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 +1,318 @@
|
|
|
1
|
+
import { InvalidArgumentError } from "../core/errors.js";
|
|
2
|
+
import { compileValueExpression } from "../transforms/expression.js";
|
|
3
|
+
import { HyperLogLog } from "./stats.js";
|
|
4
|
+
/**
|
|
5
|
+
* Parses CLI reduce specs into structured AggregationSpec objects.
|
|
6
|
+
* Supports syntax: "total_rev = sum(revenue)", "avg_margin = avg(margin)", "orders = count()"
|
|
7
|
+
*/
|
|
8
|
+
export function parseReduceSpecs(specs) {
|
|
9
|
+
const result = [];
|
|
10
|
+
for (const spec of specs) {
|
|
11
|
+
const trimmed = spec.trim();
|
|
12
|
+
if (!trimmed)
|
|
13
|
+
continue;
|
|
14
|
+
const eqIdx = trimmed.indexOf("=");
|
|
15
|
+
if (eqIdx === -1) {
|
|
16
|
+
throw new InvalidArgumentError(`Invalid reduce specification "${spec}". Expected format "target_column = func(source_expr)"`);
|
|
17
|
+
}
|
|
18
|
+
const targetField = trimmed.slice(0, eqIdx).trim();
|
|
19
|
+
const rhs = trimmed.slice(eqIdx + 1).trim();
|
|
20
|
+
const match = rhs.match(/^([a-zA-Z_]+)\s*\((.*)\)$/);
|
|
21
|
+
if (!match) {
|
|
22
|
+
throw new InvalidArgumentError(`Invalid aggregation function call in "${spec}". Expected format "func(column)" e.g. "sum(revenue)" or "count()"`);
|
|
23
|
+
}
|
|
24
|
+
const rawFunc = match[1].toLowerCase();
|
|
25
|
+
const sourceExpr = match[2].trim() || undefined;
|
|
26
|
+
const validFuncs = [
|
|
27
|
+
"sum",
|
|
28
|
+
"avg",
|
|
29
|
+
"mean",
|
|
30
|
+
"min",
|
|
31
|
+
"max",
|
|
32
|
+
"count",
|
|
33
|
+
"countdistinct",
|
|
34
|
+
"stddev",
|
|
35
|
+
"variance",
|
|
36
|
+
"first",
|
|
37
|
+
"last",
|
|
38
|
+
];
|
|
39
|
+
if (!validFuncs.includes(rawFunc)) {
|
|
40
|
+
throw new InvalidArgumentError(`Unknown aggregation function "${rawFunc}" in "${spec}". Supported functions: ${validFuncs.join(", ")}`);
|
|
41
|
+
}
|
|
42
|
+
result.push({
|
|
43
|
+
targetField,
|
|
44
|
+
func: rawFunc,
|
|
45
|
+
sourceExpr,
|
|
46
|
+
});
|
|
47
|
+
}
|
|
48
|
+
return result;
|
|
49
|
+
}
|
|
50
|
+
class SumAccumulator {
|
|
51
|
+
sum = 0;
|
|
52
|
+
add(val) {
|
|
53
|
+
if (val === null || val === undefined || val === "")
|
|
54
|
+
return;
|
|
55
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
56
|
+
if (!Number.isNaN(num)) {
|
|
57
|
+
this.sum += num;
|
|
58
|
+
}
|
|
59
|
+
}
|
|
60
|
+
result() {
|
|
61
|
+
return this.sum;
|
|
62
|
+
}
|
|
63
|
+
}
|
|
64
|
+
class AvgAccumulator {
|
|
65
|
+
sum = 0;
|
|
66
|
+
count = 0;
|
|
67
|
+
add(val) {
|
|
68
|
+
if (val === null || val === undefined || val === "")
|
|
69
|
+
return;
|
|
70
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
71
|
+
if (!Number.isNaN(num)) {
|
|
72
|
+
this.sum += num;
|
|
73
|
+
this.count++;
|
|
74
|
+
}
|
|
75
|
+
}
|
|
76
|
+
result() {
|
|
77
|
+
return this.count > 0 ? this.sum / this.count : null;
|
|
78
|
+
}
|
|
79
|
+
}
|
|
80
|
+
class MinAccumulator {
|
|
81
|
+
min = null;
|
|
82
|
+
add(val) {
|
|
83
|
+
if (val === null || val === undefined || val === "")
|
|
84
|
+
return;
|
|
85
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
86
|
+
if (!Number.isNaN(num)) {
|
|
87
|
+
if (this.min === null || num < this.min)
|
|
88
|
+
this.min = num;
|
|
89
|
+
}
|
|
90
|
+
}
|
|
91
|
+
result() {
|
|
92
|
+
return this.min;
|
|
93
|
+
}
|
|
94
|
+
}
|
|
95
|
+
class MaxAccumulator {
|
|
96
|
+
max = null;
|
|
97
|
+
add(val) {
|
|
98
|
+
if (val === null || val === undefined || val === "")
|
|
99
|
+
return;
|
|
100
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
101
|
+
if (!Number.isNaN(num)) {
|
|
102
|
+
if (this.max === null || num > this.max)
|
|
103
|
+
this.max = num;
|
|
104
|
+
}
|
|
105
|
+
}
|
|
106
|
+
result() {
|
|
107
|
+
return this.max;
|
|
108
|
+
}
|
|
109
|
+
}
|
|
110
|
+
class CountAccumulator {
|
|
111
|
+
count = 0;
|
|
112
|
+
hasSource;
|
|
113
|
+
constructor(hasSource) {
|
|
114
|
+
this.hasSource = hasSource;
|
|
115
|
+
}
|
|
116
|
+
add(val) {
|
|
117
|
+
if (!this.hasSource) {
|
|
118
|
+
this.count++;
|
|
119
|
+
}
|
|
120
|
+
else if (val !== null && val !== undefined && val !== "") {
|
|
121
|
+
this.count++;
|
|
122
|
+
}
|
|
123
|
+
}
|
|
124
|
+
result() {
|
|
125
|
+
return this.count;
|
|
126
|
+
}
|
|
127
|
+
}
|
|
128
|
+
class CountDistinctAccumulator {
|
|
129
|
+
hll = new HyperLogLog();
|
|
130
|
+
add(val) {
|
|
131
|
+
if (val !== null && val !== undefined && val !== "") {
|
|
132
|
+
this.hll.add(val);
|
|
133
|
+
}
|
|
134
|
+
}
|
|
135
|
+
result() {
|
|
136
|
+
return this.hll.count();
