
72
|
+
count: number;
|
|
73
|
+
nullCount: number;
|
|
74
|
+
emptyCount: number;
|
|
75
|
+
minLength: number | null;
|
|
76
|
+
maxLength: number | null;
|
|
77
|
+
totalLength: number;
|
|
78
|
+
hll: HyperLogLog;
|
|
79
|
+
private frequencyMap;
|
|
80
|
+
private maxTrackedValues;
|
|
81
|
+
add(val: unknown): void;
|
|
82
|
+
result(): StringColumnStats;
|
|
83
|
+
}
|
|
84
|
+
/**
|
|
85
|
+
* Dataset-wide Aggregator collecting statistics across all columns in stream.
|
|
86
|
+
*/
|
|
87
|
+
export declare class DatasetStatsAggregator implements Aggregator<DatasetStatsResult> {
|
|
88
|
+
private totalRows;
|
|
89
|
+
private targetColumn?;
|
|
90
|
+
private numericCollectors;
|
|
91
|
+
private stringCollectors;
|
|
92
|
+
private columnTypeHints;
|
|
93
|
+
private handlers;
|
|
94
|
+
private initialized;
|
|
95
|
+
constructor(options?: {
|
|
96
|
+
column?: string;
|
|
97
|
+
});
|
|
98
|
+
private initHandler;
|
|
99
|
+
add(row: Row): void;
|
|
100
|
+
result(): DatasetStatsResult;
|
|
101
|
+
}
|
|
102
|
+
//# sourceMappingURL=stats.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"stats.d.ts","sourceRoot":"","sources":["../../src/analytics/stats.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,kBAAkB,CAAC;AAsBxD;;;;GAIG;AACH,qBAAa,WAAW;IACtB,OAAO,CAAC,CAAC,CAAS;IAClB,OAAO,CAAC,CAAC,CAAS;IAClB,OAAO,CAAC,SAAS,CAAa;IAC9B,OAAO,CAAC,OAAO,CAAS;gBAEZ,CAAC,SAAK;IAYlB,GAAG,CAAC,KAAK,EAAE,OAAO,GAAG,IAAI;IAuBzB,KAAK,CAAC,KAAK,EAAE,WAAW,GAAG,IAAI;IAQ/B,KAAK,IAAI,MAAM;CAmBhB;AAED,MAAM,WAAW,kBAAkB;IACjC,KAAK,EAAE,MAAM,CAAC;IACd,SAAS,EAAE,MAAM,CAAC;IAClB,GAAG,EAAE,MAAM,GAAG,IAAI,CAAC;IACnB,GAAG,EAAE,MAAM,GAAG,IAAI,CAAC;IACnB,GAAG,EAAE,MAAM,CAAC;IACZ,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,cAAc,EAAE,MAAM,CAAC;CACxB;AAED,MAAM,WAAW,iBAAiB;IAChC,KAAK,EAAE,MAAM,CAAC;IACd,SAAS,EAAE,MAAM,CAAC;IAClB,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;IACzB,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;IACzB,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;IACzB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,cAAc,EAAE,MAAM,CAAC;IACvB,SAAS,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CACpD;AAED,MAAM,WAAW,iBAAiB;IAChC,MAAM,EAAE,MAAM,CAAC;IACf,IAAI,EAAE,SAAS,GAAG,QAAQ,CAAC;IAC3B,OAAO,CAAC,EAAE,kBAAkB,CAAC;IAC7B,MAAM,CAAC,EAAE,iBAAiB,CAAC;CAC5B;AAED,MAAM,WAAW,kBAAkB;IACjC,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAC;CAC5C;AAED;;GAEG;AACH,qBAAa,qBAAqB;IACzB,KAAK,SAAK;IACV,SAAS,SAAK;IACd,GAAG,EAAE,MAAM,GAAG,IAAI,CAAQ;IAC1B,GAAG,EAAE,MAAM,GAAG,IAAI,CAAQ;IAC1B,GAAG,SAAK;IACR,IAAI,SAAK;IACT,EAAE,SAAK;IACP,GAAG,cAAqB;IAE/B,GAAG,CAAC,GAAG,EAAE,OAAO,GAAG,IAAI;IA0BvB,KAAK,CAAC,KAAK,EAAE,qBAAqB,GAAG,IAAI;IA+BzC,MAAM,IAAI,kBAAkB;CAgB7B;AAED;;GAEG;AACH,qBAAa,oBAAoB;IACxB,KAAK,SAAK;IACV,SAAS,SAAK;IACd,UAAU,SAAK;IACf,SAAS,EAAE,MAAM,GAAG,IAAI,CAAQ;IAChC,SAAS,EAAE,MAAM,GAAG,IAAI,CAAQ;IAChC,WAAW,SAAK;IAChB,GAAG,cAAqB;IAC/B,OAAO,CAAC,YAAY,CAA6B;IACjD,OAAO,CAAC,gBAAgB,CAAQ;IAEhC,GAAG,CAAC,GAAG,EAAE,OAAO,GAAG,IAAI;IAwBvB,MAAM,IAAI,iBAAiB;CAiB5B;AASD;;GAEG;AACH,qBAAa,sBAAuB,YAAW,UAAU,CAAC,kBAAkB,CAAC;IAC3E,OAAO,CAAC,SAAS,CAAK;IACtB,OAAO,CAAC,YAAY,CAAC,CAAS;IAC9B,OAAO,CAAC,iBAAiB,CAA4C;IACrE,OAAO,CAAC,gBAAgB,CAA2C;IACnE,OAAO,CAAC,eAAe,CAAmE;IAC1F,OAAO,CAAC,QAAQ,CAAuB;IACvC,OAAO,CAAC,WAAW,CAAS;gBAEhB,OAAO,CAAC,EAAE;QAAE,MAAM,CAAC,EAAE,MAAM,CAAA;KAAE;IAIzC,OAAO,CAAC,WAAW;IAiBnB,GAAG,CAAC,GAAG,EAAE,GAAG,GAAG,IAAI;IAoCnB,MAAM,IAAI,kBAAkB;CA4B7B"}
|
|
@@ -0,0 +1,307 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Fast 32-bit FNV-1a hash function with 32-bit avalanche mixing.
|
|
3
|
+
* Zero BigInt allocations, runs at native CPU register speed.
|
|
4
|
+
*/
|
|
5
|
+
function hash32(str) {
|
|
6
|
+
let h = 0x811c9dc5;
|
|
7
|
+
const len = str.length;
|
|
8
|
+
for (let i = 0; i < len; i++) {
|
|
9
|
+
h = (h ^ str.charCodeAt(i)) >>> 0;
|
|
10
|
+
h = Math.imul(h, 0x01000193) >>> 0;
|
|
11
|
+
}
|
|
12
|
+
// Avalanche mixing
|
|
13
|
+
h ^= h >>> 16;
|
|
14
|
+
h = Math.imul(h, 0x85ebca6b) >>> 0;
|
|
15
|
+
h ^= h >>> 13;
|
|
16
|
+
h = Math.imul(h, 0xc2b2ae35) >>> 0;
|
|
17
|
+
h ^= h >>> 16;
|
|
18
|
+
return h >>> 0;
|
|
19
|
+
}
|
|
20
|
+
/**
|
|
21
|
+
* HyperLogLog (HLL) distinct count estimator with bounded O(1) memory.
