rowpipe 2.2.0 → 2.10.1
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 +35 -4
- package/dist/analytics/abtest.d.ts +39 -0
- package/dist/analytics/abtest.d.ts.map +1 -0
- package/dist/analytics/abtest.js +266 -0
- package/dist/analytics/abtest.js.map +1 -0
- package/dist/analytics/cluster.d.ts +25 -0
- package/dist/analytics/cluster.d.ts.map +1 -0
- package/dist/analytics/cluster.js +212 -0
- package/dist/analytics/cluster.js.map +1 -0
- package/dist/analytics/cohort.d.ts +22 -0
- package/dist/analytics/cohort.d.ts.map +1 -0
- package/dist/analytics/cohort.js +170 -0
- package/dist/analytics/cohort.js.map +1 -0
- package/dist/analytics/correlation.d.ts +15 -0
- package/dist/analytics/correlation.d.ts.map +1 -0
- package/dist/analytics/correlation.js +116 -0
- package/dist/analytics/correlation.js.map +1 -0
- package/dist/analytics/crosstab.d.ts +27 -0
- package/dist/analytics/crosstab.d.ts.map +1 -0
- package/dist/analytics/crosstab.js +90 -0
- package/dist/analytics/crosstab.js.map +1 -0
- package/dist/analytics/entropy.d.ts +23 -0
- package/dist/analytics/entropy.d.ts.map +1 -0
- package/dist/analytics/entropy.js +180 -0
- package/dist/analytics/entropy.js.map +1 -0
- package/dist/analytics/funnel.d.ts +29 -0
- package/dist/analytics/funnel.d.ts.map +1 -0
- package/dist/analytics/funnel.js +170 -0
- package/dist/analytics/funnel.js.map +1 -0
- package/dist/analytics/ngrams.d.ts +25 -0
- package/dist/analytics/ngrams.d.ts.map +1 -0
- package/dist/analytics/ngrams.js +110 -0
- package/dist/analytics/ngrams.js.map +1 -0
- package/dist/analytics/outliers.d.ts +25 -0
- package/dist/analytics/outliers.d.ts.map +1 -0
- package/dist/analytics/outliers.js +114 -0
- package/dist/analytics/outliers.js.map +1 -0
- package/dist/analytics/pareto.d.ts +36 -0
- package/dist/analytics/pareto.d.ts.map +1 -0
- package/dist/analytics/pareto.js +140 -0
- package/dist/analytics/pareto.js.map +1 -0
- package/dist/analytics/quantiles.d.ts +18 -0
- package/dist/analytics/quantiles.d.ts.map +1 -0
- package/dist/analytics/quantiles.js +62 -0
- package/dist/analytics/quantiles.js.map +1 -0
- package/dist/analytics/regression.d.ts +18 -0
- package/dist/analytics/regression.d.ts.map +1 -0
- package/dist/analytics/regression.js +79 -0
- package/dist/analytics/regression.js.map +1 -0
- package/dist/analytics/rfm.d.ts +41 -0
- package/dist/analytics/rfm.d.ts.map +1 -0
- package/dist/analytics/rfm.js +221 -0
- package/dist/analytics/rfm.js.map +1 -0
- package/dist/analytics/stats.d.ts.map +1 -1
- package/dist/analytics/stats.js +65 -15
- package/dist/analytics/stats.js.map +1 -1
- package/dist/analytics/technical.d.ts +19 -0
- package/dist/analytics/technical.d.ts.map +1 -0
- package/dist/analytics/technical.js +365 -0
- package/dist/analytics/technical.js.map +1 -0
- package/dist/cli/commands/abtest.d.ts +10 -0
- package/dist/cli/commands/abtest.d.ts.map +1 -0
- package/dist/cli/commands/abtest.js +68 -0
- package/dist/cli/commands/abtest.js.map +1 -0
- package/dist/cli/commands/cluster.d.ts +9 -0
- package/dist/cli/commands/cluster.d.ts.map +1 -0
- package/dist/cli/commands/cluster.js +77 -0
- package/dist/cli/commands/cluster.js.map +1 -0
- package/dist/cli/commands/cohort.d.ts +10 -0
- package/dist/cli/commands/cohort.d.ts.map +1 -0
- package/dist/cli/commands/cohort.js +66 -0
- package/dist/cli/commands/cohort.js.map +1 -0
- package/dist/cli/commands/completion.d.ts +5 -0
- package/dist/cli/commands/completion.d.ts.map +1 -0
- package/dist/cli/commands/completion.js +7 -0
- package/dist/cli/commands/completion.js.map +1 -0
- package/dist/cli/commands/concat.d.ts +9 -0
- package/dist/cli/commands/concat.d.ts.map +1 -0
- package/dist/cli/commands/concat.js +62 -0
- package/dist/cli/commands/concat.js.map +1 -0
- package/dist/cli/commands/convert.d.ts +2 -0
- package/dist/cli/commands/convert.d.ts.map +1 -1
- package/dist/cli/commands/convert.js.map +1 -1
- package/dist/cli/commands/corr.d.ts +7 -0
- package/dist/cli/commands/corr.d.ts.map +1 -0
- package/dist/cli/commands/corr.js +70 -0
- package/dist/cli/commands/corr.js.map +1 -0
- package/dist/cli/commands/crosstab.d.ts +11 -0
- package/dist/cli/commands/crosstab.d.ts.map +1 -0
- package/dist/cli/commands/crosstab.js +71 -0
- package/dist/cli/commands/crosstab.js.map +1 -0
- package/dist/cli/commands/entropy.d.ts +9 -0
- package/dist/cli/commands/entropy.d.ts.map +1 -0
- package/dist/cli/commands/entropy.js +71 -0
- package/dist/cli/commands/entropy.js.map +1 -0
- package/dist/cli/commands/explode.d.ts +8 -0
- package/dist/cli/commands/explode.d.ts.map +1 -0
- package/dist/cli/commands/explode.js +61 -0
- package/dist/cli/commands/explode.js.map +1 -0
- package/dist/cli/commands/fetch.d.ts +8 -0
- package/dist/cli/commands/fetch.d.ts.map +1 -0
- package/dist/cli/commands/fetch.js +41 -0
- package/dist/cli/commands/fetch.js.map +1 -0
- package/dist/cli/commands/files.d.ts +1 -1
- package/dist/cli/commands/files.d.ts.map +1 -1
- package/dist/cli/commands/flatten.d.ts +8 -0
- package/dist/cli/commands/flatten.d.ts.map +1 -0
- package/dist/cli/commands/flatten.js +57 -0
- package/dist/cli/commands/flatten.js.map +1 -0
- package/dist/cli/commands/freq.d.ts +10 -0
- package/dist/cli/commands/freq.d.ts.map +1 -0
- package/dist/cli/commands/freq.js +122 -0
- package/dist/cli/commands/freq.js.map +1 -0
- package/dist/cli/commands/funnel.d.ts +11 -0
- package/dist/cli/commands/funnel.d.ts.map +1 -0
- package/dist/cli/commands/funnel.js +69 -0
- package/dist/cli/commands/funnel.js.map +1 -0
- package/dist/cli/commands/fuzzy-join.d.ts +25 -0
- package/dist/cli/commands/fuzzy-join.d.ts.map +1 -0
- package/dist/cli/commands/fuzzy-join.js +85 -0
- package/dist/cli/commands/fuzzy-join.js.map +1 -0
- package/dist/cli/commands/generate.d.ts +8 -0
- package/dist/cli/commands/generate.d.ts.map +1 -0
- package/dist/cli/commands/generate.js +34 -0
- package/dist/cli/commands/generate.js.map +1 -0
- package/dist/cli/commands/mask.d.ts +6 -0
- package/dist/cli/commands/mask.d.ts.map +1 -0
- package/dist/cli/commands/mask.js +58 -0
- package/dist/cli/commands/mask.js.map +1 -0
- package/dist/cli/commands/ngrams.d.ts +11 -0
- package/dist/cli/commands/ngrams.d.ts.map +1 -0
- package/dist/cli/commands/ngrams.js +70 -0
- package/dist/cli/commands/ngrams.js.map +1 -0
- package/dist/cli/commands/outliers.d.ts +13 -0
- package/dist/cli/commands/outliers.d.ts.map +1 -0
- package/dist/cli/commands/outliers.js +58 -0
- package/dist/cli/commands/outliers.js.map +1 -0
- package/dist/cli/commands/pareto.d.ts +8 -0
- package/dist/cli/commands/pareto.d.ts.map +1 -0
- package/dist/cli/commands/pareto.js +80 -0
- package/dist/cli/commands/pareto.js.map +1 -0
- package/dist/cli/commands/partition.d.ts +8 -0