|
|
137
|
+
}
|
|
138
|
+
}
|
|
139
|
+
class WelfordAccumulator {
|
|
140
|
+
count = 0;
|
|
141
|
+
mean = 0;
|
|
142
|
+
M2 = 0;
|
|
143
|
+
mode;
|
|
144
|
+
constructor(mode) {
|
|
145
|
+
this.mode = mode;
|
|
146
|
+
}
|
|
147
|
+
add(val) {
|
|
148
|
+
if (val === null || val === undefined || val === "")
|
|
149
|
+
return;
|
|
150
|
+
const num = typeof val === "number" ? val : Number(val);
|
|
151
|
+
if (Number.isNaN(num))
|
|
152
|
+
return;
|
|
153
|
+
this.count++;
|
|
154
|
+
const delta = num - this.mean;
|
|
155
|
+
this.mean += delta / this.count;
|
|
156
|
+
const delta2 = num - this.mean;
|
|
157
|
+
this.M2 += delta * delta2;
|
|
158
|
+
}
|
|
159
|
+
result() {
|
|
160
|
+
if (this.count <= 1)
|
|
161
|
+
return this.count === 1 ? 0 : null;
|
|
162
|
+
const variance = this.M2 / (this.count - 1);
|
|
163
|
+
return this.mode === "stddev" ? Math.sqrt(variance) : variance;
|
|
164
|
+
}
|
|
165
|
+
}
|
|
166
|
+
class FirstAccumulator {
|
|
167
|
+
value = null;
|
|
168
|
+
found = false;
|
|
169
|
+
add(val) {
|
|
170
|
+
if (!this.found && val !== null && val !== undefined && val !== "") {
|
|
171
|
+
this.value = val;
|
|
172
|
+
this.found = true;
|
|
173
|
+
}
|
|
174
|
+
}
|
|
175
|
+
result() {
|
|
176
|
+
return this.value;
|
|
177
|
+
}
|
|
178
|
+
}
|
|
179
|
+
class LastAccumulator {
|
|
180
|
+
value = null;
|
|
181
|
+
add(val) {
|
|
182
|
+
if (val !== null && val !== undefined && val !== "") {
|
|
183
|
+
this.value = val;
|
|
184
|
+
}
|
|
185
|
+
}
|
|
186
|
+
result() {
|
|
187
|
+
return this.value;
|
|
188
|
+
}
|
|
189
|
+
}
|
|
190
|
+
function createAccumulatorFor(func, hasSource) {
|
|
191
|
+
switch (func) {
|
|
192
|
+
case "sum":
|
|
193
|
+
return new SumAccumulator();
|
|
194
|
+
case "avg":
|
|
195
|
+
case "mean":
|
|
196
|
+
return new AvgAccumulator();
|
|
197
|
+
case "min":
|
|
198
|
+
return new MinAccumulator();
|
|
199
|
+
case "max":
|
|
200
|
+
return new MaxAccumulator();
|
|
201
|
+
case "count":
|
|
202
|
+
return new CountAccumulator(hasSource);
|
|
203
|
+
case "countdistinct":
|
|
204
|
+
return new CountDistinctAccumulator();
|
|
205
|
+
case "stddev":
|
|
206
|
+
return new WelfordAccumulator("stddev");
|
|
207
|
+
case "variance":
|
|
208
|
+
return new WelfordAccumulator("variance");
|
|
209
|
+
case "first":
|
|
210
|
+
return new FirstAccumulator();
|
|
211
|
+
case "last":
|
|
212
|
+
return new LastAccumulator();
|
|
213
|
+
}
|
|
214
|
+
}
|
|
215
|
+
/**
|
|
216
|
+
* Streaming Reduce & Group-by Aggregator.
|
|
217
|
+
*/
|
|
218
|
+
export class ReduceAggregator {
|
|
219
|
+
byCols;
|
|
220
|
+
compiled;
|
|
221
|
+
globalAccs;
|
|
222
|
+
groupMap;
|
|
223
|
+
constructor(options) {
|
|
224
|
+
this.byCols = options.by && options.by.length > 0 ? options.by : [];
|
|
225
|
+
const rawSpecs = typeof options.aggregations[0] === "string"
|
|
226
|
+
? parseReduceSpecs(options.aggregations)
|
|
227
|
+
: options.aggregations;
|
|
228
|
+
this.compiled = rawSpecs.map((spec) => ({
|
|
229
|
+
targetField: spec.targetField,
|
|
230
|
+
func: spec.func,
|
|
231
|
+
evaluator: spec.sourceExpr ? compileValueExpression(spec.sourceExpr) : undefined,
|
|
232
|
+
createAccumulator: () => createAccumulatorFor(spec.func, Boolean(spec.sourceExpr)),
|
|
233
|
+
}));
|
|
234
|
+
if (this.byCols.length === 0) {
|
|
235
|
+
this.globalAccs = this.compiled.map((c) => c.createAccumulator());
|
|
236
|
+
}
|
|
237
|
+
else {
|
|
238
|
+
this.groupMap = new Map();
|
|
239
|
+
}
|
|
240
|
+
}
|
|
241
|
+
add(row) {
|
|
242
|
+
if (this.globalAccs) {
|
|
243
|
+
// Global reduction
|
|
244
|
+
for (let i = 0; i < this.compiled.length; i++) {
|
|
245
|
+
const item = this.compiled[i];
|
|
246
|
+
const val = item.evaluator ? item.evaluator(row) : null;
|
|
247
|
+
this.globalAccs[i].add(val);
|
|
248
|
+
}
|
|
249
|
+
}
|
|
250
|
+
else if (this.groupMap) {
|
|
251
|
+
// Group-by reduction
|
|
252
|
+
let key = "";
|
|
253
|
+
const groupValues = {};
|
|
254
|
+
for (let i = 0; i < this.byCols.length; i++) {
|
|
255
|
+
const col = this.byCols[i];
|
|
256
|
+
const val = row[col] ?? null;
|
|
257
|
+
groupValues[col] = val;
|
|
258
|
+
key += (i > 0 ? "\x1f" : "") + String(val ?? "");
|
|
259
|
+
}
|
|
260
|
+
let group = this.groupMap.get(key);
|
|
261
|
+
if (!group) {
|
|
262
|
+
group = {
|
|
263
|
+
groupValues,
|
|
264
|
+
accs: this.compiled.map((c) => c.createAccumulator()),
|
|
265
|
+
};
|
|
266
|
+
this.groupMap.set(key, group);
|
|
267
|
+
}
|
|
268
|
+
for (let i = 0; i < this.compiled.length; i++) {