|
|
22
|
+
* Uses m = 1024 registers (p = 10), standard error ~3.25%.
|
|
23
|
+
* Hardware Math.clz32 for instantaneous zero counting.
|
|
24
|
+
*/
|
|
25
|
+
export class HyperLogLog {
|
|
26
|
+
p;
|
|
27
|
+
m;
|
|
28
|
+
registers;
|
|
29
|
+
alphaMM;
|
|
30
|
+
constructor(p = 10) {
|
|
31
|
+
this.p = p;
|
|
32
|
+
this.m = 1 << p;
|
|
33
|
+
this.registers = new Uint8Array(this.m);
|
|
34
|
+
// Alpha constant calculation
|
|
35
|
+
if (this.m === 16)
|
|
36
|
+
this.alphaMM = 0.673 * this.m * this.m;
|
|
37
|
+
else if (this.m === 32)
|
|
38
|
+
this.alphaMM = 0.697 * this.m * this.m;
|
|
39
|
+
else if (this.m === 64)
|
|
40
|
+
this.alphaMM = 0.709 * this.m * this.m;
|
|
41
|
+
else
|
|
42
|
+
this.alphaMM = (0.7213 / (1 + 1.079 / this.m)) * this.m * this.m;
|
|
43
|
+
}
|
|
44
|
+
add(value) {
|
|
45
|
+
if (value === null || value === undefined)
|
|
46
|
+
return;
|
|
47
|
+
const str = String(value);
|
|
48
|
+
const hash = hash32(str);
|
|
49
|
+
// Low p bits for register index
|
|
50
|
+
const index = hash & (this.m - 1);
|
|
51
|
+
// Remaining (32 - p) bits for leading zeros
|
|
52
|
+
const w = hash >>> this.p;
|
|
53
|
+
// Single-cycle CPU instruction Math.clz32 for leading zero count
|
|
54
|
+
let leadingZeros;
|
|
55
|
+
if (w === 0) {
|
|
56
|
+
leadingZeros = 32 - this.p + 1;
|
|
57
|
+
}
|
|
58
|
+
else {
|
|
59
|
+
leadingZeros = Math.clz32(w) - this.p + 1;
|
|
60
|
+
}
|
|
61
|
+
if (leadingZeros > this.registers[index]) {
|
|
62
|
+
this.registers[index] = leadingZeros;
|
|
63
|
+
}
|
|
64
|
+
}
|
|
65
|
+
merge(other) {
|
|
66
|
+
for (let i = 0; i < this.m; i++) {
|
|
67
|
+
if (other.registers[i] > this.registers[i]) {
|
|
68
|
+
this.registers[i] = other.registers[i];
|
|
69
|
+
}
|
|
70
|
+
}
|
|
71
|
+
}
|
|
72
|
+
count() {
|
|
73
|
+
let sum = 0;
|
|
74
|
+
let zeros = 0;
|
|
75
|
+
for (let i = 0; i < this.m; i++) {
|
|
76
|
+
const val = this.registers[i];
|
|
77
|
+
sum += 2 ** -val;
|
|
78
|
+
if (val === 0)
|
|
79
|
+
zeros++;
|
|
80
|
+
}
|
|
81
|
+
let estimate = this.alphaMM / sum;
|
|
82
|
+
// Small range correction (Linear Counting)
|
|
83
|
+
if (estimate <= 2.5 * this.m && zeros > 0) {
|
|
84
|
+
estimate = this.m * Math.log(this.m / zeros);
|
|
85
|
+
}
|
|
86
|
+
return Math.round(estimate);
|
|
87
|
+
}
|
|
88
|
+
}
|
|
89
|
+
/**
|
|
90
|
+
* Online Welford statistics collector for a numeric column.
|
|
91
|
+
*/
|
|
92
|
+
export class NumericStatsCollector {
|
|
93
|
+
count = 0;
|
|
94
|
+
nullCount = 0;
|
|
95
|
+
min = null;
|
|
96
|
+
max = null;
|
|
97
|
+
sum = 0;
|
|
98
|
+
mean = 0;
|
|
99
|
+
M2 = 0; // sum of squared differences from the mean
|
|
100
|
+
hll = new HyperLogLog();
|
|
101
|
+
add(val) {
|
|
102
|
+
if (val === null || val === undefined || val === "") {
|
|
103
|
+
this.nullCount++;
|
|
104
|
+
return;
|
|
105
|
+
}
|
|
106
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
107
|
+
if (Number.isNaN(num)) {
|
|
108
|
+
this.nullCount++;
|
|
109
|
+
return;
|
|
110
|
+
}
|
|
111
|
+
this.count++;
|
|
112
|
+
this.sum += num;
|
|
113
|
+
this.hll.add(num);
|
|
114
|
+
if (this.min === null || num < this.min)
|
|
115
|
+
this.min = num;
|
|
116
|
+
if (this.max === null || num > this.max)
|
|
117
|
+
this.max = num;
|
|
118
|
+
// Welford's online algorithm
|
|
119
|
+
const delta = num - this.mean;
|
|
120
|
+
this.mean += delta / this.count;
|
|
121
|
+
const delta2 = num - this.mean;
|
|
122
|
+
this.M2 += delta * delta2;
|
|
123
|
+
}
|
|
124
|
+
merge(other) {
|
|
125
|
+
if (other.count === 0) {
|
|
126
|
+
this.nullCount += other.nullCount;
|
|
127
|
+
return;
|
|
128
|
+
}
|
|
129
|
+
if (this.count === 0) {
|
|
130
|
+
this.count = other.count;
|
|
131
|
+
this.nullCount = other.nullCount;
|
|
132
|
+
this.min = other.min;
|
|
133
|
+
this.max = other.max;
|
|
134
|
+
this.sum = other.sum;
|
|
135
|
+
this.mean = other.mean;
|
|
136
|
+
this.M2 = other.M2;
|
|
137
|
+
this.hll.merge(other.hll);
|
|
138
|
+
return;
|
|
139
|
+
}
|
|
140
|
+
const totalCount = this.count + other.count;
|
|
141
|
+
const delta = other.mean - this.mean;
|
|
142
|
+
this.mean = (this.count * this.mean + other.count * other.mean) / totalCount;
|
|
143
|
+
this.M2 = this.M2 + other.M2 + (delta * delta * this.count * other.count) / totalCount;
|
|
144
|
+
this.count = totalCount;
|
|
145
|
+
this.nullCount += other.nullCount;
|
|
146
|
+
this.sum += other.sum;
|
|
147
|
+
if (this.min === null || (other.min !== null && other.min < this.min))
|
|
148
|
+
this.min = other.min;
|
|
149
|
+
if (this.max === null || (other.max !== null && other.max > this.max))
|
|
150
|
+
this.max = other.max;
|
|
151
|
+
this.hll.merge(other.hll);
|
|
152
|
+
}
|
|
153
|
+
result() {
|
|
154
|
+
const variance = this.count > 1 ? this.M2 / (this.count - 1) : this.count === 1 ? 0 : null;
|
|
155
|
+
const stddev = variance !== null ? Math.sqrt(variance) : null;
|
|
156
|
+
return {
|
|
157
|
+
count: this.count,
|
|
158
|
+
nullCount: this.nullCount,
|
|
159
|
+
min: this.min,
|
|
160
|
+
max: this.max,
|
|
161
|
+
sum: this.sum,
|
|
162
|
+
mean: this.count > 0 ? this.mean : null,
|
|
163
|
+
variance,
|
|
164
|
+
stddev,
|
|
165
|
+
approxDistinct: this.hll.count(),
|
|
166
|
+
};
|
|
167
|
+
}
|
|
168
|
+
}
|
|
169
|
+
/**
|
|
170
|
+
* Online statistics collector for a string column.