- package/dist/cli/commands/partition.d.ts.map +1 -0
- package/dist/cli/commands/partition.js +44 -0
- package/dist/cli/commands/partition.js.map +1 -0
- package/dist/cli/commands/pipeline.d.ts +6 -0
- package/dist/cli/commands/pipeline.d.ts.map +1 -1
- package/dist/cli/commands/pipeline.js +22 -0
- package/dist/cli/commands/pipeline.js.map +1 -1
- package/dist/cli/commands/pivot.d.ts +11 -0
- package/dist/cli/commands/pivot.d.ts.map +1 -0
- package/dist/cli/commands/pivot.js +58 -0
- package/dist/cli/commands/pivot.js.map +1 -0
- package/dist/cli/commands/plot.d.ts +12 -0
- package/dist/cli/commands/plot.d.ts.map +1 -0
- package/dist/cli/commands/plot.js +123 -0
- package/dist/cli/commands/plot.js.map +1 -0
- package/dist/cli/commands/quantiles.d.ts +7 -0
- package/dist/cli/commands/quantiles.d.ts.map +1 -0
- package/dist/cli/commands/quantiles.js +94 -0
- package/dist/cli/commands/quantiles.js.map +1 -0
- package/dist/cli/commands/regression.d.ts +7 -0
- package/dist/cli/commands/regression.d.ts.map +1 -0
- package/dist/cli/commands/regression.js +86 -0
- package/dist/cli/commands/regression.js.map +1 -0
- package/dist/cli/commands/report.d.ts +8 -0
- package/dist/cli/commands/report.d.ts.map +1 -0
- package/dist/cli/commands/report.js +36 -0
- package/dist/cli/commands/report.js.map +1 -0
- package/dist/cli/commands/rfm.d.ts +11 -0
- package/dist/cli/commands/rfm.d.ts.map +1 -0
- package/dist/cli/commands/rfm.js +78 -0
- package/dist/cli/commands/rfm.js.map +1 -0
- package/dist/cli/commands/serve.d.ts +7 -0
- package/dist/cli/commands/serve.d.ts.map +1 -0
- package/dist/cli/commands/serve.js +43 -0
- package/dist/cli/commands/serve.js.map +1 -0
- package/dist/cli/commands/split.d.ts +8 -0
- package/dist/cli/commands/split.d.ts.map +1 -0
- package/dist/cli/commands/split.js +43 -0
- package/dist/cli/commands/split.js.map +1 -0
- package/dist/cli/commands/table.d.ts +9 -0
- package/dist/cli/commands/table.d.ts.map +1 -0
- package/dist/cli/commands/table.js +9 -0
- package/dist/cli/commands/table.js.map +1 -0
- package/dist/cli/commands/technical.d.ts +8 -0
- package/dist/cli/commands/technical.d.ts.map +1 -0
- package/dist/cli/commands/technical.js +54 -0
- package/dist/cli/commands/technical.js.map +1 -0
- package/dist/cli/commands/test.d.ts +7 -0
- package/dist/cli/commands/test.d.ts.map +1 -0
- package/dist/cli/commands/test.js +53 -0
- package/dist/cli/commands/test.js.map +1 -0
- package/dist/cli/commands/timeseries.d.ts +11 -0
- package/dist/cli/commands/timeseries.d.ts.map +1 -0
- package/dist/cli/commands/timeseries.js +82 -0
- package/dist/cli/commands/timeseries.js.map +1 -0
- package/dist/cli/commands/unpivot.d.ts +12 -0
- package/dist/cli/commands/unpivot.d.ts.map +1 -0
- package/dist/cli/commands/unpivot.js +59 -0
- package/dist/cli/commands/unpivot.js.map +1 -0
- package/dist/cli/completion.d.ts +5 -0
- package/dist/cli/completion.d.ts.map +1 -0
- package/dist/cli/completion.js +194 -0
- package/dist/cli/completion.js.map +1 -0
- package/dist/cli/index.js +816 -15
- package/dist/cli/index.js.map +1 -1
- package/dist/core/error-handler.d.ts +38 -0
- package/dist/core/error-handler.d.ts.map +1 -0
- package/dist/core/error-handler.js +81 -0
- package/dist/core/error-handler.js.map +1 -0
- package/dist/core/pipeline.d.ts +10 -0
- package/dist/core/pipeline.d.ts.map +1 -1
- package/dist/core/pipeline.js +14 -0
- package/dist/core/pipeline.js.map +1 -1
- package/dist/core/types.d.ts +4 -0
- package/dist/core/types.d.ts.map +1 -1
- package/dist/dataops/fetch.d.ts +27 -0
- package/dist/dataops/fetch.d.ts.map +1 -0
- package/dist/dataops/fetch.js +150 -0
- package/dist/dataops/fetch.js.map +1 -0
- package/dist/dataops/generate.d.ts +16 -0
- package/dist/dataops/generate.d.ts.map +1 -0
- package/dist/dataops/generate.js +201 -0
- package/dist/dataops/generate.js.map +1 -0
- package/dist/dataops/mask.d.ts +25 -0
- package/dist/dataops/mask.d.ts.map +1 -0
- package/dist/dataops/mask.js +133 -0
- package/dist/dataops/mask.js.map +1 -0
- package/dist/dataops/report.d.ts +7 -0
- package/dist/dataops/report.d.ts.map +1 -0
- package/dist/dataops/report.js +139 -0
- package/dist/dataops/report.js.map +1 -0
- package/dist/dataops/server.d.ts +10 -0
- package/dist/dataops/server.d.ts.map +1 -0
- package/dist/dataops/server.js +137 -0
- package/dist/dataops/server.js.map +1 -0
- package/dist/dataops/test-runner.d.ts +34 -0
- package/dist/dataops/test-runner.d.ts.map +1 -0
- package/dist/dataops/test-runner.js +128 -0
- package/dist/dataops/test-runner.js.map +1 -0
- package/dist/db/adapters/sqlite.d.ts +2 -1
- package/dist/db/adapters/sqlite.d.ts.map +1 -1
- package/dist/db/adapters/sqlite.js +54 -20
- package/dist/db/adapters/sqlite.js.map +1 -1
- package/dist/diff/engine.d.ts.map +1 -1
- package/dist/diff/engine.js +73 -16
- package/dist/diff/engine.js.map +1 -1
- package/dist/diff/storage/disk-index.d.ts +10 -1
- package/dist/diff/storage/disk-index.d.ts.map +1 -1
- package/dist/diff/storage/disk-index.js +95 -13
- package/dist/diff/storage/disk-index.js.map +1 -1
- package/dist/files/hash.d.ts +9 -2
- package/dist/files/hash.d.ts.map +1 -1
- package/dist/files/hash.js +9 -10
- package/dist/files/hash.js.map +1 -1
- package/dist/files/reader.d.ts.map +1 -1
- package/dist/files/reader.js +17 -10
- package/dist/files/reader.js.map +1 -1
- package/dist/files/types.d.ts +2 -1
- package/dist/files/types.d.ts.map +1 -1
- package/dist/index.d.ts +65 -1
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +66 -1
- package/dist/index.js.map +1 -1
- package/dist/planner/executor.d.ts.map +1 -1
- package/dist/planner/executor.js +17 -0
- package/dist/planner/executor.js.map +1 -1
- package/dist/planner/types.d.ts +11 -0
- package/dist/planner/types.d.ts.map +1 -1
- package/dist/readers/csv.d.ts.map +1 -1
- package/dist/readers/csv.js +206 -164
- package/dist/readers/csv.js.map +1 -1
- package/dist/readers/json.d.ts.map +1 -1
- package/dist/readers/json.js +145 -121
- package/dist/readers/json.js.map +1 -1
- package/dist/readers/jsonl.d.ts.map +1 -1
- package/dist/readers/jsonl.js +51 -31
- package/dist/readers/jsonl.js.map +1 -1
- package/dist/rowpipe +4 -0
- package/dist/transforms/concat.d.ts +26 -0
- package/dist/transforms/concat.d.ts.map +1 -0
- package/dist/transforms/concat.js +57 -0
- package/dist/transforms/concat.js.map +1 -0
- package/dist/transforms/explode.d.ts +14 -0
- package/dist/transforms/explode.d.ts.map +1 -0
- package/dist/transforms/explode.js +81 -0
- package/dist/transforms/explode.js.map +1 -0
- package/dist/transforms/flatten.d.ts +12 -0
- package/dist/transforms/flatten.d.ts.map +1 -0
- package/dist/transforms/flatten.js +65 -0