|
|
269
|
+
const item = this.compiled[i];
|
|
270
|
+
const val = item.evaluator ? item.evaluator(row) : null;
|
|
271
|
+
group.accs[i].add(val);
|
|
272
|
+
}
|
|
273
|
+
}
|
|
274
|
+
}
|
|
275
|
+
result() {
|
|
276
|
+
if (this.globalAccs) {
|
|
277
|
+
const summaryRow = {};
|
|
278
|
+
for (let i = 0; i < this.compiled.length; i++) {
|
|
279
|
+
const item = this.compiled[i];
|
|
280
|
+
summaryRow[item.targetField] = this.globalAccs[i].result();
|
|
281
|
+
}
|
|
282
|
+
return [summaryRow];
|
|
283
|
+
}
|
|
284
|
+
if (this.groupMap) {
|
|
285
|
+
const rows = [];
|
|
286
|
+
for (const group of this.groupMap.values()) {
|
|
287
|
+
const outRow = { ...group.groupValues };
|
|
288
|
+
for (let i = 0; i < this.compiled.length; i++) {
|
|
289
|
+
const item = this.compiled[i];
|
|
290
|
+
outRow[item.targetField] = group.accs[i].result();
|
|
291
|
+
}
|
|
292
|
+
rows.push(outRow);
|
|
293
|
+
}
|
|
294
|
+
return rows;
|
|
295
|
+
}
|
|
296
|
+
return [];
|
|
297
|
+
}
|
|
298
|
+
}
|
|
299
|
+
/**
|
|
300
|
+
* Creates a high-performance streaming transform that reduces/aggregates rows.
|
|
301
|
+
*/
|
|
302
|
+
export function reduceRows(options) {
|
|
303
|
+
return async function* (stream) {
|
|
304
|
+
const aggregator = new ReduceAggregator(options);
|
|
305
|
+
for await (const batch of stream) {
|
|
306
|
+
const len = batch.rows.length;
|
|
307
|
+
for (let i = 0; i < len; i++) {
|
|
308
|
+
aggregator.add(batch.rows[i]);
|
|
309
|
+
}
|
|
310
|
+
}
|
|
311
|
+
const finalRows = aggregator.result();
|
|
312
|
+
yield {
|
|
313
|
+
rows: finalRows,
|
|
314
|
+
offset: 0,
|
|
315
|
+
};
|
|
316
|
+
};
|
|
317
|
+
}
|
|
318
|
+
//# sourceMappingURL=reduce.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"reduce.js","sourceRoot":"","sources":["../../src/analytics/reduce.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,mBAAmB,CAAC;AAEzD,OAAO,EAAE,sBAAsB,EAAE,MAAM,6BAA6B,CAAC;AACrE,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAqBzC;;;GAGG;AACH,MAAM,UAAU,gBAAgB,CAAC,KAAe;IAC9C,MAAM,MAAM,GAAsB,EAAE,CAAC;IAErC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,OAAO;YAAE,SAAS;QAEvB,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,KAAK,KAAK,CAAC,CAAC,EAAE,CAAC;YACjB,MAAM,IAAI,oBAAoB,CAC5B,iCAAiC,IAAI,wDAAwD,CAC9F,CAAC;QACJ,CAAC;QAED,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;QACnD,MAAM,GAAG,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAE5C,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,2BAA2B,CAAC,CAAC;QACrD,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,MAAM,IAAI,oBAAoB,CAC5B,yCAAyC,IAAI,oEAAoE,CAClH,CAAC;QACJ,CAAC;QAED,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,WAAW,EAAE,CAAC;QACxC,MAAM,UAAU,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,IAAI,SAAS,CAAC;QAEjD,MAAM,UAAU,GAAG;YACjB,KAAK;YACL,KAAK;YACL,MAAM;YACN,KAAK;YACL,KAAK;YACL,OAAO;YACP,eAAe;YACf,QAAQ;YACR,UAAU;YACV,OAAO;YACP,MAAM;SACP,CAAC;QAEF,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;YAClC,MAAM,IAAI,oBAAoB,CAC5B,iCAAiC,OAAO,SAAS,IAAI,2BAA2B,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACxG,CAAC;QACJ,CAAC;QAED,MAAM,CAAC,IAAI,CAAC;YACV,WAAW;YACX,IAAI,EAAE,OAA8B;YACpC,UAAU;SACX,CAAC,CAAC;IACL,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAOD,MAAM,cAAc;IACV,GAAG,GAAG,CAAC,CAAC;IAChB,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;YAAE,OAAO;QAC5D,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC;QAClB,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,GAAG,CAAC;IAClB,CAAC;CACF;AAED,MAAM,cAAc;IACV,GAAG,GAAG,CAAC,CAAC;IACR,KAAK,GAAG,CAAC,CAAC;IAClB,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;YAAE,OAAO;QAC5D,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC;YAChB,IAAI,CAAC,KAAK,EAAE,CAAC;QACf,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC;IACvD,CAAC;CACF;AAED,MAAM,cAAc;IACV,GAAG,GAAkB,IAAI,CAAC;IAClC,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;YAAE,OAAO;QAC5D,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,IAAI,GAAG,GAAG,IAAI,CAAC,GAAG;gBAAE,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;QAC1D,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,GAAG,CAAC;IAClB,CAAC;CACF;AAED,MAAM,cAAc;IACV,GAAG,GAAkB,IAAI,CAAC;IAClC,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;YAAE,OAAO;QAC5D,MAAM,GAAG,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;YACvB,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,IAAI,GAAG,GAAG,IAAI,CAAC,GAAG;gBAAE,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;QAC1D,