|
|
171
|
+
*/
|
|
172
|
+
export class StringStatsCollector {
|
|
173
|
+
count = 0;
|
|
174
|
+
nullCount = 0;
|
|
175
|
+
emptyCount = 0;
|
|
176
|
+
minLength = null;
|
|
177
|
+
maxLength = null;
|
|
178
|
+
totalLength = 0;
|
|
179
|
+
hll = new HyperLogLog();
|
|
180
|
+
frequencyMap = new Map();
|
|
181
|
+
maxTrackedValues = 1000;
|
|
182
|
+
add(val) {
|
|
183
|
+
if (val === null || val === undefined) {
|
|
184
|
+
this.nullCount++;
|
|
185
|
+
return;
|
|
186
|
+
}
|
|
187
|
+
const str = String(val);
|
|
188
|
+
if (str === "") {
|
|
189
|
+
this.emptyCount++;
|
|
190
|
+
}
|
|
191
|
+
this.count++;
|
|
192
|
+
this.totalLength += str.length;
|
|
193
|
+
this.hll.add(str);
|
|
194
|
+
if (this.minLength === null || str.length < this.minLength)
|
|
195
|
+
this.minLength = str.length;
|
|
196
|
+
if (this.maxLength === null || str.length > this.maxLength)
|
|
197
|
+
this.maxLength = str.length;
|
|
198
|
+
// Track top values bounded
|
|
199
|
+
if (this.frequencyMap.size < this.maxTrackedValues || this.frequencyMap.has(str)) {
|
|
200
|
+
this.frequencyMap.set(str, (this.frequencyMap.get(str) || 0) + 1);
|
|
201
|
+
}
|
|
202
|
+
}
|
|
203
|
+
result() {
|
|
204
|
+
const sorted = Array.from(this.frequencyMap.entries())
|
|
205
|
+
.sort((a, b) => b[1] - a[1])
|
|
206
|
+
.slice(0, 5)
|
|
207
|
+
.map(([value, count]) => ({ value, count }));
|
|
208
|
+
return {
|
|
209
|
+
count: this.count,
|
|
210
|
+
nullCount: this.nullCount,
|
|
211
|
+
emptyCount: this.emptyCount,
|
|
212
|
+
minLength: this.minLength,
|
|
213
|
+
maxLength: this.maxLength,
|
|
214
|
+
avgLength: this.count > 0 ? this.totalLength / this.count : null,
|
|
215
|
+
approxDistinct: this.hll.count(),
|
|
216
|
+
topValues: sorted,
|
|
217
|
+
};
|
|
218
|
+
}
|
|
219
|
+
}
|
|
220
|
+
/**
|
|
221
|
+
* Dataset-wide Aggregator collecting statistics across all columns in stream.
|
|
222
|
+
*/
|
|
223
|
+
export class DatasetStatsAggregator {
|
|
224
|
+
totalRows = 0;
|
|
225
|
+
targetColumn;
|
|
226
|
+
numericCollectors = new Map();
|
|
227
|
+
stringCollectors = new Map();
|
|
228
|
+
columnTypeHints = new Map();
|
|
229
|
+
handlers = [];
|
|
230
|
+
initialized = false;
|
|
231
|
+
constructor(options) {
|
|
232
|
+
this.targetColumn = options?.column;
|
|
233
|
+
}
|
|
234
|
+
initHandler(col) {
|
|
235
|
+
let numCollector = this.numericCollectors.get(col);
|
|
236
|
+
let strCollector = this.stringCollectors.get(col);
|
|
237
|
+
let hint = this.columnTypeHints.get(col);
|
|
238
|
+
if (!numCollector) {
|
|
239
|
+
numCollector = new NumericStatsCollector();
|
|
240
|
+
strCollector = new StringStatsCollector();
|
|
241
|
+
hint = { numericVotes: 0, totalVotes: 0 };
|
|
242
|
+
this.numericCollectors.set(col, numCollector);
|
|
243
|
+
this.stringCollectors.set(col, strCollector);
|
|
244
|
+
this.columnTypeHints.set(col, hint);
|
|
245
|
+
}
|
|
246
|
+
return { col, numCollector: numCollector, strCollector: strCollector, hint: hint };
|
|
247
|
+
}
|
|
248
|
+
add(row) {
|
|
249
|
+
this.totalRows++;
|
|
250
|
+
if (!this.initialized) {
|
|
251
|
+
const columns = this.targetColumn ? [this.targetColumn] : Object.keys(row);
|
|
252
|
+
this.handlers = new Array(columns.length);
|
|
253
|
+
for (let i = 0; i < columns.length; i++) {
|
|
254
|
+
this.handlers[i] = this.initHandler(columns[i]);
|
|
255
|
+
}
|
|
256
|
+
this.initialized = true;
|
|
257
|
+
}
|
|
258
|
+
const handlers = this.handlers;
|
|
259
|
+
const len = handlers.length;
|
|
260
|
+
for (let i = 0; i < len; i++) {
|
|
261
|
+
const h = handlers[i];
|
|
262
|
+
const val = row[h.col];
|
|
263
|
+
h.strCollector.add(val);
|
|
264
|
+
if (val !== null && val !== undefined && val !== "") {
|
|
265
|
+
h.hint.totalVotes++;
|
|
266
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
267
|
+
if (!Number.isNaN(num) && typeof val !== "boolean") {
|
|
268
|
+
h.hint.numericVotes++;
|
|
269
|
+
h.numCollector.add(num);
|
|
270
|
+
}
|
|
271
|
+
else {
|
|
272
|
+
h.numCollector.add(null);
|
|
273
|
+
}
|
|
274
|
+
}
|
|
275
|
+
else {
|
|
276
|
+
h.numCollector.add(null);
|
|
277
|
+
}
|
|
278
|
+
}
|
|
279
|
+
}
|
|
280
|
+
result() {
|
|
281
|
+
const columns = {};
|
|
282
|
+
for (const [col, hint] of this.columnTypeHints.entries()) {
|
|
283
|
+
const isNumeric = hint.totalVotes > 0 && hint.numericVotes / hint.totalVotes >= 0.8;
|
|
284
|
+
const numCollector = this.numericCollectors.get(col);
|
|
285
|
+
const strCollector = this.stringCollectors.get(col);
|
|
286
|
+
if (isNumeric) {
|
|
287
|
+
columns[col] = {
|
|
288
|
+
column: col,
|
|
289
|
+
type: "numeric",
|
|
290
|
+
numeric: numCollector.result(),
|
|
291
|
+
};
|
|
292
|
+
}
|
|
293
|
+
else {
|
|
294
|
+
columns[col] = {
|
|
295
|
+
column: col,
|
|
296
|
+
type: "string",
|
|
297
|
+
string: strCollector.result(),
|
|
298
|
+
};
|
|
299
|
+
}
|
|
300
|
+
}
|
|
301
|
+
return {
|
|
302
|
+
totalRows: this.totalRows,
|
|
303
|
+
columns,
|
|
304
|
+
};
|
|
305
|
+
}
|
|
306
|
+
}
|
|
307
|
+