- package/dist/transforms/flatten.js.map +1 -0
- package/dist/transforms/fuzzy-join.d.ts +53 -0
- package/dist/transforms/fuzzy-join.d.ts.map +1 -0
- package/dist/transforms/fuzzy-join.js +347 -0
- package/dist/transforms/fuzzy-join.js.map +1 -0
- package/dist/transforms/jit-compiler.d.ts.map +1 -1
- package/dist/transforms/jit-compiler.js +103 -14
- package/dist/transforms/jit-compiler.js.map +1 -1
- package/dist/transforms/join/index-storage.d.ts +8 -5
- package/dist/transforms/join/index-storage.d.ts.map +1 -1
- package/dist/transforms/join/index-storage.js +91 -54
- package/dist/transforms/join/index-storage.js.map +1 -1
- package/dist/transforms/map.d.ts +6 -2
- package/dist/transforms/map.d.ts.map +1 -1
- package/dist/transforms/map.js +64 -26
- package/dist/transforms/map.js.map +1 -1
- package/dist/transforms/partition.d.ts +35 -0
- package/dist/transforms/partition.d.ts.map +1 -0
- package/dist/transforms/partition.js +150 -0
- package/dist/transforms/partition.js.map +1 -0
- package/dist/transforms/pivot.d.ts +16 -0
- package/dist/transforms/pivot.d.ts.map +1 -0
- package/dist/transforms/pivot.js +194 -0
- package/dist/transforms/pivot.js.map +1 -0
- package/dist/transforms/sort/comparator.d.ts.map +1 -1
- package/dist/transforms/sort/comparator.js +153 -24
- package/dist/transforms/sort/comparator.js.map +1 -1
- package/dist/transforms/sort/external-sort.js +1 -1
- package/dist/transforms/sort/external-sort.js.map +1 -1
- package/dist/transforms/split.d.ts +17 -0
- package/dist/transforms/split.d.ts.map +1 -0
- package/dist/transforms/split.js +88 -0
- package/dist/transforms/split.js.map +1 -0
- package/dist/transforms/timeseries.d.ts +42 -0
- package/dist/transforms/timeseries.d.ts.map +1 -0
- package/dist/transforms/timeseries.js +276 -0
- package/dist/transforms/timeseries.js.map +1 -0
- package/dist/transforms/unpivot.d.ts +14 -0
- package/dist/transforms/unpivot.d.ts.map +1 -0
- package/dist/transforms/unpivot.js +56 -0
- package/dist/transforms/unpivot.js.map +1 -0
- package/dist/ui/chart.d.ts +30 -0
- package/dist/ui/chart.d.ts.map +1 -0
- package/dist/ui/chart.js +91 -0
- package/dist/ui/chart.js.map +1 -0
- package/dist/writers/index.d.ts +1 -0
- package/dist/writers/index.d.ts.map +1 -1
- package/dist/writers/index.js +10 -0
- package/dist/writers/index.js.map +1 -1
- package/dist/writers/table.d.ts +23 -0
- package/dist/writers/table.d.ts.map +1 -0
- package/dist/writers/table.js +156 -0
- package/dist/writers/table.js.map +1 -0
- package/package.json +6 -3
|
@@ -0,0 +1,180 @@
|
|
|
1
|
+
function discretizeNumeric(val, min, max, bins) {
|
|
2
|
+
if (min === max)
|
|
3
|
+
return "bin_0";
|
|
4
|
+
const normalized = Math.max(0, Math.min(0.9999, (val - min) / (max - min)));
|
|
5
|
+
const binIdx = Math.floor(normalized * bins);
|
|
6
|
+
return `bin_${binIdx}`;
|
|
7
|
+
}
|
|
8
|
+
export async function computeEntropy(batchIterator, options = {}) {
|
|
9
|
+
const bins = options.bins ?? 10;
|
|
10
|
+
let targetCol = options.targetCol;
|
|
11
|
+
let targetCols;
|
|
12
|
+
if (options.cols) {
|
|
13
|
+
targetCols = Array.isArray(options.cols)
|
|
14
|
+
? options.cols.flatMap((c) => c.split(",").map((s) => s.trim()))
|
|
15
|
+
: options.cols.split(",").map((s) => s.trim());
|
|
16
|
+
}
|
|
17
|
+
// First pass: collect distinct value counts, min/max for numerics, and joint counts
|
|
18
|
+
const singleCounts = new Map();
|
|
19
|
+
const jointCounts = new Map(); // feature -> targetVal -> featVal -> count
|
|
20
|
+
const numericRanges = new Map();
|
|
21
|
+
let totalSamples = 0;
|
|
22
|
+
for await (const batch of batchIterator) {
|
|
23
|
+
if (batch.rows.length === 0)
|
|
24
|
+
continue;
|
|
25
|
+
if (!targetCols) {
|
|
26
|
+
const firstRow = batch.rows[0];
|
|
27
|
+
const keys = Object.keys(firstRow);
|
|
28
|
+
if (!targetCol && keys.length > 1) {
|
|
29
|
+
targetCol =
|
|
30
|
+
keys.find((k) => /(target|label|class|churn|churned|survived|y|outcome|status)/i.test(k)) ??
|
|
31
|
+
keys[keys.length - 1];
|
|
32
|
+
}
|
|
33
|
+
targetCols = keys.filter((k) => k !== targetCol);
|
|
34
|
+
}
|
|
35
|
+
for (const row of batch.rows) {
|
|
36
|
+
totalSamples++;
|
|
37
|
+
for (const col of [targetCol, ...targetCols].filter(Boolean)) {
|
|
38
|
+
const val = row[col];
|
|
39
|
+
if (typeof val === "number") {
|
|
40
|
+
const rng = numericRanges.get(col);
|
|
41
|
+
if (!rng) {
|
|
42
|
+
numericRanges.set(col, { min: val, max: val });
|
|
43
|
+
}
|
|
44
|
+
else {
|
|
45
|
+
if (val < rng.min)
|
|
46
|
+
rng.min = val;
|
|
47
|
+
if (val > rng.max)
|
|
48
|
+
rng.max = val;
|
|
49
|
+
}
|
|
50
|
+
}
|
|
51
|
+
}
|
|
52
|
+
// Track target single counts
|
|
53
|
+
let targetValStr = "";
|
|
54
|
+
if (targetCol) {
|
|
55
|
+
const rawT = row[targetCol];
|
|
56
|
+
targetValStr = rawT !== null && rawT !== undefined ? String(rawT) : "<null>";
|
|
57
|
+
let tMap = singleCounts.get(targetCol);
|
|
58
|
+
if (!tMap) {
|
|
59
|
+
tMap = new Map();
|
|
60
|
+
singleCounts.set(targetCol, tMap);
|
|
61
|
+
}
|
|
62
|
+
tMap.set(targetValStr, (tMap.get(targetValStr) || 0) + 1);
|
|
63
|
+
}
|
|
64
|
+
// Track feature single & joint counts
|
|
65
|
+
for (const col of targetCols) {
|
|
66
|
+
const rawVal = row[col];
|
|
67
|
+
let valStr = "";
|
|
68
|
+
const rng = numericRanges.get(col);
|
|
69
|
+
if (typeof rawVal === "number" && rng) {
|
|
70
|
+
valStr = discretizeNumeric(rawVal, rng.min, rng.max, bins);
|
|
71
|
+
}
|
|
72
|
+
else {
|
|
73
|
+
valStr = rawVal !== null && rawVal !== undefined ? String(rawVal) : "<null>";
|
|
74
|
+
}
|
|
75
|
+
let fMap = singleCounts.get(col);
|
|
76
|
+
if (!fMap) {
|
|
77
|
+
fMap = new Map();
|
|
78
|
+
singleCounts.set(col, fMap);
|
|
79
|
+
}
|
|
80
|
+
fMap.set(valStr, (fMap.get(valStr) || 0) + 1);
|
|
81
|
+
if (targetCol) {
|
|
82
|
+
let jFeature = jointCounts.get(col);
|
|
83
|
+
if (!jFeature) {
|
|
84
|
+
jFeature = new Map();
|
|
85
|
+
jointCounts.set(col, jFeature);
|
|
86
|
+
}
|
|
87
|
+
let tMap = jFeature.get(targetValStr);
|
|
88
|
+
if (!tMap) {
|
|
89
|
+
tMap = new Map();
|
|
90
|
+
jFeature.set(targetValStr, tMap);
|
|
91
|
+
}
|
|
92
|
+
tMap.set(valStr, (tMap.get(valStr) || 0) + 1);
|
|
93
|
+
}
|
|
94
|
+
}
|
|
95
|
+
}
|
|
96
|
+
}
|
|
97
|
+
if (totalSamples === 0 || !targetCols) {
|
|
98
|
+
throw new Error("No data available for entropy computation.");
|
|
99
|
+
}
|
|
100
|
+
// Calculate Target Entropy H(Y)
|
|
101
|
+
let targetEntropy = 0;
|
|
102
|
+
if (targetCol) {
|
|
103
|
+
const tMap = singleCounts.get(targetCol);
|
|
104
|
+
if (tMap) {
|
|
105
|
+
for (const count of tMap.values()) {
|
|
106
|
+
const p = count / totalSamples;
|
|
107
|
+
if (p > 0) {
|
|
108
|
+