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,GAAG,CAAC;IAClB,CAAC;CACF;AAED,MAAM,gBAAgB;IACZ,KAAK,GAAG,CAAC,CAAC;IACV,SAAS,CAAU;IAC3B,YAAY,SAAkB;QAC5B,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;IAC7B,CAAC;IACD,GAAG,CAAC,GAAY;QACd,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;YACpB,IAAI,CAAC,KAAK,EAAE,CAAC;QACf,CAAC;aAAM,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;YAC3D,IAAI,CAAC,KAAK,EAAE,CAAC;QACf,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;CACF;AAED,MAAM,wBAAwB;IACpB,GAAG,GAAG,IAAI,WAAW,EAAE,CAAC;IAChC,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;YACpD,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACpB,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC;IAC1B,CAAC;CACF;AAED,MAAM,kBAAkB;IACd,KAAK,GAAG,CAAC,CAAC;IACV,IAAI,GAAG,CAAC,CAAC;IACT,EAAE,GAAG,CAAC,CAAC;IACP,IAAI,CAAwB;IAEpC,YAAY,IAA2B;QACrC,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC;IACnB,CAAC;IAED,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;YAAE,OAAO;QAC5D,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;YAAE,OAAO;QAE9B,IAAI,CAAC,KAAK,EAAE,CAAC;QACb,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,MAAM;QACJ,IAAI,IAAI,CAAC,KAAK,IAAI,CAAC;YAAE,OAAO,IAAI,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QACxD,MAAM,QAAQ,GAAG,IAAI,CAAC,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QAC5C,OAAO,IAAI,CAAC,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;IACjE,CAAC;CACF;AAED,MAAM,gBAAgB;IACZ,KAAK,GAAY,IAAI,CAAC;IACtB,KAAK,GAAG,KAAK,CAAC;IACtB,GAAG,CAAC,GAAY;QACd,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;YACnE,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC;YACjB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC;QACpB,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;CACF;AAED,MAAM,eAAe;IACX,KAAK,GAAY,IAAI,CAAC;IAC9B,GAAG,CAAC,GAAY;QACd,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;YACpD,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC;QACnB,CAAC;IACH,CAAC;IACD,MAAM;QACJ,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;CACF;AAED,SAAS,oBAAoB,CAAC,IAAyB,EAAE,SAAkB;IACzE,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,KAAK;YACR,OAAO,IAAI,cAAc,EAAE,CAAC;QAC9B,KAAK,KAAK,CAAC;QACX,KAAK,MAAM;YACT,OAAO,IAAI,cAAc,EAAE,CAAC;QAC9B,KAAK,KAAK;YACR,OAAO,IAAI,cAAc,EAAE,CAAC;QAC9B,KAAK,KAAK;YACR,OAAO,IAAI,cAAc,EAAE,CAAC;QAC9B,KAAK,OAAO;YACV,OAAO,IAAI,gBAAgB,CAAC,SAAS,CAAC,CAAC;QACzC,KAAK,eAAe;YAClB,OAAO,IAAI,wBAAwB,EAAE,CAAC;QACxC,KAAK,QAAQ;YACX,OAAO,IAAI,kBAAkB,CAAC,QAAQ,CAAC,CAAC;QAC1C,KAAK,UAAU;YACb,OAAO,IAAI,kBAAkB,CAAC,UAAU,CAAC,CAAC;QAC5C,KAAK,OAAO;YACV,OAAO,IAAI,gBAAgB,EAAE,CAAC;QAChC,KAAK,MAAM;YACT,OAAO,IAAI,eAAe,EAAE,CAAC;IACjC,CAAC;AACH,CAAC;AAcD;;GAEG;AACH,MAAM,OAAO,gBAAgB;IACnB,MAAM,CAAW;IACjB,QAAQ,CAAwB;IAChC,UAAU,CAAiB;IAC3B,QAAQ,CAA8E;IAE9F,YAAY,OAAsB;QAChC,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,EAAE,IAAI,OAAO,CAAC,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAEpE,MAAM,QAAQ,GACZ,OAAO,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,QAAQ;YACzC,CAAC,CAAC,gBAAgB,CAAC,OAAO,CAAC,YAAwB,CAAC;YACpD,CAAC,CAAE,OAAO,CAAC,YAAkC,CAAC;QAElD,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACtC,WAAW,EAAE,IAAI,CAAC,WAAW;YAC7B,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,SAAS,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,sBAAsB,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS;YAChF,iBAAiB,EAAE,GAAG,EAAE,CAAC,oBAAoB,CAAC,IAAI,CAAC,IAAI,EAAE,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;SACnF,CAAC,CAAC,CAAC;QAEJ,IAAI,IAAI,CAAC,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC7B,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,iBAAiB,EAAE,CAAC,CAAC;QACpE,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,QAAQ,GAAG,IAAI,GAAG,EAAE,CAAC;QAC5B,CAAC;IACH,CAAC;IAED,GAAG,CAAC,GAAQ;QACV,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACpB,mBAAmB;YACnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAC;gBAC/B,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;gBACxD,IAAI,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAC/B,CAAC;QACH,CAAC;aAAM,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YACzB,qBAAqB;YACrB,IAAI,GAAG,GAAG,EAAE,CAAC;YACb,MAAM,WAAW,GAA4B,EAAE,CAAC;YAEhD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC5C,MAAM,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC;gBAC5B,MAAM,GAAG,GAAG,GAAG,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC;gBAC7B,WAAW,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC;gBACvB,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC;YACnD,CAAC;