//# sourceMappingURL=stats.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"stats.js","sourceRoot":"","sources":["../../src/analytics/stats.ts"],"names":[],"mappings":"AAEA;;;GAGG;AACH,SAAS,MAAM,CAAC,GAAW;IACzB,IAAI,CAAC,GAAG,UAAU,CAAC;IACnB,MAAM,GAAG,GAAG,GAAG,CAAC,MAAM,CAAC;IACvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7B,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;QAClC,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,CAAC,KAAK,CAAC,CAAC;IACrC,CAAC;IACD,mBAAmB;IACnB,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;IACd,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,CAAC,KAAK,CAAC,CAAC;IACnC,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;IACd,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,CAAC,KAAK,CAAC,CAAC;IACnC,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;IACd,OAAO,CAAC,KAAK,CAAC,CAAC;AACjB,CAAC;AAED;;;;GAIG;AACH,MAAM,OAAO,WAAW;IACd,CAAC,CAAS;IACV,CAAC,CAAS;IACV,SAAS,CAAa;IACtB,OAAO,CAAS;IAExB,YAAY,CAAC,GAAG,EAAE;QAChB,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QACX,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAChB,IAAI,CAAC,SAAS,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAExC,6BAA6B;QAC7B,IAAI,IAAI,CAAC,CAAC,KAAK,EAAE;YAAE,IAAI,CAAC,OAAO,GAAG,KAAK,GAAG,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;aACrD,IAAI,IAAI,CAAC,CAAC,KAAK,EAAE;YAAE,IAAI,CAAC,OAAO,GAAG,KAAK,GAAG,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;aAC1D,IAAI,IAAI,CAAC,CAAC,KAAK,EAAE;YAAE,IAAI,CAAC,OAAO,GAAG,KAAK,GAAG,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;;YAC1D,IAAI,CAAC,OAAO,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;IACxE,CAAC;IAED,GAAG,CAAC,KAAc;QAChB,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS;YAAE,OAAO;QAClD,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;QAC1B,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QAEzB,gCAAgC;QAChC,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAClC,4CAA4C;QAC5C,MAAM,CAAC,GAAG,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC;QAE1B,iEAAiE;QACjE,IAAI,YAAoB,CAAC;QACzB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;YACZ,YAAY,GAAG,EAAE,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QACjC,CAAC;aAAM,CAAC;YACN,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QAC5C,CAAC;QAED,IAAI,YAAY,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAE,EAAE,CAAC;YAC1C,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,GAAG,YAAY,CAAC;QACvC,CAAC;IACH,CAAC;IAED,KAAK,CAAC,KAAkB;QACtB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAChC,IAAI,KAAK,CAAC,SAAS,CAAC,CAAC,CAAE,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAE,EAAE,CAAC;gBAC7C,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,CAAE,CAAC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IAED,KAAK;QACH,IAAI,GAAG,GAAG,CAAC,CAAC;QACZ,IAAI,KAAK,GAAG,CAAC,CAAC;QAEd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAChC,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAE,CAAC;YAC/B,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;YACjB,IAAI,GAAG,KAAK,CAAC;gBAAE,KAAK,EAAE,CAAC;QACzB,CAAC;QAED,IAAI,QAAQ,GAAG,IAAI,CAAC,OAAO,GAAG,GAAG,CAAC;QAElC,2CAA2C;QAC3C,IAAI,QAAQ,IAAI,GAAG,GAAG,IAAI,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YAC1C,QAAQ,GAAG,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC;QAC/C,CAAC;QAED,OAAO,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IAC9B,CAAC;CACF;AAsCD;;GAEG;AACH,MAAM,OAAO,qBAAqB;IACzB,KAAK,GAAG,CAAC,CAAC;IACV,SAAS,GAAG,CAAC,CAAC;IACd,GAAG,GAAkB,IAAI,CAAC;IAC1B,GAAG,GAAkB,IAAI,CAAC;IAC1B,GAAG,GAAG,CAAC,CAAC;IACR,IAAI,GAAG,CAAC,CAAC;IACT,EAAE,GAAG,CAAC,CAAC,CAAC,2CAA2C;IACnD,GAAG,GAAG,IAAI,WAAW,EAAE,CAAC;IAE/B,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;YACpD,IAAI,CAAC,SAAS,EAAE,CAAC;YACjB,OAAO;QACT,CAAC;QAED,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;YACtB,IAAI,CAAC,SAAS,EAAE,CAAC;YACjB,OAAO;QACT,CAAC;QAED,IAAI,CAAC,KAAK,EAAE,CAAC;QACb,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC;QAChB,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAElB,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,IAAI,GAAG,GAAG,IAAI,CAAC,GAAG;YAAE,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;QACxD,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,IAAI,GAAG,GAAG,IAAI,CAAC,GAAG;YAAE,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;QAExD,6BAA6B;QAC7B,MAAM,KAAK,GAAG,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC;QAC9B,IAAI,CAAC,IAAI,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC;QAChC,MAAM,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC;QAC/B,IAAI,CAAC,EAAE,IAAI,KAAK,GAAG,MAAM,CAAC;IAC5B,CAAC;IAED,KAAK,CAAC,KAA4B;QAChC,IAAI,KAAK,CAAC,KAAK,KAAK,CAAC,EAAE,CAAC;YACtB,IAAI,CAAC,SAAS,IAAI,KAAK,CAAC,SAAS,CAAC;YAClC,OAAO;QACT,CAAC;QACD,IAAI,IAAI,CAAC,KAAK,KAAK,CAAC,EAAE,CAAC;YACrB,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;YACzB,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC,SAAS,CAAC;YACjC,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;YACrB,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;YACrB,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;YACrB,IAAI,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;YACvB,IAAI,CAAC,EAAE,GAAG,KAAK,CAAC,EAAE,CAAC;YACnB,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;YAC1B,OAAO;QACT,CAAC;QAED,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;QAC5C,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;QAErC,IAAI,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,IAAI,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,UAAU,CAAC;QAC7E,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,EAAE,GAAG,KAAK,CAAC,EAAE,GAAG,CAAC,KAAK,GAAG,KAAK,GAAG,