targetEntropy -= p * Math.log2(p);
|
|
109
|
+
}
|
|
110
|
+
}
|
|
111
|
+
}
|
|
112
|
+
}
|
|
113
|
+
const features = [];
|
|
114
|
+
for (const col of targetCols) {
|
|
115
|
+
const fMap = singleCounts.get(col);
|
|
116
|
+
if (!fMap)
|
|
117
|
+
continue;
|
|
118
|
+
let featureEntropy = 0;
|
|
119
|
+
for (const count of fMap.values()) {
|
|
120
|
+
const p = count / totalSamples;
|
|
121
|
+
if (p > 0) {
|
|
122
|
+
featureEntropy -= p * Math.log2(p);
|
|
123
|
+
}
|
|
124
|
+
}
|
|
125
|
+
let infoGain = 0;
|
|
126
|
+
let gainRatio = 0;
|
|
127
|
+
if (targetCol && targetEntropy > 0) {
|
|
128
|
+
const jFeature = jointCounts.get(col);
|
|
129
|
+
let condEntropy = 0;
|
|
130
|
+
if (jFeature) {
|
|
131
|
+
// H(Y|X) = - sum_{x,y} p(x,y) log2(p(x,y) / p(x))
|
|
132
|
+
for (const [tVal, tMap] of jFeature.entries()) {
|
|
133
|
+
for (const [fVal, count] of tMap.entries()) {
|
|
134
|
+
const pXY = count / totalSamples;
|
|
135
|
+
const pX = (fMap.get(fVal) || 0) / totalSamples;
|
|
136
|
+
if (pXY > 0 && pX > 0) {
|
|
137
|
+
condEntropy -= pXY * Math.log2(pXY / pX);
|
|
138
|
+
}
|
|
139
|
+
}
|
|
140
|
+
}
|
|
141
|
+
}
|
|
142
|
+
infoGain = Math.max(0, targetEntropy - condEntropy);
|
|
143
|
+
gainRatio = featureEntropy > 0 ? infoGain / featureEntropy : 0;
|
|
144
|
+
}
|
|
145
|
+
features.push({
|
|
146
|
+
feature: col,
|
|
147
|
+
entropy: Math.round(featureEntropy * 10000) / 10000,
|
|
148
|
+
information_gain: Math.round(infoGain * 10000) / 10000,
|
|
149
|
+
gain_ratio: Math.round(gainRatio * 10000) / 10000,
|
|
150
|
+
importance_score: targetEntropy > 0 ? Math.round((infoGain / targetEntropy) * 1000) / 10 : 0,
|
|
151
|
+
distinct_values: fMap.size,
|
|
152
|
+
});
|
|
153
|
+
}
|
|
154
|
+
// Sort by Information Gain / Importance descending
|
|
155
|
+
features.sort((a, b) => b.information_gain - a.information_gain || b.entropy - a.entropy);
|
|
156
|
+
const formattedRows = targetCol
|
|
157
|
+
? features.map((f, idx) => ({
|
|
158
|
+
rank: idx + 1,
|
|
159
|
+
feature: f.feature,
|
|
160
|
+
info_gain: f.information_gain,
|
|
161
|
+
gain_ratio: f.gain_ratio,
|
|
162
|
+
entropy: f.entropy,
|
|
163
|
+
importance: `${f.importance_score.toFixed(1)}%`,
|
|
164
|
+
cardinality: f.distinct_values,
|
|
165
|
+
}))
|
|
166
|
+
: features.map((f, idx) => ({
|
|
167
|
+
rank: idx + 1,
|
|
168
|
+
feature: f.feature,
|
|
169
|
+
entropy: f.entropy,
|
|
170
|
+
cardinality: f.distinct_values,
|
|
171
|
+
}));
|
|
172
|
+
return {
|
|
173
|
+
target_column: targetCol,
|
|
174
|
+
target_entropy: targetCol ? Math.round(targetEntropy * 10000) / 10000 : undefined,
|
|
175
|
+
total_samples: totalSamples,
|
|
176
|
+
features,
|
|
177
|
+
formattedRows,
|
|
178
|
+
};
|
|
179
|
+
}
|
|
180
|
+
//# sourceMappingURL=entropy.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"entropy.js","sourceRoot":"","sources":["../../src/analytics/entropy.ts"],"names":[],"mappings":"AAyBA,SAAS,iBAAiB,CAAC,GAAW,EAAE,GAAW,EAAE,GAAW,EAAE,IAAY;IAC5E,IAAI,GAAG,KAAK,GAAG;QAAE,OAAO,OAAO,CAAC;IAChC,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,GAAG,GAAG,GAAG,CAAC,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;IAC5E,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,GAAG,IAAI,CAAC,CAAC;IAC7C,OAAO,OAAO,MAAM,EAAE,CAAC;AACzB,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,aAAuC,EACvC,UAA0B,EAAE;IAE5B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC;IAChC,IAAI,SAAS,GAAG,OAAO,CAAC,SAAS,CAAC;IAClC,IAAI,UAAgC,CAAC;IAErC,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC;QACjB,UAAU,GAAG,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC;YACtC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YAChE,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IACnD,CAAC;IAED,oFAAoF;IACpF,MAAM,YAAY,GAAG,IAAI,GAAG,EAA+B,CAAC;IAC5D,MAAM,WAAW,GAAG,IAAI,GAAG,EAA4C,CAAC,CAAC,2CAA2C;IACpH,MAAM,aAAa,GAAG,IAAI,GAAG,EAAwC,CAAC;IACtE,IAAI,YAAY,GAAG,CAAC,CAAC;IAErB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;QACxC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAEtC,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;YAChC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YACnC,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClC,SAAS;oBACP,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,+DAA+D,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;wBACzF,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAC1B,CAAC;YACD,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC,CAAC;QACnD,CAAC;QAED,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;YAC7B,YAAY,EAAE,CAAC;YAEf,KAAK,MAAM,GAAG,IAAI,CAAC,SAAS,EAAE,GAAG,UAAU,CAAC,CAAC,MAAM,CAAC,OAAO,CAAa,EAAE,CAAC;gBACzE,MAAM,GAAG,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC;gBACrB,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;oBAC5B,MAAM,GAAG,GAAG,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;oBACnC,IAAI,CAAC,GAAG,EAAE,CAAC;wBACT,aAAa,CAAC,GAAG,CAAC,GAAG,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC;oBACjD,CAAC;yBAAM,CAAC;wBACN,IAAI,GAAG,GAAG,GAAG,CAAC,GAAG;4BAAE,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC;wBACjC,IAAI,GAAG,GAAG,GAAG,CAAC,GAAG;4BAAE,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC;oBACnC,CAAC;gBACH,CAAC;YACH,CAAC;YAED,6BAA6B;YAC7B,IAAI,YAAY,GAAG,EAAE,CAAC;YACtB,IAAI,SAAS,EAAE,CAAC;gBACd,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC,CAAC;gBAC5B,YAAY,GAAG,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;gBAC7E,IAAI,IAAI,GAAG,YAAY,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;gBACvC,IAAI,CAAC,IAAI,EAAE,CAAC;oBACV,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAC;oBACjC,YAAY,CAAC,GAAG,CAAC,SAAS,EAAE,IAAI,CAAC,CAAC;gBACpC,CAAC;gBACD,IAAI,CAAC,GAAG,CAAC,YAAY,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAC5D,CAAC;YAED,sCAAsC;YACtC,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;gBAC7B,MAAM,MAAM,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC;gBACxB,IAAI,MAAM,GAAG,EAAE,CAAC;gBAChB,MAAM,GAAG,GAAG,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;gBACnC,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,GAAG,EAAE,CAAC;oBACtC,MAAM,GAAG,iBAAiB,CAAC,MAAM,EAAE,GAAG,CAAC,GAAG,EAAE,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;gBAC7D,CAAC;qBAAM,CAAC;oBACN,MAAM,GAAG,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;gBAC/E,CAAC;gBAED,IAAI,IAAI,GAAG,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;gBACjC,IAAI,CAAC,IAAI,EAAE,CAAC;oBACV,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAC;oBACjC,YAAY,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;gBAC9B,CAAC;gBACD,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAE9C,IAAI,SAAS,EAAE,CAAC;oBACd,IAAI,QAAQ,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