YAED,IAAI,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACnC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,KAAK,GAAG;oBACN,WAAW;oBACX,IAAI,EAAE,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,iBAAiB,EAAE,CAAC;iBACtD,CAAC;gBACF,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;YAChC,CAAC;YAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAC;gBAC/B,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;gBACxD,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAC1B,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM;QACJ,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACpB,MAAM,UAAU,GAAQ,EAAE,CAAC;YAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAC;gBAC/B,UAAU,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,IAAI,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC,MAAM,EAAE,CAAC;YAC9D,CAAC;YACD,OAAO,CAAC,UAAU,CAAC,CAAC;QACtB,CAAC;QAED,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,MAAM,IAAI,GAAU,EAAE,CAAC;YACvB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,EAAE,CAAC;gBAC3C,MAAM,MAAM,GAAQ,EAAE,GAAG,KAAK,CAAC,WAAW,EAAE,CAAC;gBAC7C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAC;oBAC/B,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,MAAM,EAAE,CAAC;gBACrD,CAAC;gBACD,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpB,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;QAED,OAAO,EAAE,CAAC;IACZ,CAAC;CACF;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,OAAsB;IAC/C,OAAO,KAAK,SAAS,CAAC,EAAE,MAAkB;QACxC,MAAM,UAAU,GAAG,IAAI,gBAAgB,CAAC,OAAO,CAAC,CAAC;QAEjD,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YACjC,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC;YAC9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC7B,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,CAAC;YACjC,CAAC;QACH,CAAC;QAED,MAAM,SAAS,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;QAEtC,MAAM;YACJ,IAAI,EAAE,SAAS;YACf,MAAM,EAAE,CAAC;SACV,CAAC;IACJ,CAAC,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,31 @@
|
|
|
1
|
+
import type { Aggregator, ColumnType, Row, SemanticType } from "../core/types.js";
|
|
2
|
+
export interface ColumnInferenceResult {
|
|
3
|
+
name: string;
|
|
4
|
+
type: ColumnType;
|
|
5
|
+
nullable: boolean;
|
|
6
|
+
confidence: number;
|
|
7
|
+
semantic?: SemanticType;
|
|
8
|
+
semanticConfidence?: number;
|
|
9
|
+
typeBreakdown: Record<string, number>;
|
|
10
|
+
sampleCount: number;
|
|
11
|
+
nullCount: number;
|
|
12
|
+
}
|
|
13
|
+
export interface InferredSchemaResult {
|
|
14
|
+
totalRowsScanned: number;
|
|
15
|
+
columns: ColumnInferenceResult[];
|
|
16
|
+
}
|
|
17
|
+
/**
|
|
18
|
+
* Tests primitive type for an individual value.
|
|
19
|
+
*/
|
|
20
|
+
export declare function classifyPrimitiveType(value: unknown): ColumnType;
|
|
21
|
+
export declare class SchemaInferenceAggregator implements Aggregator<InferredSchemaResult> {
|
|
22
|
+
private totalRows;
|
|
23
|
+
private maxSample;
|
|
24
|
+
private columnStats;
|
|
25
|
+
constructor(options?: {
|
|
26
|
+
sample?: number;
|
|
27
|
+
});
|
|
28
|
+
add(row: Row): void;
|
|
29
|
+
result(): InferredSchemaResult;
|
|
30
|
+
}
|
|
31
|
+
//# sourceMappingURL=schema-inference.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"schema-inference.d.ts","sourceRoot":"","sources":["../../src/analytics/schema-inference.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,GAAG,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAGlF,MAAM,WAAW,qBAAqB;IACpC,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,UAAU,CAAC;IACjB,QAAQ,EAAE,OAAO,CAAC;IAClB,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,CAAC,EAAE,YAAY,CAAC;IACxB,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACtC,WAAW,EAAE,MAAM,CAAC;IACpB,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,oBAAoB;IACnC,gBAAgB,EAAE,MAAM,CAAC;IACzB,OAAO,EAAE,qBAAqB,EAAE,CAAC;CAClC;AAKD;;GAEG;AACH,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,OAAO,GAAG,UAAU,CAuChE;AAED,qBAAa,yBAA0B,YAAW,UAAU,CAAC,oBAAoB,CAAC;IAChF,OAAO,CAAC,SAAS,CAAK;IACtB,OAAO,CAAC,SAAS,CAAS;IAC1B,OAAO,CAAC,WAAW,CAQf;gBAEQ,OAAO,CAAC,EAAE;QAAE,MAAM,CAAC,EAAE,MAAM,CAAA;KAAE;IAIzC,GAAG,CAAC,GAAG,EAAE,GAAG,GAAG,IAAI;IAuCnB,MAAM,IAAI,oBAAoB;CA6F/B"}
|
|
@@ -0,0 +1,166 @@
|
|
|
1
|
+
import { detectSemanticType } from "./semantic-types.js";
|
|
2
|
+
const DATE_REGEX = /^\d{4}-\d{2}-\d{2}$/;
|
|
3
|
+
const DATETIME_REGEX = /^\d{4}-\d{2}-\d{2}[T\s]\d{2}:\d{2}:\d{2}(?:\.\d+)?(?:Z|[+-]\d{2}:?\d{2})?$/;
|
|
4
|
+
/**
|
|
5
|
+
* Tests primitive type for an individual value.