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,UAAU,CAAC;QACvF,IAAI,CAAC,KAAK,GAAG,UAAU,CAAC;QACxB,IAAI,CAAC,SAAS,IAAI,KAAK,CAAC,SAAS,CAAC;QAClC,IAAI,CAAC,GAAG,IAAI,KAAK,CAAC,GAAG,CAAC;QAEtB,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,KAAK,IAAI,IAAI,KAAK,CAAC,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC;YAAE,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;QAC5F,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,KAAK,IAAI,IAAI,KAAK,CAAC,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC;YAAE,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC;QAC5F,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC5B,CAAC;IAED,MAAM;QACJ,MAAM,QAAQ,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QAC3F,MAAM,MAAM,GAAG,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QAE9D,OAAO;YACL,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,GAAG,EAAE,IAAI,CAAC,GAAG;YACb,GAAG,EAAE,IAAI,CAAC,GAAG;YACb,GAAG,EAAE,IAAI,CAAC,GAAG;YACb,IAAI,EAAE,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI;YACvC,QAAQ;YACR,MAAM;YACN,cAAc,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE;SACjC,CAAC;IACJ,CAAC;CACF;AAED;;GAEG;AACH,MAAM,OAAO,oBAAoB;IACxB,KAAK,GAAG,CAAC,CAAC;IACV,SAAS,GAAG,CAAC,CAAC;IACd,UAAU,GAAG,CAAC,CAAC;IACf,SAAS,GAAkB,IAAI,CAAC;IAChC,SAAS,GAAkB,IAAI,CAAC;IAChC,WAAW,GAAG,CAAC,CAAC;IAChB,GAAG,GAAG,IAAI,WAAW,EAAE,CAAC;IACvB,YAAY,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,gBAAgB,GAAG,IAAI,CAAC;IAEhC,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACtC,IAAI,CAAC,SAAS,EAAE,CAAC;YACjB,OAAO;QACT,CAAC;QAED,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QACxB,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;YACf,IAAI,CAAC,UAAU,EAAE,CAAC;QACpB,CAAC;QAED,IAAI,CAAC,KAAK,EAAE,CAAC;QACb,IAAI,CAAC,WAAW,IAAI,GAAG,CAAC,MAAM,CAAC;QAC/B,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAElB,IAAI,IAAI,CAAC,SAAS,KAAK,IAAI,IAAI,GAAG,CAAC,MAAM,GAAG,IAAI,CAAC,SAAS;YAAE,IAAI,CAAC,SAAS,GAAG,GAAG,CAAC,MAAM,CAAC;QACxF,IAAI,IAAI,CAAC,SAAS,KAAK,IAAI,IAAI,GAAG,CAAC,MAAM,GAAG,IAAI,CAAC,SAAS;YAAE,IAAI,CAAC,SAAS,GAAG,GAAG,CAAC,MAAM,CAAC;QAExF,2BAA2B;QAC3B,IAAI,IAAI,CAAC,YAAY,CAAC,IAAI,GAAG,IAAI,CAAC,gBAAgB,IAAI,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACjF,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QACpE,CAAC;IACH,CAAC;IAED,MAAM;QACJ,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,YAAY,CAAC,OAAO,EAAE,CAAC;aACnD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;aAC3B,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;aACX,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;QAE/C,OAAO;YACL,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,UAAU,EAAE,IAAI,CAAC,UAAU;YAC3B,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,SAAS,EAAE,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI;YAChE,cAAc,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE;YAChC,SAAS,EAAE,MAAM;SAClB,CAAC;IACJ,CAAC;CACF;AASD;;GAEG;AACH,MAAM,OAAO,sBAAsB;IACzB,SAAS,GAAG,CAAC,CAAC;IACd,YAAY,CAAU;IACtB,iBAAiB,GAAG,IAAI,GAAG,EAAiC,CAAC;IAC7D,gBAAgB,GAAG,IAAI,GAAG,EAAgC,CAAC;IAC3D,eAAe,GAAG,IAAI,GAAG,EAAwD,CAAC;IAClF,QAAQ,GAAoB,EAAE,CAAC;IAC/B,WAAW,GAAG,KAAK,CAAC;IAE5B,YAAY,OAA6B;QACvC,IAAI,CAAC,YAAY,GAAG,OAAO,EAAE,MAAM,CAAC;IACtC,CAAC;IAEO,WAAW,CAAC,GAAW;QAC7B,IAAI,YAAY,GAAG,IAAI,CAAC,iBAAiB,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACnD,IAAI,YAAY,GAAG,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAClD,IAAI,IAAI,GAAG,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAEzC,IAAI,CAAC,YAAY,EAAE,CAAC;YAClB,YAAY,GAAG,IAAI,qBAAqB,EAAE,CAAC;YAC3C,YAAY,GAAG,IAAI,oBAAoB,EAAE,CAAC;YAC1C,IAAI,GAAG,EAAE,YAAY,EAAE,CAAC,EAAE,UAAU,EAAE,CAAC,EAAE,CAAC;YAC1C,IAAI,CAAC,iBAAiB,CAAC,GAAG,CAAC,GAAG,EAAE,YAAY,CAAC,CAAC;YAC9C,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,GAAG,EAAE,YAAY,CAAC,CAAC;YAC7C,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,EAAE,GAAG,EAAE,YAAY,EAAE,YAAa,EAAE,YAAY,EAAE,YAAa,EAAE,IAAI,EAAE,IAAK,EAAE,CAAC;IACxF,CAAC;IAED,GAAG,CAAC,GAAQ;QACV,IAAI,CAAC,SAAS,EAAE,CAAC;QAEjB,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;YACtB,MAAM,OAAO,GAAG,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAC3E,IAAI,CAAC,QAAQ,GAAG,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;YAC1C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACxC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,CAAE,CAAC,CAAC;YACnD,CAAC;YACD,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC;QAC1B,CAAC;QAED,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;QAC/B,MAAM,GAAG,GAAG,QAAQ,CAAC,MAAM,CAAC;QAE5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7B,MAAM,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC;YACvB,MAAM,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YAEvB,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAExB,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;gBACpD,CAAC,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;gBACpB,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;gBACxD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,KAAK,SAAS,EAAE,CAAC;oBACnD,CAAC,CAAC,IAAI,CAAC,YAAY,EAAE,CAAC;oBACtB,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;gBAC1B,CAAC;qBAAM,CAAC;oBACN,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;gBAC3B,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAC3B,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM;QACJ,MAAM,OAAO,GAAsC,EAAE,CAAC;QAEtD,KAAK,MAAM,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,eAAe,CAAC,OAAO,EAAE,EAAE,CAAC;YACzD,MAAM,SAAS,GAAG,IAAI,CAAC,UAAU,GAAG,CAAC,IAAI,IAAI,CAAC,YAAY,GAAG,IAAI,CAAC,UAAU,IAAI,GAAG,CAAC;YACpF,MAAM,YAAY,GAAG,IAAI,CAAC,iBAAiB,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC;YACtD,MAAM,YAAY,GAAG,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC;YAErD,IAAI,SAAS,EAAE,CAAC;gBACd,OAAO,CAAC,GAAG,CAAC,GAAG;oBACb,MAAM,EAAE,GAAG;oBACX,IAAI,EAAE,SAAS;oBACf,OAAO,EAAE,YAAY,CAAC,MAAM,EAAE;iBAC/B,CAAC;YACJ,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,GAAG,CAAC,GAAG;oBACb,MAAM,EAAE,GAAG;oBACX,IAAI,EAAE,QAAQ;oBACd,MAAM,EAAE,YAAY,CAAC,MAAM,EAAE;iBAC9B,CAAC;YACJ,CAAC;QACH,CAAC;QAED,OAAO;YACL,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,OAAO;SACR,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,32 @@
|
|
|
1
|
+
import type { Aggregator, Row } from "../core/types.js";
|
|
2
|
+
export interface ColumnSchemaRule {
|
|
3
|
+
type?: string;
|
|
4
|
+
nullable?: boolean;
|
|
5
|
+
format?: string;
|
|
6
|
+
}
|
|
7
|
+
export type ValidationSchemaDefinition = Record<string, ColumnSchemaRule>;
|
|
8
|
+
export interface ColumnViolationSummary {
|
|
9
|
+
column: string;
|
|
10
|
+
rule: string;
|
|
11
|
+
violationsCount: number;
|
|
12
|
+
sampleInvalidValues: unknown[];
|
|
13
|
+
}
|
|
14
|
+
export interface ValidationReport {
|
|
15
|
+
totalRows: number;
|
|
16
|
+
validRows: number;
|
|
17
|
+
invalidRows: number;
|
|
18
|
+
isValid: boolean;
|
|
19
|
+
violations: ColumnViolationSummary[];
|
|
20
|
+
}
|
|
21
|
+
export declare class SchemaValidatorAggregator implements Aggregator<ValidationReport> {
|
|
22
|
+
private schema;
|
|
23
|
+
private totalRows;
|
|
24
|
+
private validRows;
|
|
25
|
+
private invalidRows;
|
|
26
|
+
private violationsMap;
|
|
27
|
+
constructor(schema: ValidationSchemaDefinition);
|
|
28
|
+
private addViolation;
|
|
29
|
+
add(row: Row): void;
|
|
30
|
+
result(): ValidationReport;
|
|
31
|
+
}
|
|
32
|
+
//# sourceMappingURL=validator.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"validator.d.ts","sourceRoot":"","sources":["../../src/analytics/validator.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,kBAAkB,CAAC;AAIxD,MAAM,WAAW,gBAAgB;IAC/B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB;AAED,MAAM,MAAM,0BAA0B,GAAG,MAAM,CAAC,MAAM,EAAE,gBAAgB,CAAC,CAAC;AAE1E,MAAM,WAAW,sBAAsB;IACrC,MAAM,EAAE,MAAM,CAAC;IACf,IAAI,EAAE,MAAM,CAAC;IACb,eAAe,EAAE,MAAM,CAAC;IACxB,mBAAmB,EAAE,OAAO,EAAE,CAAC;CAChC;AAED,MAAM,WAAW,gBAAgB;IAC/B,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,OAAO,EAAE,OAAO,CAAC;IACjB,UAAU,EAAE,sBAAsB,EAAE,CAAC;CACtC;AAED,qBAAa,yBAA0B,YAAW,UAAU,CAAC,gBAAgB,CAAC;IAC5E,OAAO,CAAC,MAAM,CAA6B;IAC3C,OAAO,CAAC,SAAS,CAAK;IACtB,OAAO,CAAC,SAAS,CAAK;IACtB,OAAO,CAAC,WAAW,CAAK;IACxB,OAAO,CAAC,aAAa,CAA6C;gBAEtD,MAAM,EAAE,0BAA0B;IAI9C,OAAO,CAAC,YAAY;IAkBpB,GAAG,CAAC,GAAG,EAAE,GAAG,GAAG,IAAI;IAoEnB,MAAM,IAAI,gBAAgB;CAS3B"}
|
|
@@ -0,0 +1,96 @@
|
|
|
1
|
+
import { classifyPrimitiveType } from "./schema-inference.js";
|
|
2
|
+
import { detectSemanticType } from "./semantic-types.js";
|
|
3
|
+
export class SchemaValidatorAggregator {
|
|
4
|
+
schema;
|
|
5
|
+
totalRows = 0;
|
|
6
|
+
validRows = 0;
|
|
7
|
+
invalidRows = 0;
|
|
8
|
+
violationsMap = new Map();
|
|
9
|
+
constructor(schema) {
|
|
10
|
+
this.schema = schema;
|
|
11
|
+
}
|
|
12
|
+
addViolation(column, rule, value) {
|
|
13
|
+
const key = `${column}::${rule}`;
|
|
14
|
+
if (!this.violationsMap.has(key)) {
|
|
15
|
+
this.violationsMap.set(key, {
|
|
16
|
+
column,
|
|
17
|
+
rule,
|
|
18
|
+
violationsCount: 0,
|
|
19
|
+
sampleInvalidValues: [],
|
|
20
|
+
});
|
|
21
|
+
}
|
|
22
|
+
const v = this.violationsMap.get(key);
|
|
23
|
+
v.violationsCount++;
|
|
24
|
+
if (v.sampleInvalidValues.length < 5) {
|
|
25
|
+
v.sampleInvalidValues.push(value);
|
|
26
|
+
}
|
|
27
|
+
}
|
|
28
|
+
add(row) {
|
|
29
|
+
this.totalRows++;
|
|
30
|
+
let rowHasError = false;
|
|
31
|
+
for (const [colName, rule] of Object.entries(this.schema)) {
|
|
32
|
+
const val = row[colName];
|
|
33
|
+
const isNull = val === null || val === undefined || val === "";
|
|
34
|
+
// Nullability check
|
|
35
|
+
if (isNull) {
|
|
36
|
+
if (rule.nullable === false) {
|
|
37
|
+
this.addViolation(colName, "null not allowed", val);
|
|
38
|
+
rowHasError = true;
|
|
39
|
+
}
|
|
40
|
+
continue;
|
|
41
|
+
}
|
|
42
|
+
// Type check
|
|
43
|
+
if (rule.type) {
|
|
44
|
+
const actualType = classifyPrimitiveType(val);
|
|
45
|
+
const expectedType = rule.type.toLowerCase();
|
|
46
|
+
let matchesType = false;
|
|
47
|
+
if (expectedType === "string") {
|
|
48
|
+
matchesType = true;
|
|
49
|
+
}
|
|
50
|
+
else if (expectedType === "integer" || expectedType === "int") {
|
|
51
|
+
matchesType = actualType === "integer";
|
|
52
|
+
}
|
|
53
|
+
else if (expectedType === "number" || expectedType === "float") {
|
|
54
|
+
matchesType = actualType === "integer" || actualType === "number";
|
|
55
|
+
}
|
|
56
|
+
else if (expectedType === "boolean" || expectedType === "bool") {
|
|
57
|
+
matchesType = actualType === "boolean";
|
|
58
|
+
}
|
|
59
|
+
else if (expectedType === "date") {
|
|
60
|
+
matchesType = actualType === "date" || actualType === "datetime";
|
|
61
|
+
}
|
|
62
|
+
else if (expectedType === "datetime") {