;oBACpC,IAAI,CAAC,QAAQ,EAAE,CAAC;wBACd,QAAQ,GAAG,IAAI,GAAG,EAA+B,CAAC;wBAClD,WAAW,CAAC,GAAG,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;oBACjC,CAAC;oBACD,IAAI,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;oBACtC,IAAI,CAAC,IAAI,EAAE,CAAC;wBACV,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAC;wBACjC,QAAQ,CAAC,GAAG,CAAC,YAAY,EAAE,IAAI,CAAC,CAAC;oBACnC,CAAC;oBACD,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAChD,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,YAAY,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;QACtC,MAAM,IAAI,KAAK,CAAC,4CAA4C,CAAC,CAAC;IAChE,CAAC;IAED,gCAAgC;IAChC,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,IAAI,SAAS,EAAE,CAAC;QACd,MAAM,IAAI,GAAG,YAAY,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QACzC,IAAI,IAAI,EAAE,CAAC;YACT,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,MAAM,EAAE,EAAE,CAAC;gBAClC,MAAM,CAAC,GAAG,KAAK,GAAG,YAAY,CAAC;gBAC/B,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;oBACV,aAAa,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBACpC,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,QAAQ,GAA8B,EAAE,CAAC;IAE/C,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI;YAAE,SAAS;QAEpB,IAAI,cAAc,GAAG,CAAC,CAAC;QACvB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,MAAM,EAAE,EAAE,CAAC;YAClC,MAAM,CAAC,GAAG,KAAK,GAAG,YAAY,CAAC;YAC/B,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;gBACV,cAAc,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAED,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,IAAI,SAAS,GAAG,CAAC,CAAC;QAElB,IAAI,SAAS,IAAI,aAAa,GAAG,CAAC,EAAE,CAAC;YACnC,MAAM,QAAQ,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACtC,IAAI,WAAW,GAAG,CAAC,CAAC;YAEpB,IAAI,QAAQ,EAAE,CAAC;gBACb,kDAAkD;gBAClD,KAAK,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,QAAQ,CAAC,OAAO,EAAE,EAAE,CAAC;oBAC9C,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;wBAC3C,MAAM,GAAG,GAAG,KAAK,GAAG,YAAY,CAAC;wBACjC,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,YAAY,CAAC;wBAChD,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,EAAE,CAAC;4BACtB,WAAW,IAAI,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,GAAG,EAAE,CAAC,CAAC;wBAC3C,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;YACD,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,aAAa,GAAG,WAAW,CAAC,CAAC;YACpD,SAAS,GAAG,cAAc,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,GAAG,cAAc,CAAC,CAAC,CAAC,CAAC,CAAC;QACjE,CAAC;QAED,QAAQ,CAAC,IAAI,CAAC;YACZ,OAAO,EAAE,GAAG;YACZ,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,cAAc,GAAG,KAAK,CAAC,GAAG,KAAK;YACnD,gBAAgB,EAAE,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,KAAK,CAAC,GAAG,KAAK;YACtD,UAAU,EAAE,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC,GAAG,KAAK;YACjD,gBAAgB,EAAE,aAAa,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,GAAG,aAAa,CAAC,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;YAC5F,eAAe,EAAE,IAAI,CAAC,IAAI;SAC3B,CAAC,CAAC;IACL,CAAC;IAED,mDAAmD;IACnD,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,gBAAgB,GAAG,CAAC,CAAC,gBAAgB,IAAI,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC;IAE1F,MAAM,aAAa,GAAU,SAAS;QACpC,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC;YACxB,IAAI,EAAE,GAAG,GAAG,CAAC;YACb,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,SAAS,EAAE,CAAC,CAAC,gBAAgB;YAC7B,UAAU,EAAE,CAAC,CAAC,UAAU;YACxB,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,UAAU,EAAE,GAAG,CAAC,CAAC,gBAAgB,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG;YAC/C,WAAW,EAAE,CAAC,CAAC,eAAe;SAC/B,CAAC,CAAC;QACL,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC;YACxB,IAAI,EAAE,GAAG,GAAG,CAAC;YACb,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,WAAW,EAAE,CAAC,CAAC,eAAe;SAC/B,CAAC,CAAC,CAAC;IAER,OAAO;QACL,aAAa,EAAE,SAAS;QACxB,cAAc,EAAE,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,aAAa,GAAG,KAAK,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,SAAS;QACjF,aAAa,EAAE,YAAY;QAC3B,QAAQ;QACR,aAAa;KACd,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,29 @@
|
|
|
1
|
+
import type { Row, DataBatch } from "../core/types.js";
|
|
2
|
+
export interface FunnelOptions {
|
|
3
|
+
steps?: string | string[];
|
|
4
|
+
userIdCol?: string;
|
|
5
|
+
stepCol?: string;
|
|
6
|
+
timeCol?: string;
|
|
7
|
+
strict?: boolean;
|
|
8
|
+
windowMs?: number;
|
|
9
|
+
}
|
|
10
|
+
export interface FunnelStepMetric {
|
|
11
|
+
step: string;
|
|
12
|
+
step_index: number;
|
|
13
|
+
users: number;
|
|
14
|
+
step_conversion_pct: number;
|
|
15
|
+
overall_conversion_pct: number;
|
|
16
|
+
dropoff_users: number;
|
|
17
|
+
dropoff_pct: number;
|
|
18
|
+
visual_bar: string;
|
|
19
|
+
}
|
|
20
|
+
export interface FunnelResult {
|
|
21
|
+
steps: FunnelStepMetric[];
|
|
22
|
+
total_entry_users: number;
|
|
23
|
+
final_converted_users: number;
|
|
24
|
+
overall_conversion_rate: number;
|
|
25
|
+
formattedRows: Row[];
|
|
26
|
+
}
|
|
27
|
+
export declare function parseStepList(stepInput?: string | string[]): string[];
|
|
28
|
+
export declare function computeFunnel(batchIterator: AsyncIterable<DataBatch>, options?: FunnelOptions): Promise<FunnelResult>;
|
|
29
|
+
//# sourceMappingURL=funnel.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"funnel.d.ts","sourceRoot":"","sources":["../../src/analytics/funnel.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,GAAG,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAEvD,MAAM,WAAW,aAAa;IAC5B,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAAC;IAC1B,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,MAAM,CAAC;IACnB,KAAK,EAAE,MAAM,CAAC;IACd,mBAAmB,EAAE,MAAM,CAAC;IAC5B,sBAAsB,EAAE,MAAM,CAAC;IAC/B,aAAa,EAAE,MAAM,CAAC;IACtB,WAAW,EAAE,MAAM,CAAC;IACpB,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,YAAY;IAC3B,KAAK,EAAE,gBAAgB,EAAE,CAAC;IAC1B,iBAAiB,EAAE,MAAM,CAAC;IAC1B,qBAAqB,EAAE,MAAM,CAAC;IAC9B,uBAAuB,EAAE,MAAM,CAAC;IAChC,aAAa,EAAE,GAAG,EAAE,CAAC;CACtB;AA6BD,wBAAgB,aAAa,CAAC,SAAS,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,GAAG,MAAM,EAAE,CAYrE;AAED,wBAAsB,aAAa,CACjC,aAAa,EAAE,aAAa,CAAC,SAAS,CAAC,EACvC,OAAO,GAAE,aAAkB,GAC1B,OAAO,CAAC,YAAY,CAAC,CAqJvB"}
|
|
@@ -0,0 +1,170 @@
|
|
|
1
|
+
function parseDateToMs(val) {
|
|
2
|
+
if (val === null || val === undefined || val === "")
|
|
3
|
+
return null;
|
|
4
|
+
if (val instanceof Date && !isNaN(val.getTime()))
|
|
5
|
+
return val.getTime();
|
|
6
|
+
if (typeof val === "number" && Number.isFinite(val)) {
|
|
7
|
+
return val < 100_000_000_000 ? Math.floor(val * 1000) : Math.floor(val);
|
|
8
|
+
}
|
|
9
|
+
if (typeof val === "string") {
|
|
10
|
+
const trimmed = val.trim();
|
|
11
|
+
if (!trimmed)
|
|
12
|
+
return null;
|
|
13
|
+
const num = Number(trimmed);