|
|
6
|
+
*/
|
|
7
|
+
export function classifyPrimitiveType(value) {
|
|
8
|
+
if (value === null || value === undefined || value === "") {
|
|
9
|
+
return "null";
|
|
10
|
+
}
|
|
11
|
+
if (typeof value === "boolean")
|
|
12
|
+
return "boolean";
|
|
13
|
+
if (typeof value === "number") {
|
|
14
|
+
return Number.isInteger(value) ? "integer" : "number";
|
|
15
|
+
}
|
|
16
|
+
const str = String(value).trim();
|
|
17
|
+
// Boolean strings
|
|
18
|
+
if (["true", "false"].includes(str.toLowerCase())) {
|
|
19
|
+
return "boolean";
|
|
20
|
+
}
|
|
21
|
+
// Integer regex
|
|
22
|
+
if (/^-?\d+$/.test(str)) {
|
|
23
|
+
return "integer";
|
|
24
|
+
}
|
|
25
|
+
// Floating point regex
|
|
26
|
+
if (/^-?\d+\.\d+$/.test(str) || /^-?\d+(?:\.\d+)?e[+-]?\d+$/i.test(str)) {
|
|
27
|
+
return "number";
|
|
28
|
+
}
|
|
29
|
+
// Date / Datetime
|
|
30
|
+
if (DATE_REGEX.test(str)) {
|
|
31
|
+
const d = new Date(str);
|
|
32
|
+
if (!Number.isNaN(d.getTime()))
|
|
33
|
+
return "date";
|
|
34
|
+
}
|
|
35
|
+
if (DATETIME_REGEX.test(str)) {
|
|
36
|
+
const d = new Date(str);
|
|
37
|
+
if (!Number.isNaN(d.getTime()))
|
|
38
|
+
return "datetime";
|
|
39
|
+
}
|
|
40
|
+
return "string";
|
|
41
|
+
}
|
|
42
|
+
export class SchemaInferenceAggregator {
|
|
43
|
+
totalRows = 0;
|
|
44
|
+
maxSample;
|
|
45
|
+
columnStats = new Map();
|
|
46
|
+
constructor(options) {
|
|
47
|
+
this.maxSample = options?.sample ?? Number.POSITIVE_INFINITY;
|
|
48
|
+
}
|
|
49
|
+
add(row) {
|
|
50
|
+
if (this.totalRows >= this.maxSample)
|
|
51
|
+
return;
|
|
52
|
+
this.totalRows++;
|
|
53
|
+
for (const [col, val] of Object.entries(row)) {
|
|
54
|
+
if (!this.columnStats.has(col)) {
|
|
55
|
+
this.columnStats.set(col, {
|
|
56
|
+
counts: {
|
|
57
|
+
string: 0,
|
|
58
|
+
integer: 0,
|
|
59
|
+
number: 0,
|
|
60
|
+
boolean: 0,
|
|
61
|
+
date: 0,
|
|
62
|
+
datetime: 0,
|
|
63
|
+
null: 0,
|
|
64
|
+
mixed: 0,
|
|
65
|
+
},
|
|
66
|
+
semanticCounts: {},
|
|
67
|
+
nullCount: 0,
|
|
68
|
+
totalCount: 0,
|
|
69
|
+
});
|
|
70
|
+
}
|
|
71
|
+
const stat = this.columnStats.get(col);
|
|
72
|
+
stat.totalCount++;
|
|
73
|
+
const pType = classifyPrimitiveType(val);
|
|
74
|
+
stat.counts[pType]++;
|
|
75
|
+
if (pType === "null") {
|
|
76
|
+
stat.nullCount++;
|
|
77
|
+
}
|
|
78
|
+
else {
|
|
79
|
+
const sem = detectSemanticType(val);
|
|
80
|
+
if (sem) {
|
|
81
|
+
stat.semanticCounts[sem] = (stat.semanticCounts[sem] || 0) + 1;
|
|
82
|
+
}
|
|
83
|
+
}
|
|
84
|
+
}
|
|
85
|
+
}
|
|
86
|
+
result() {
|
|
87
|
+
const columns = [];
|
|
88
|
+
for (const [col, stat] of this.columnStats.entries()) {
|
|
89
|
+
const nonNullCount = stat.totalCount - stat.nullCount;
|
|
90
|
+
const nullable = stat.nullCount > 0;
|
|
91
|
+
if (nonNullCount === 0) {
|
|
92
|
+
columns.push({
|
|
93
|
+
name: col,
|
|
94
|
+
type: "null",
|
|
95
|
+
nullable: true,
|
|
96
|
+
confidence: 100,
|
|
97
|
+
typeBreakdown: { null: 100 },
|
|
98
|
+
sampleCount: stat.totalCount,
|
|
99
|
+
nullCount: stat.nullCount,
|
|
100
|
+
});
|
|
101
|
+
continue;
|
|
102
|
+
}
|
|
103
|
+
// Calculate confidence percentages
|
|
104
|
+
const breakdown = {};
|
|
105
|
+
let dominantType = "string";
|
|
106
|
+
let maxPercentage = 0;
|
|
107
|
+
const candidates = [
|
|
108
|
+
"integer",
|
|
109
|
+
"number",
|
|
110
|
+
"boolean",
|
|
111
|
+
"datetime",
|
|
112
|
+
"date",
|
|
113
|
+
"string",
|
|
114
|
+
];
|
|
115
|
+
for (const t of candidates) {
|
|
116
|
+
const pct = (stat.counts[t] / nonNullCount) * 100;
|
|
117
|
+
if (pct > 0) {
|
|
118
|
+
breakdown[t] = Math.round(pct * 10) / 10;
|
|
119
|
+
}
|
|
120
|
+
if (pct > maxPercentage) {
|
|
121
|
+
maxPercentage = pct;
|
|
122
|
+
dominantType = t;
|
|
123
|
+
}
|
|
124
|
+
}
|
|
125
|
+
// If integer + number together make 100%, consider number
|
|
126
|
+
if (dominantType === "integer" &&
|
|
127
|
+
stat.counts["number"] > 0 &&
|
|
128
|
+
(stat.counts["integer"] + stat.counts["number"]) / nonNullCount >= 0.95) {
|
|
129
|
+
dominantType = "number";
|
|
130
|
+
maxPercentage = ((stat.counts["integer"] + stat.counts["number"]) / nonNullCount) * 100;
|
|
131
|
+
}
|
|
132
|
+
let inferredType = dominantType;
|
|
133
|
+
let confidence = Math.round(maxPercentage * 10) / 10;
|
|
134
|
+
if (maxPercentage < 90 && Object.keys(breakdown).length > 1) {
|
|
135
|
+
inferredType = "mixed";
|
|
136
|
+
}
|
|
137
|
+
// Semantic type detection
|
|
138
|
+
let semantic;
|
|
139
|
+
let semanticConfidence;
|
|
140
|
+
for (const [sem, count] of Object.entries(stat.semanticCounts)) {
|
|
141
|
+
const semPct = (count / nonNullCount) * 100;
|
|
142
|
+
if (semPct >= 80) {
|
|
143
|
+
semantic = sem;
|
|
144
|
+
semanticConfidence = Math.round(semPct * 10) / 10;
|
|
145
|
+
break;
|
|
146
|
+
}
|
|
147
|
+
}
|
|
148
|
+
columns.push({
|
|
149
|
+
name: col,
|
|
150
|
+
type: inferredType,
|
|
151
|
+
nullable,
|
|
152
|
+
confidence,
|
|
153
|
+
semantic,
|
|
154