|
|
63
|
+
matchesType = actualType === "datetime";
|
|
64
|
+
}
|
|
65
|
+
if (!matchesType) {
|
|
66
|
+
this.addViolation(colName, `expected ${rule.type}`, val);
|
|
67
|
+
rowHasError = true;
|
|
68
|
+
}
|
|
69
|
+
}
|
|
70
|
+
// Format check (semantic type)
|
|
71
|
+
if (rule.format) {
|
|
72
|
+
const detected = detectSemanticType(val);
|
|
73
|
+
if (detected !== rule.format.toLowerCase()) {
|
|
74
|
+
this.addViolation(colName, `expected format ${rule.format}`, val);
|
|
75
|
+
rowHasError = true;
|
|
76
|
+
}
|
|
77
|
+
}
|
|
78
|
+
}
|
|
79
|
+
if (rowHasError) {
|
|
80
|
+
this.invalidRows++;
|
|
81
|
+
}
|
|
82
|
+
else {
|
|
83
|
+
this.validRows++;
|
|
84
|
+
}
|
|
85
|
+
}
|
|
86
|
+
result() {
|
|
87
|
+
return {
|
|
88
|
+
totalRows: this.totalRows,
|
|
89
|
+
validRows: this.validRows,
|
|
90
|
+
invalidRows: this.invalidRows,
|
|
91
|
+
isValid: this.invalidRows === 0,
|
|
92
|
+
violations: Array.from(this.violationsMap.values()),
|
|
93
|
+
};
|
|
94
|
+
}
|
|
95
|
+
}
|
|
96
|
+
//# sourceMappingURL=validator.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"validator.js","sourceRoot":"","sources":["../../src/analytics/validator.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAC9D,OAAO,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAC;AAyBzD,MAAM,OAAO,yBAAyB;IAC5B,MAAM,CAA6B;IACnC,SAAS,GAAG,CAAC,CAAC;IACd,SAAS,GAAG,CAAC,CAAC;IACd,WAAW,GAAG,CAAC,CAAC;IAChB,aAAa,GAAG,IAAI,GAAG,EAAkC,CAAC;IAElE,YAAY,MAAkC;QAC5C,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAEO,YAAY,CAAC,MAAc,EAAE,IAAY,EAAE,KAAc;QAC/D,MAAM,GAAG,GAAG,GAAG,MAAM,KAAK,IAAI,EAAE,CAAC;QACjC,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACjC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,GAAG,EAAE;gBAC1B,MAAM;gBACN,IAAI;gBACJ,eAAe,EAAE,CAAC;gBAClB,mBAAmB,EAAE,EAAE;aACxB,CAAC,CAAC;QACL,CAAC;QAED,MAAM,CAAC,GAAG,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC;QACvC,CAAC,CAAC,eAAe,EAAE,CAAC;QACpB,IAAI,CAAC,CAAC,mBAAmB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrC,CAAC,CAAC,mBAAmB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IAED,GAAG,CAAC,GAAQ;QACV,IAAI,CAAC,SAAS,EAAE,CAAC;QACjB,IAAI,WAAW,GAAG,KAAK,CAAC;QAExB,KAAK,MAAM,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;YAC1D,MAAM,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,CAAC;YACzB,MAAM,MAAM,GAAG,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,CAAC;YAE/D,oBAAoB;YACpB,IAAI,MAAM,EAAE,CAAC;gBACX,IAAI,IAAI,CAAC,QAAQ,KAAK,KAAK,EAAE,CAAC;oBAC5B,IAAI,CAAC,YAAY,CAAC,OAAO,EAAE,kBAAkB,EAAE,GAAG,CAAC,CAAC;oBACpD,WAAW,GAAG,IAAI,CAAC;gBACrB,CAAC;gBACD,SAAS;YACX,CAAC;YAED,aAAa;YACb,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;gBACd,MAAM,UAAU,GAAG,qBAAqB,CAAC,GAAG,CAAC,CAAC;gBAC9C,MAAM,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;gBAE7C,IAAI,WAAW,GAAG,KAAK,CAAC;gBACxB,IAAI,YAAY,KAAK,QAAQ,EAAE,CAAC;oBAC9B,WAAW,GAAG,IAAI,CAAC;gBACrB,CAAC;qBAAM,IAAI,YAAY,KAAK,SAAS,IAAI,YAAY,KAAK,KAAK,EAAE,CAAC;oBAChE,WAAW,GAAG,UAAU,KAAK,SAAS,CAAC;gBACzC,CAAC;qBAAM,IAAI,YAAY,KAAK,QAAQ,IAAI,YAAY,KAAK,OAAO,EAAE,CAAC;oBACjE,WAAW,GAAG,UAAU,KAAK,SAAS,IAAI,UAAU,KAAK,QAAQ,CAAC;gBACpE,CAAC;qBAAM,IAAI,YAAY,KAAK,SAAS,IAAI,YAAY,KAAK,MAAM,EAAE,CAAC;oBACjE,WAAW,GAAG,UAAU,KAAK,SAAS,CAAC;gBACzC,CAAC;qBAAM,IAAI,YAAY,KAAK,MAAM,EAAE,CAAC;oBACnC,WAAW,GAAG,UAAU,KAAK,MAAM,IAAI,UAAU,KAAK,UAAU,CAAC;gBACnE,CAAC;qBAAM,IAAI,YAAY,KAAK,UAAU,EAAE,CAAC;oBACvC,WAAW,GAAG,UAAU,KAAK,UAAU,CAAC;gBAC1C,CAAC;gBAED,IAAI,CAAC,WAAW,EAAE,CAAC;oBACjB,IAAI,CAAC,YAAY,CACf,OAAO,EACP,YAAY,IAAI,CAAC,IAAI,EAAE,EACvB,GAAG,CACJ,CAAC;oBACF,WAAW,GAAG,IAAI,CAAC;gBACrB,CAAC;YACH,CAAC;YAED,+BAA+B;YAC/B,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;gBAChB,MAAM,QAAQ,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC;gBACzC,IAAI,QAAQ,KAAK,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,EAAE,CAAC;oBAC3C,IAAI,CAAC,YAAY,CACf,OAAO,EACP,mBAAmB,IAAI,CAAC,MAAM,EAAE,EAChC,GAAG,CACJ,CAAC;oBACF,WAAW,GAAG,IAAI,CAAC;gBACrB,CAAC;YACH,CAAC;QACH,CAAC;QAED,IAAI,WAAW,EAAE,CAAC;YAChB,IAAI,CAAC,WAAW,EAAE,CAAC;QACrB,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,SAAS,EAAE,CAAC;QACnB,CAAC;IACH,CAAC;IAED,MAAM;QACJ,OAAO;YACL,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,WAAW,EAAE,IAAI,CAAC,WAAW;YAC7B,OAAO,EAAE,IAAI,CAAC,WAAW,KAAK,CAAC;YAC/B,UAAU,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,MAAM,EAAE,CAAC;SACpD,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
import { type CastOnError } from "../../transforms/cast.js";
|
|
2
|
+
export interface CastCommandOptions {
|
|
3
|
+
onError?: CastOnError;
|
|
4
|
+
from?: string;
|
|
5
|
+
to?: string;
|
|
6
|
+
sheet?: string;
|
|
7
|
+
delimiter?: string;
|
|
8
|
+
batchSize?: string | number;
|
|
9
|
+
quiet?: boolean;
|
|
10
|
+
noProgress?: boolean;
|
|
11
|
+
}
|
|
12
|
+
export declare function castCommand(inputPath: string | undefined, specs: string[], options?: CastCommandOptions): Promise<void>;
|
|
13
|
+
//# sourceMappingURL=cast.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cast.d.ts","sourceRoot":"","sources":["../../../src/cli/commands/cast.ts"],"names":[],"mappings":"AAGA,OAAO,EAAe,KAAK,WAAW,EAAkB,MAAM,0BAA0B,CAAC;AAKzF,MAAM,WAAW,kBAAkB;IACjC,OAAO,CAAC,EAAE,WAAW,CAAC;IACtB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,EAAE,CAAC,EAAE,MAAM,CAAC;IACZ,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,SAAS,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAC5B,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,UAAU,CAAC,EAAE,OAAO,CAAC;CACtB;AAED,wBAAsB,WAAW,CAC/B,SAAS,oBAAM,EACf,KAAK,EAAE,MAAM,EAAE,EACf,OAAO,GAAE,kBAAuB,GAC/B,OAAO,CAAC,IAAI,CAAC,CAoCf"}