|
|
14
|
+
if (!Number.isNaN(num) && /^\d{10,13}$/.test(trimmed)) {
|
|
15
|
+
return num < 100_000_000_000 ? Math.floor(num * 1000) : Math.floor(num);
|
|
16
|
+
}
|
|
17
|
+
const parsed = Date.parse(trimmed);
|
|
18
|
+
if (!Number.isNaN(parsed)) {
|
|
19
|
+
return parsed;
|
|
20
|
+
}
|
|
21
|
+
}
|
|
22
|
+
return null;
|
|
23
|
+
}
|
|
24
|
+
function renderBar(pct, width = 20) {
|
|
25
|
+
const filled = Math.max(0, Math.min(width, Math.round((pct / 100) * width)));
|
|
26
|
+
const empty = width - filled;
|
|
27
|
+
return "█".repeat(filled) + "░".repeat(empty);
|
|
28
|
+
}
|
|
29
|
+
export function parseStepList(stepInput) {
|
|
30
|
+
if (!stepInput)
|
|
31
|
+
return [];
|
|
32
|
+
if (Array.isArray(stepInput)) {
|
|
33
|
+
return stepInput
|
|
34
|
+
.flatMap((s) => s.split(/->|,|>|\|/))
|
|
35
|
+
.map((s) => s.trim())
|
|
36
|
+
.filter(Boolean);
|
|
37
|
+
}
|
|
38
|
+
return stepInput
|
|
39
|
+
.split(/->|,|>|\|/)
|
|
40
|
+
.map((s) => s.trim())
|
|
41
|
+
.filter(Boolean);
|
|
42
|
+
}
|
|
43
|
+
export async function computeFunnel(batchIterator, options = {}) {
|
|
44
|
+
const steps = parseStepList(options.steps);
|
|
45
|
+
if (steps.length === 0) {
|
|
46
|
+
throw new Error("At least two funnel steps are required (e.g. --steps 'view,cart,checkout,purchase' or 'view->cart->checkout')");
|
|
47
|
+
}
|
|
48
|
+
const stepIndexMap = new Map();
|
|
49
|
+
for (let i = 0; i < steps.length; i++) {
|
|
50
|
+
stepIndexMap.set(steps[i].toLowerCase(), i);
|
|
51
|
+
}
|
|
52
|
+
let userIdCol = options.userIdCol;
|
|
53
|
+
let stepCol = options.stepCol;
|
|
54
|
+
let timeCol = options.timeCol;
|
|
55
|
+
const isStrict = options.strict ?? false;
|
|
56
|
+
// Map: userId -> Array of events { stepIndex: number, timeMs: number }
|
|
57
|
+
const userEvents = new Map();
|
|
58
|
+
for await (const batch of batchIterator) {
|
|
59
|
+
if (batch.rows.length === 0)
|
|
60
|
+
continue;
|
|
61
|
+
if (!userIdCol || !stepCol) {
|
|
62
|
+
const firstRow = batch.rows[0];
|
|
63
|
+
const keys = Object.keys(firstRow);
|
|
64
|
+
if (!userIdCol) {
|
|
65
|
+
userIdCol =
|
|
66
|
+
keys.find((k) => /^(user(_?id)?|customer(_?id)?|client(_?id)?|account(_?id)?|uid|id)$/i.test(k)) ??
|
|
67
|
+
keys[0];
|
|
68
|
+
}
|
|
69
|
+
if (!stepCol) {
|
|
70
|
+
stepCol =
|
|
71
|
+
keys.find((k) => /(step|event(_?name)?|action|page|stage|status)/i.test(k)) ??
|
|
72
|
+
keys[1];
|
|
73
|
+
}
|
|
74
|
+
if (!timeCol) {
|
|
75
|
+
timeCol = keys.find((k) => /(date|time|timestamp|created_at|occurred_at)/i.test(k));
|
|
76
|
+
}
|
|
77
|
+
}
|
|
78
|
+
for (const row of batch.rows) {
|
|
79
|
+
const rawUid = row[userIdCol];
|
|
80
|
+
if (rawUid === null || rawUid === undefined || rawUid === "")
|
|
81
|
+
continue;
|
|
82
|
+
const uid = String(rawUid);
|
|
83
|
+
const rawStep = row[stepCol];
|
|
84
|
+
if (rawStep === null || rawStep === undefined || rawStep === "")
|
|
85
|
+
continue;
|
|
86
|
+
const stepStr = String(rawStep).toLowerCase().trim();
|
|
87
|
+
const stepIdx = stepIndexMap.get(stepStr);
|
|
88
|
+
if (stepIdx === undefined)
|
|
89
|
+
continue;
|
|
90
|
+
let tMs = 0;
|
|
91
|
+
if (timeCol) {
|
|
92
|
+
tMs = parseDateToMs(row[timeCol]) ?? 0;
|
|
93
|
+
}
|
|
94
|
+
let events = userEvents.get(uid);
|
|
95
|
+
if (!events) {
|
|
96
|
+
events = [];
|
|
97
|
+
userEvents.set(uid, events);
|
|
98
|
+
}
|
|
99
|
+
events.push({ stepIndex: stepIdx, timeMs: tMs });
|
|
100
|
+
}
|
|
101
|
+
}
|
|
102
|
+
// Count how many users completed each stage in the funnel
|
|
103
|
+
const completedStepCounts = new Array(steps.length).fill(0);
|
|
104
|
+
for (const events of userEvents.values()) {
|
|
105
|
+
if (events.length === 0)
|
|
106
|
+
continue;
|
|
107
|
+
if (timeCol || isStrict) {
|
|
108
|
+
events.sort((a, b) => a.timeMs - b.timeMs);
|
|
109
|
+
}
|
|
110
|
+
let currentStepTarget = 0;
|
|
111
|
+
let firstStepTime = 0;
|
|
112
|
+
for (const ev of events) {
|
|
113
|
+
if (ev.stepIndex === currentStepTarget) {
|
|
114
|
+
if (currentStepTarget === 0) {
|
|
115
|
+
firstStepTime = ev.timeMs;
|
|
116
|
+
}
|
|
117
|
+
if (options.windowMs && firstStepTime > 0 && ev.timeMs - firstStepTime > options.windowMs) {
|
|
118
|
+
// Out of conversion window
|
|
119
|
+
break;
|
|
120
|
+
}
|
|
121
|
+
completedStepCounts[currentStepTarget] = (completedStepCounts[currentStepTarget] || 0) + 1;
|
|
122
|
+
currentStepTarget += 1;
|
|
123
|
+
if (currentStepTarget >= steps.length) {
|
|
124
|
+
break;
|
|
125
|
+
}
|
|
126
|
+
}
|
|
127
|
+
}
|
|
128
|
+
}
|
|
129
|
+
const baseUsers = completedStepCounts[0] || 0;
|
|
130
|
+
const metrics = [];
|
|
131
|
+
const formattedRows = [];
|
|
132
|
+
for (let i = 0; i < steps.length; i++) {
|
|
133
|
+
const stepName = steps[i];
|
|
134
|
+
const users = completedStepCounts[i] || 0;
|
|
135
|
+
const prevUsers = i === 0 ? users : completedStepCounts[i - 1] || 0;
|
|
136
|
+
const stepConv = prevUsers > 0 ? (users / prevUsers) * 100 : 0;
|
|
137
|
+
const overallConv = baseUsers > 0 ? (users / baseUsers) * 100 : 0;
|
|
138
|
+
const dropoff = i === 0 ? 0 : Math.max(0, prevUsers - users);
|
|
139
|
+
const dropoffPct = i === 0 ? 0 : 100 - stepConv;
|
|
140
|
+
const metric = {
|
|
141
|
+
step: stepName,
|
|
142
|
+
step_index: i + 1,
|
|
143
|
+
users,
|
|
144
|
+
step_conversion_pct: Math.round(stepConv * 10) / 10,
|
|
145
|
+
overall_conversion_pct: Math.round(overallConv * 10) / 10,
|
|
146
|
+
dropoff_users: dropoff,
|
|
147
|
+
dropoff_pct: Math.round(dropoffPct * 10) / 10,
|
|
148
|
+
visual_bar: `${renderBar(overallConv, 15)} ${(Math.round(overallConv * 10) / 10).toFixed(1)}%`,
|
|
149
|
+
};
|
|
150
|
+
metrics.push(metric);
|
|
151
|
+
formattedRows.push({
|
|
152
|
+
step: stepName,
|
|
153
|
+
users,
|
|
154
|
+
step_conv: i === 0 ? "100.0%" : `${metric.step_conversion_pct.toFixed(1)}%`,
|
|
155
|
+
overall_conv: `${metric.overall_conversion_pct.toFixed(1)}%`,
|
|
156
|
+
dropoff: i === 0 ? "-" : `${dropoff} (${metric.dropoff_pct.toFixed(1)}%)`,
|
|
157
|
+
funnel: metric.visual_bar,
|
|
158
|
+
});
|
|
159
|
+
}
|
|
160
|
+
const finalUsers = completedStepCounts[steps.length - 1] || 0;
|
|
161
|
+
const totalRate = baseUsers > 0 ? (finalUsers / baseUsers) * 100 : 0;
|
|
162
|
+
return {
|
|
163
|
+
steps: metrics,
|
|
164
|
+
total_entry_users: baseUsers,
|
|
165
|
+
final_converted_users: finalUsers,
|
|
166
|
+
overall_conversion_rate: Math.round(totalRate * 10) / 10,
|
|
167
|
+
formattedRows,
|
|
168
|
+
};
|
|
169
|
+
}
|
|
170
|
+
//# sourceMappingURL=funnel.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"funnel.js","sourceRoot":"","sources":["../../src/analytics/funnel.ts"],"names":[],"mappings":"AA8BA,SAAS,aAAa,CAAC,GAAY;IACjC,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,IAAI,CAAC;IACjE,IAAI,GAAG,YAAY,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC;QAAE,OAAO,GAAG,CAAC,OAAO,EAAE,CAAC;IACvE,IAAI,OAAO,GAAG,KAAK,QAAQ,IAAI,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QACpD,OAAO,GAAG,GAAG,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC1E,CAAC;IACD,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;QAC5B,MAAM,OAAO,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;QAC3B,IAAI,CAAC,OAAO;YAAE,OAAO,IAAI,CAAC;QAC1B,MAAM,GAAG,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;QAC5B,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;YACtD,OAAO,GAAG,GAAG,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC1E,CAAC;QACD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QACnC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,SAAS,CAAC,GAAW,EAAE,KAAK,GAAG,EAAE;IACxC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IAC7E,MAAM,KAAK,GAAG,KAAK,GAAG,MAAM,CAAC;IAC7B,OAAO,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AAChD,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,SAA6B;IACzD,IAAI,CAAC,SAAS;QAAE,OAAO,EAAE,CAAC;IAC1B,IAAI,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;QAC7B,OAAO,SAAS;aACb,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC;aACpC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;aACpB,MAAM,CAAC,OAAO,CAAC,CAAC;IACrB,CAAC;IACD,OAAO,SAAS;SACb,KAAK,CAAC,WAAW,CAAC;SAClB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,OAAO,CAAC,CAAC;AACrB,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CACjC,aAAuC,EACvC,UAAyB,EAAE;IAE3B,MAAM,KAAK,GAAG,aAAa,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;IAC3C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,MAAM,IAAI,KAAK,CACb,+GAA+G,CAChH,CAAC;IACJ,CAAC;IAED,MAAM,YAAY,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC/C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC;IAC/C,CAAC;IAED,IAAI,SAAS,GAAG,OAAO,CAAC,SAAS,CAAC;IAClC,IAAI,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC;IAC9B,IAAI,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC;IAC9B,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,IAAI,KAAK,CAAC;IAEzC,uEAAuE;IACvE,MAAM,UAAU,GAAG,IAAI,GAAG,EAAwD,CAAC;IAEnF,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;QACxC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAEtC,IAAI,CAAC,SAAS,IAAI,CAAC,OAAO,EAAE,CAAC;YAC3B,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;YAChC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YAEnC,IAAI,CAAC,SAAS,EAAE,CAAC;gBACf,SAAS;oBACP,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,sEAAsE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;wBAChG,IAAI,CAAC,CAAC,CAAC,CAAC;YACZ,CAAC;YACD,IAAI,CAAC,OAAO,EAAE,CAAC;gBACb,OAAO;oBACL,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,iDAAiD,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;wBAC3E,IAAI,CAAC,CAAC,CAAC,CAAC;YACZ,CAAC;YACD,IAAI,CAAC,OAAO,EAAE,CAAC;gBACb,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,+CAA+C,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;YACtF,CAAC;QACH,CAAC;QAED,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;YAC7B,MAAM,MAAM,GAAG,GAAG,CAAC,SAAU,CAAC,CAAC;YAC/B,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,EAAE;gBAAE,SAAS;YACvE,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;YAE3B,MAAM,OAAO,GAAG,GAAG,CAAC,OAAQ,CAAC,CAAC;YAC9B,IAAI,OAAO,KAAK,IAAI,IAAI,OAAO,KAAK,SAAS,IAAI,OAAO,KAAK,EAAE;gBAAE,SAAS;YAC1E,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;YAErD,MAAM,OAAO,GAAG,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;YAC1C,IAAI,OAAO,KAAK,SAAS;gBAAE,SAAS;YAEpC,IAAI,GAAG,GAAG,CAAC,CAAC;YACZ,IAAI,OAAO,EAAE,CAAC;gBACZ,GAAG,GAAG,aAAa,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC;YACzC,CAAC;YAED,IAAI,MAAM,GAAG,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACjC,IAAI,CAAC,MAAM,EAAE,CAAC;gBACZ,MAAM,GAAG,EAAE,CAAC;gBACZ,UAAU,CAAC,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;YAC9B,CAAC;YACD,MAAM,CAAC,IAAI,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;QACnD,CAAC;IACH,CAAC;IAED,0DAA0D;IAC1D,MAAM,mBAAmB,GAAG,IAAI,KAAK,CAAS,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAEpE,KAAK,MAAM,MAAM,IAAI,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC;QACzC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAElC,IAAI,OAAO,IAAI,QAAQ,EAAE,CAAC;YACxB,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC;QAC7C,CAAC;QAED,IAAI,iBAAiB,GAAG,CAAC,CAAC;QAC1B,IAAI,aAAa,GAAG,CAAC,CAAC;QAEtB,KAAK,MAAM,EAAE,IAAI,MAAM,EAAE,CAAC;YACxB,IAAI,EAAE,CAAC,SAAS,KAAK,iBAAiB,EAAE,CAAC;gBACvC,IAAI,iBAAiB,KAAK,CAAC,EAAE,CAAC;oBAC5B,aAAa,GAAG,EAAE,CAAC,MAAM,CAAC;gBAC5B,CAAC;gBAED,IAAI,OAAO,CAAC,QAAQ,IAAI,aAAa,GAAG,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,aAAa,GAAG,OAAO,CAAC,QAAQ,EAAE,CAAC;oBAC1F,2BAA2B;oBAC3B,MAAM;gBACR,CAAC;gBAED,mBAAmB,CAAC,iBAAiB,CAAC,GAAG,CAAC,mBAAmB,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;gBAC3F,iBAAiB,IAAI,CAAC,CAAC;gBAEvB,IAAI,iBAAiB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;oBACtC,MAAM;gBACR,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,SAAS,GAAG,mBAAmB,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IAC9C,MAAM,OAAO,GAAuB,EAAE,CAAC;IACvC,MAAM,aAAa,GAAU,EAAE,CAAC;IAEhC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,mBAAmB,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC1C,MAAM,SAAS,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC;QAEpE,MAAM,QAAQ,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,SAAS,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAC/D,MAAM,WAAW,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,SAAS,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAClE,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,GAAG,KAAK,CAAC,CAAC;QAC7D,MAAM,UAAU,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,QAAQ,CAAC;QAEhD,MAAM,MAAM,GAAqB;YAC/B,IAAI,EAAE,QAAQ;YACd,UAAU,EAAE,CAAC,GAAG,CAAC;YACjB,KAAK;YACL,mBAAmB,EAAE,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,EAAE,CAAC,GAAG,EAAE;YACnD,sBAAsB,EAAE,IAAI,CAAC,KAAK,CAAC,WAAW,GAAG,EAAE,CAAC,GAAG,EAAE;YACzD,aAAa,EAAE,OAAO;YACtB,WAAW,EAAE,IAAI,CAAC,KAAK,CAAC,UAAU,GAAG,EAAE,CAAC,GAAG,EAAE;YAC7C,UAAU,EAAE,GAAG,SAAS,CAAC,WAAW,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,WAAW,GAAG,EAAE,CAAC,GAAG,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG;SAC/F,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAErB,aAAa,CAAC,IAAI,CAAC;YACjB,IAAI,EAAE,QAAQ;YACd,KAAK;YACL,SAAS,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,mBAAmB,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG;YAC3E,YAAY,EAAE,GAAG,MAAM,CAAC,sBAAsB,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG;YAC5D,OAAO,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,OAAO,KAAK,MAAM,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI;YACzE,MAAM,EAAE,MAAM,CAAC,UAAU;SAC1B,CAAC,CAAC;IACL,CAAC;IAED,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC;IAC9D,MAAM,SAAS,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,GAAG,SAAS,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAErE,OAAO;QACL,KAAK,EAAE,OAAO;QACd,iBAAiB,EAAE,SAAS;QAC5B,qBAAqB,EAAE,UAAU;QACjC,uBAAuB,EAAE,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,EAAE,CAAC,GAAG,EAAE;QACxD,aAAa;KACd,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,25 @@