|
+
semanticConfidence,
|
|
155
|
+
typeBreakdown: breakdown,
|
|
156
|
+
sampleCount: stat.totalCount,
|
|
157
|
+
nullCount: stat.nullCount,
|
|
158
|
+
});
|
|
159
|
+
}
|
|
160
|
+
return {
|
|
161
|
+
totalRowsScanned: this.totalRows,
|
|
162
|
+
columns,
|
|
163
|
+
};
|
|
164
|
+
}
|
|
165
|
+
}
|
|
166
|
+
//# sourceMappingURL=schema-inference.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"schema-inference.js","sourceRoot":"","sources":["../../src/analytics/schema-inference.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAC;AAmBzD,MAAM,UAAU,GAAG,qBAAqB,CAAC;AACzC,MAAM,cAAc,GAAG,4EAA4E,CAAC;AAEpG;;GAEG;AACH,MAAM,UAAU,qBAAqB,CAAC,KAAc;IAClD,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,EAAE,EAAE,CAAC;QAC1D,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,IAAI,OAAO,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IACjD,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC9B,OAAO,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;IACxD,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;IAEjC,kBAAkB;IAClB,IAAI,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,EAAE,CAAC;QAClD,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,gBAAgB;IAChB,IAAI,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACxB,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,uBAAuB;IACvB,IAAI,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,6BAA6B,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACxE,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,kBAAkB;IAClB,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACzB,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;QACxB,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;YAAE,OAAO,MAAM,CAAC;IAChD,CAAC;IAED,IAAI,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QAC7B,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;QACxB,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC;YAAE,OAAO,UAAU,CAAC;IACpD,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,MAAM,OAAO,yBAAyB;IAC5B,SAAS,GAAG,CAAC,CAAC;IACd,SAAS,CAAS;IAClB,WAAW,GAAG,IAAI,GAAG,EAQ1B,CAAC;IAEJ,YAAY,OAA6B;QACvC,IAAI,CAAC,SAAS,GAAG,OAAO,EAAE,MAAM,IAAI,MAAM,CAAC,iBAAiB,CAAC;IAC/D,CAAC;IAED,GAAG,CAAC,GAAQ;QACV,IAAI,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,SAAS;YAAE,OAAO;QAC7C,IAAI,CAAC,SAAS,EAAE,CAAC;QAEjB,KAAK,MAAM,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;YAC7C,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC/B,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,EAAE;oBACxB,MAAM,EAAE;wBACN,MAAM,EAAE,CAAC;wBACT,OAAO,EAAE,CAAC;wBACV,MAAM,EAAE,CAAC;wBACT,OAAO,EAAE,CAAC;wBACV,IAAI,EAAE,CAAC;wBACP,QAAQ,EAAE,CAAC;wBACX,IAAI,EAAE,CAAC;wBACP,KAAK,EAAE,CAAC;qBACT;oBACD,cAAc,EAAE,EAAE;oBAClB,SAAS,EAAE,CAAC;oBACZ,UAAU,EAAE,CAAC;iBACd,CAAC,CAAC;YACL,CAAC;YAED,MAAM,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC;YACxC,IAAI,CAAC,UAAU,EAAE,CAAC;YAElB,MAAM,KAAK,GAAG,qBAAqB,CAAC,GAAG,CAAC,CAAC;YACzC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC;YACrB,IAAI,KAAK,KAAK,MAAM,EAAE,CAAC;gBACrB,IAAI,CAAC,SAAS,EAAE,CAAC;YACnB,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC;gBACpC,IAAI,GAAG,EAAE,CAAC;oBACR,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;gBACjE,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM;QACJ,MAAM,OAAO,GAA4B,EAAE,CAAC;QAE5C,KAAK,MAAM,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,WAAW,CAAC,OAAO,EAAE,EAAE,CAAC;YACrD,MAAM,YAAY,GAAG,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,SAAS,CAAC;YACtD,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,GAAG,CAAC,CAAC;YAEpC,IAAI,YAAY,KAAK,CAAC,EAAE,CAAC;gBACvB,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,GAAG;oBACT,IAAI,EAAE,MAAM;oBACZ,QAAQ,EAAE,IAAI;oBACd,UAAU,EAAE,GAAG;oBACf,aAAa,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE;oBAC5B,WAAW,EAAE,IAAI,CAAC,UAAU;oBAC5B,SAAS,EAAE,IAAI,CAAC,SAAS;iBAC1B,CAAC,CAAC;gBACH,SAAS;YACX,CAAC;YAED,mCAAmC;YACnC,MAAM,SAAS,GAA2B,EAAE,CAAC;YAC7C,IAAI,YAAY,GAAe,QAAQ,CAAC;YACxC,IAAI,aAAa,GAAG,CAAC,CAAC;YAEtB,MAAM,UAAU,GAAiB;gBAC/B,SAAS;gBACT,QAAQ;gBACR,SAAS;gBACT,UAAU;gBACV,MAAM;gBACN,QAAQ;aACT,CAAC;YAEF,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;gBAC3B,MAAM,GAAG,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,YAAY,CAAC,GAAG,GAAG,CAAC;gBAClD,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;oBACZ,SAAS,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,EAAE,CAAC,GAAG,EAAE,CAAC;gBAC3C,CAAC;gBACD,IAAI,GAAG,GAAG,aAAa,EAAE,CAAC;oBACxB,aAAa,GAAG,GAAG,CAAC;oBACpB,YAAY,GAAG,CAAC,CAAC;gBACnB,CAAC;YACH,CAAC;YAED,0DAA0D;YAC1D,IACE,YAAY,KAAK,SAAS;gBAC1B,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC;gBACzB,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,GAAG,YAAY,IAAI,IAAI,EACvE,CAAC;gBACD,YAAY,GAAG,QAAQ,CAAC;gBACxB,aAAa,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,GAAG,YAAY,CAAC,GAAG,GAAG,CAAC;YAC1F,CAAC;YAED,IAAI,YAAY,GAAe,YAAY,CAAC;YAC5C,IAAI,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,aAAa,GAAG,