|
|
@@ -0,0 +1,39 @@
|
|
|
1
|
+
import { createPipeline } from "../../core/pipeline.js";
|
|
2
|
+
import { createReader, inferFormatFromPath as inferReaderFormat } from "../../readers/index.js";
|
|
3
|
+
import { createWriter } from "../../writers/index.js";
|
|
4
|
+
import { castColumns, parseCastSpecs } from "../../transforms/cast.js";
|
|
5
|
+
import { openReadableStream } from "../../utils/compression.js";
|
|
6
|
+
import { logMemoryDebug } from "../../utils/formatting.js";
|
|
7
|
+
import { ProgressReporter } from "../../utils/progress.js";
|
|
8
|
+
export async function castCommand(inputPath = "-", specs, options = {}) {
|
|
9
|
+
const effectiveBatchSize = Number(options.batchSize) || 1000;
|
|
10
|
+
const progress = new ProgressReporter(options);
|
|
11
|
+
let fromFormat = options.from?.toLowerCase();
|
|
12
|
+
if (!fromFormat && inputPath !== "-") {
|
|
13
|
+
fromFormat = inferReaderFormat(inputPath) ?? undefined;
|
|
14
|
+
}
|
|
15
|
+
if (!fromFormat) {
|
|
16
|
+
fromFormat = "csv";
|
|
17
|
+
}
|
|
18
|
+
const toFormat = options.to?.toLowerCase() || (fromFormat === "xlsx" ? "csv" : fromFormat);
|
|
19
|
+
const mapping = parseCastSpecs(specs);
|
|
20
|
+
const inputStream = openReadableStream(inputPath);
|
|
21
|
+
const reader = createReader(inputStream, {
|
|
22
|
+
format: fromFormat,
|
|
23
|
+
sheet: options.sheet,
|
|
24
|
+
delimiter: options.delimiter,
|
|
25
|
+
batchSize: effectiveBatchSize,
|
|
26
|
+
filePath: inputPath,
|
|
27
|
+
});
|
|
28
|
+
const writer = createWriter(process.stdout, {
|
|
29
|
+
format: toFormat,
|
|
30
|
+
delimiter: options.delimiter,
|
|
31
|
+
});
|
|
32
|
+
const pipeline = createPipeline(reader, { batchSize: effectiveBatchSize });
|
|
33
|
+
pipeline.pipe(castColumns(mapping, { onError: options.onError }));
|
|
34
|
+
pipeline.onProgress((info) => progress.update(info));
|
|
35
|
+
await pipeline.to(writer);
|
|
36
|
+
progress.done();
|
|
37
|
+
logMemoryDebug();
|
|
38
|
+
}
|
|
39
|
+
//# sourceMappingURL=cast.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cast.js","sourceRoot":"","sources":["../../../src/cli/commands/cast.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,YAAY,EAAE,mBAAmB,IAAI,iBAAiB,EAAE,MAAM,wBAAwB,CAAC;AAChG,OAAO,EAAE,YAAY,EAAE,MAAM,wBAAwB,CAAC;AACtD,OAAO,EAAE,WAAW,EAAoB,cAAc,EAAE,MAAM,0BAA0B,CAAC;AACzF,OAAO,EAAE,kBAAkB,EAAE,MAAM,4BAA4B,CAAC;AAChE,OAAO,EAAE,cAAc,EAAE,MAAM,2BAA2B,CAAC;AAC3D,OAAO,EAAE,gBAAgB,EAAE,MAAM,yBAAyB,CAAC;AAa3D,MAAM,CAAC,KAAK,UAAU,WAAW,CAC/B,SAAS,GAAG,GAAG,EACf,KAAe,EACf,UAA8B,EAAE;IAEhC,MAAM,kBAAkB,GAAG,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,IAAI,IAAI,CAAC;IAC7D,MAAM,QAAQ,GAAG,IAAI,gBAAgB,CAAC,OAAO,CAAC,CAAC;IAE/C,IAAI,UAAU,GAAG,OAAO,CAAC,IAAI,EAAE,WAAW,EAAE,CAAC;IAC7C,IAAI,CAAC,UAAU,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;QACrC,UAAU,GAAG,iBAAiB,CAAC,SAAS,CAAC,IAAI,SAAS,CAAC;IACzD,CAAC;IACD,IAAI,CAAC,UAAU,EAAE,CAAC;QAChB,UAAU,GAAG,KAAK,CAAC;IACrB,CAAC;IAED,MAAM,QAAQ,GAAG,OAAO,CAAC,EAAE,EAAE,WAAW,EAAE,IAAI,CAAC,UAAU,KAAK,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAC3F,MAAM,OAAO,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC;IAEtC,MAAM,WAAW,GAAG,kBAAkB,CAAC,SAAS,CAAC,CAAC;IAClD,MAAM,MAAM,GAAG,YAAY,CAAC,WAAW,EAAE;QACvC,MAAM,EAAE,UAAU;QAClB,KAAK,EAAE,OAAO,CAAC,KAAK;QACpB,SAAS,EAAE,OAAO,CAAC,SAAS;QAC5B,SAAS,EAAE,kBAAkB;QAC7B,QAAQ,EAAE,SAAS;KACpB,CAAC,CAAC;IAEH,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,CAAC,MAAM,EAAE;QAC1C,MAAM,EAAE,QAAQ;QAChB,SAAS,EAAE,OAAO,CAAC,SAAS;KAC7B,CAAC,CAAC;IAEH,MAAM,QAAQ,GAAG,cAAc,CAAC,MAAM,EAAE,EAAE,SAAS,EAAE,kBAAkB,EAAE,CAAC,CAAC;IAC3E,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,EAAE,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IAClE,QAAQ,CAAC,UAAU,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC;IAErD,MAAM,QAAQ,CAAC,EAAE,CAAC,MAAM,CAAC,CAAC;IAC1B,QAAQ,CAAC,IAAI,EAAE,CAAC;IAChB,cAAc,EAAE,CAAC;AACnB,CAAC"}
|
|
@@ -0,0 +1,15 @@
|
|
|
1
|
+
export interface ConvertCommandOptions {
|
|
2
|
+
from?: string;
|
|
3
|
+
to?: string;
|
|
4
|
+
sheet?: string;
|
|
5
|
+
allSheets?: boolean;
|
|
6
|
+
outDir?: string;
|
|
7
|
+
path?: string;
|
|
8
|
+
delimiter?: string;
|
|
9
|
+
header?: boolean;
|
|
10
|
+
batchSize?: string | number;
|
|
11
|
+
quiet?: boolean;
|
|
12
|
+
noProgress?: boolean;
|
|
13
|
+
}
|
|
14
|
+
export declare function convertCommand(inputPath?: string, outputPath?: string, options?: ConvertCommandOptions): Promise<void>;
|
|
15
|
+
//# sourceMappingURL=convert.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"convert.d.ts","sourceRoot":"","sources":["../../../src/cli/commands/convert.ts"],"names":[],"mappings":"AAUA,MAAM,WAAW,qBAAqB;IACpC,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,EAAE,CAAC,EAAE,MAAM,CAAC;IACZ,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,SAAS,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAC5B,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,UAAU,CAAC,EAAE,OAAO,CAAC;CACtB;AAED,wBAAsB,cAAc,CAClC,SAAS,SAAM,EACf,UAAU,CAAC,EAAE,MAAM,EACnB,OAAO,GAAE,qBAA0B,GAClC,OAAO,CAAC,IAAI,CAAC,CAmGf"}
|