EAAE,CAAC,GAAG,EAAE,CAAC;YAErD,IAAI,aAAa,GAAG,EAAE,IAAI,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5D,YAAY,GAAG,OAAO,CAAC;YACzB,CAAC;YAED,0BAA0B;YAC1B,IAAI,QAAkC,CAAC;YACvC,IAAI,kBAAsC,CAAC;YAE3C,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,cAAc,CAAC,EAAE,CAAC;gBAC/D,MAAM,MAAM,GAAG,CAAC,KAAK,GAAG,YAAY,CAAC,GAAG,GAAG,CAAC;gBAC5C,IAAI,MAAM,IAAI,EAAE,EAAE,CAAC;oBACjB,QAAQ,GAAG,GAAmB,CAAC;oBAC/B,kBAAkB,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,EAAE,CAAC,GAAG,EAAE,CAAC;oBAClD,MAAM;gBACR,CAAC;YACH,CAAC;YAED,OAAO,CAAC,IAAI,CAAC;gBACX,IAAI,EAAE,GAAG;gBACT,IAAI,EAAE,YAAY;gBAClB,QAAQ;gBACR,UAAU;gBACV,QAAQ;gBACR,kBAAkB;gBAClB,aAAa,EAAE,SAAS;gBACxB,WAAW,EAAE,IAAI,CAAC,UAAU;gBAC5B,SAAS,EAAE,IAAI,CAAC,SAAS;aAC1B,CAAC,CAAC;QACL,CAAC;QAED,OAAO;YACL,gBAAgB,EAAE,IAAI,CAAC,SAAS;YAChC,OAAO;SACR,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"semantic-types.d.ts","sourceRoot":"","sources":["../../src/analytics/semantic-types.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAiBrD;;GAEG;AACH,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,OAAO,GAAG,YAAY,GAAG,IAAI,CAetE"}
|
|
@@ -0,0 +1,40 @@
|
|
|
1
|
+
const EMAIL_REGEX = /^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$/;
|
|
2
|
+
const URL_REGEX = /^https?:\/\/[^\s/$.?#].[^\s]*$/i;
|
|
3
|
+
const UUID_REGEX = /^[0-9a-f]{8}-[0-9a-f]{4}-[1-5][0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i;
|
|
4
|
+
const IPV4_REGEX = /^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$/;
|
|
5
|
+
const IPV6_REGEX = /^(?:[A-F0-9]{1,4}:){7}[A-F0-9]{1,4}$/i;
|
|
6
|
+
const PHONE_REGEX = /^\+?[0-9\s\-().]{7,20}$/;
|
|
7
|
+
const ISO_COUNTRY_CODES = new Set([
|
|
8
|
+
"US", "GB", "DE", "FR", "TR", "CA", "AU", "JP", "CN", "IN", "BR", "IT", "ES", "NL", "SE", "CH", "PL", "MX", "KR", "RU", "ZA", "EG", "SA", "AE", "SG", "NZ", "IE", "NO", "DK", "FI", "PT", "GR", "AT", "BE", "IL", "AR", "CL", "CO", "PE", "TH", "MY", "ID", "PH", "VN", "PK", "NG", "KE", "ZA"
|
|
9
|
+
]);
|
|
10
|
+
const ISO_CURRENCIES = new Set([
|
|
11
|
+
"USD", "EUR", "GBP", "TRY", "JPY", "CAD", "AUD", "CHF", "CNY", "INR", "BRL", "RUB", "KRW", "SEK", "NOK", "MXN", "SGD", "NZD", "HKD", "PLN", "ZAR", "AED", "SAR", "THB", "IDR"
|
|
12
|
+
]);
|
|
13
|
+
/**
|
|
14
|
+
* Tests if a string value matches a specific semantic pattern.
|
|
15
|
+
*/
|
|
16
|
+
export function detectSemanticType(value) {
|
|
17
|
+
if (value === null || value === undefined)
|
|
18
|
+
return null;
|
|
19
|
+
const str = String(value).trim();
|
|
20
|
+
if (str.length === 0)
|
|
21
|
+
return null;
|
|
22
|
+
if (UUID_REGEX.test(str))
|
|
23
|
+
return "uuid";
|
|
24
|
+
if (EMAIL_REGEX.test(str))
|
|
25
|
+
return "email";
|
|
26
|
+
if (URL_REGEX.test(str))
|
|
27
|
+
return "url";
|
|
28
|
+
if (IPV4_REGEX.test(str))
|
|
29
|
+
return "ipv4";
|
|
30
|
+
if (IPV6_REGEX.test(str))
|
|
31
|
+
return "ipv6";
|
|
32
|
+
if (str.length >= 2 && str.length <= 3 && ISO_COUNTRY_CODES.has(str.toUpperCase()))
|
|
33
|
+
return "country-code";
|
|
34
|
+
if (str.length === 3 && ISO_CURRENCIES.has(str.toUpperCase()))
|
|
35
|
+
return "currency";
|
|
36
|
+
if (PHONE_REGEX.test(str) && /\d{4,}/.test(str) && (str.startsWith("+") || str.includes("-") || str.includes("(")))
|
|
37
|
+
return "phone";
|
|
38
|
+
return null;
|
|
39
|
+
}
|
|
40
|
+
//# sourceMappingURL=semantic-types.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"semantic-types.js","sourceRoot":"","sources":["../../src/analytics/semantic-types.ts"],"names":[],"mappings":"AAEA,MAAM,WAAW,GAAG,kDAAkD,CAAC;AACvE,MAAM,SAAS,GAAG,iCAAiC,CAAC;AACpD,MAAM,UAAU,GAAG,4EAA4E,CAAC;AAChG,MAAM,UAAU,GAAG,6FAA6F,CAAC;AACjH,MAAM,UAAU,GAAG,uCAAuC,CAAC;AAC3D,MAAM,WAAW,GAAG,yBAAyB,CAAC;AAE9C,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAC;IAChC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;CAC/R,CAAC,CAAC;AAEH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC;IAC7B,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK;CAC9K,CAAC,CAAC;AAEH;;GAEG;AACH,MAAM,UAAU,kBAAkB,CAAC,KAAc;IAC/C,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,IAAI,CAAC;IACvD,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;IACjC,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAElC,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,MAAM,CAAC;IACxC,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IAC1C,IAAI,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IACtC,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,MAAM,CAAC;IACxC,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,MAAM,CAAC;IACxC,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,IAAI,iBAAiB,CAAC,GAAG,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC;QAAE,OAAO,cAAc,CAAC;IAC1G,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,IAAI,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC;QAAE,OAAO,UAAU,CAAC;IACjF,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;QAAE,OAAO,OAAO,CAAC;IAEnI,OAAO,IAAI,CAAC;AACd,CAAC"}
|