|
|
|
1
|
+
import type { Row, DataBatch } from "../core/types.js";
|
|
2
|
+
export interface NgramOptions {
|
|
3
|
+
col?: string;
|
|
4
|
+
n?: number;
|
|
5
|
+
top?: number;
|
|
6
|
+
stopwords?: boolean;
|
|
7
|
+
minFreq?: number;
|
|
8
|
+
caseSensitive?: boolean;
|
|
9
|
+
}
|
|
10
|
+
export interface NgramFrequencyItem {
|
|
11
|
+
ngram: string;
|
|
12
|
+
count: number;
|
|
13
|
+
percentage: number;
|
|
14
|
+
bar: string;
|
|
15
|
+
}
|
|
16
|
+
export interface NgramResult {
|
|
17
|
+
n: number;
|
|
18
|
+
column: string;
|
|
19
|
+
total_ngrams: number;
|
|
20
|
+
unique_ngrams: number;
|
|
21
|
+
items: NgramFrequencyItem[];
|
|
22
|
+
formattedRows: Row[];
|
|
23
|
+
}
|
|
24
|
+
export declare function computeNgrams(batchIterator: AsyncIterable<DataBatch>, options?: NgramOptions): Promise<NgramResult>;
|
|
25
|
+
//# sourceMappingURL=ngrams.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"ngrams.d.ts","sourceRoot":"","sources":["../../src/analytics/ngrams.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,GAAG,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAEvD,MAAM,WAAW,YAAY;IAC3B,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,CAAC,CAAC,EAAE,MAAM,CAAC;IACX,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,aAAa,CAAC,EAAE,OAAO,CAAC;CACzB;AAED,MAAM,WAAW,kBAAkB;IACjC,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;IACnB,GAAG,EAAE,MAAM,CAAC;CACb;AAED,MAAM,WAAW,WAAW;IAC1B,CAAC,EAAE,MAAM,CAAC;IACV,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,EAAE,MAAM,CAAC;IACrB,aAAa,EAAE,MAAM,CAAC;IACtB,KAAK,EAAE,kBAAkB,EAAE,CAAC;IAC5B,aAAa,EAAE,GAAG,EAAE,CAAC;CACtB;AA6CD,wBAAsB,aAAa,CACjC,aAAa,EAAE,aAAa,CAAC,SAAS,CAAC,EACvC,OAAO,GAAE,YAAiB,GACzB,OAAO,CAAC,WAAW,CAAC,CA8EtB"}
|
|
@@ -0,0 +1,110 @@
|
|
|
1
|
+
const DEFAULT_STOPWORDS = new Set([
|
|
2
|
+
"a", "about", "above", "after", "again", "against", "all", "am", "an", "and",
|
|
3
|
+
"any", "are", "aren't", "as", "at", "be", "because", "been", "before", "being",
|
|
4
|
+
"below", "between", "both", "but", "by", "can't", "cannot", "could", "couldn't",
|
|
5
|
+
"did", "didn't", "do", "does", "doesn't", "doing", "don't", "down", "during",
|
|
6
|
+
"each", "few", "for", "from", "further", "had", "hadn't", "has", "hasn't",
|
|
7
|
+
"have", "haven't", "having", "he", "he'd", "he'll", "he's", "her", "here",
|
|
8
|
+
"here's", "hers", "herself", "him", "himself", "his", "how", "how's", "i",
|
|
9
|
+
"i'd", "i'll", "i'm", "i've", "if", "in", "into", "is", "isn't", "it", "it's",
|
|
10
|
+
"its", "itself", "let's", "me", "more", "most", "mustn't", "my", "myself",
|
|
11
|
+
"no", "nor", "not", "of", "off", "on", "once", "only", "or", "other", "ought",
|
|
12
|
+
"our", "ours", "ourselves", "out", "over", "own", "same", "shan't", "she",
|
|
13
|
+
"she'd", "she'll", "she's", "should", "shouldn't", "so", "some", "such",
|
|
14
|
+
"than", "that", "that's", "the", "their", "theirs", "them", "themselves",
|
|
15
|
+
"then", "there", "there's", "these", "they", "they'd", "they'll", "they're",
|
|
16
|
+
"they've", "this", "those", "through", "to", "too", "under", "until", "up",
|
|
17
|
+
"very", "was", "wasn't", "we", "we'd", "we'll", "we're", "we've", "were",
|
|
18
|
+
"weren't", "what", "what's", "when", "when's", "where", "where's", "which",
|
|
19
|
+
"while", "who", "who's", "whom", "why", "why's", "with", "won't", "would",
|
|
20
|
+
"wouldn't", "you", "you'd", "you'll", "you're", "you've", "your", "yours",
|
|
21
|
+
"yourself", "yourselves"
|
|
22
|
+
]);
|
|
23
|
+
function tokenizeText(text, caseSensitive, removeStopwords) {
|
|
24
|
+
const clean = caseSensitive ? text : text.toLowerCase();
|
|
25
|
+
const words = clean
|
|
26
|
+
.replace(/[^\p{L}\p{N}\s_]/gu, " ")
|
|
27
|
+
.split(/\s+/)
|
|
28
|
+
.map((w) => w.trim())
|
|
29
|
+
.filter((w) => w.length > 0);
|
|
30
|
+
if (removeStopwords) {
|
|
31
|
+
return words.filter((w) => !DEFAULT_STOPWORDS.has(w.toLowerCase()));
|
|
32
|
+
}
|
|
33
|
+
return words;
|
|
34
|
+
}
|
|
35
|
+
function renderBar(pct, width = 15) {
|
|
36
|
+
const filled = Math.max(0, Math.min(width, Math.round((pct / 100) * width)));
|
|
37
|
+
const empty = width - filled;
|
|
38
|
+
return "█".repeat(filled) + "░".repeat(empty);
|
|
39
|
+
}
|
|
40
|
+
export async function computeNgrams(batchIterator, options = {}) {
|
|
41
|
+
const n = Math.max(1, options.n ?? 2);
|
|
42
|
+
const top = options.top ?? 20;
|
|
43
|
+
const minFreq = options.minFreq ?? 1;
|
|
44
|
+
const removeStopwords = options.stopwords ?? false;
|
|
45
|
+
const caseSensitive = options.caseSensitive ?? false;
|
|
46
|
+
let targetCol = options.col;
|
|
47
|
+
const freqMap = new Map();
|
|
48
|
+
let totalNgrams = 0;
|
|
49
|
+
for await (const batch of batchIterator) {
|
|
50
|
+
if (batch.rows.length === 0)
|
|
51
|
+
continue;
|
|
52
|
+
if (!targetCol) {
|
|
53
|
+
const firstRow = batch.rows[0];
|
|
54
|
+
const keys = Object.keys(firstRow);
|
|
55
|
+
targetCol =
|
|
56
|
+
keys.find((k) => /(text|title|name|description|body|review|content|message|summary)/i.test(k)) ??
|
|
57
|
+
keys.find((k) => typeof firstRow[k] === "string") ??
|
|
58
|
+
keys[0];
|
|
59
|
+
}
|
|
60
|
+
for (const row of batch.rows) {
|
|
61
|
+
const rawVal = row[targetCol];
|
|
62
|
+
if (rawVal === null || rawVal === undefined || rawVal === "")
|
|
63
|
+
continue;
|
|
64
|
+
const text = String(rawVal);
|
|
65
|
+
const tokens = tokenizeText(text, caseSensitive, removeStopwords);
|
|
66
|
+
if (tokens.length < n)
|
|
67
|
+
continue;
|
|
68
|
+
for (let i = 0; i <= tokens.length - n; i++) {
|
|
69
|
+
const gram = tokens.slice(i, i + n).join(" ");
|
|
70
|
+
freqMap.set(gram, (freqMap.get(gram) || 0) + 1);
|
|
71
|
+
totalNgrams++;
|
|
72
|
+
}
|
|
73
|
+
}
|
|
74
|
+
}
|
|
75
|
+
if (!targetCol) {
|
|
76
|
+
throw new Error("No text column found for N-gram extraction.");
|
|
77
|
+
}
|
|
78
|
+
// Filter and sort top results
|
|
79
|
+
const sorted = Array.from(freqMap.entries())
|
|
80
|
+
.filter(([, count]) => count >= minFreq)
|
|
81
|
+
.sort((a, b) => b[1] - a[1])
|
|
82
|
+
.slice(0, top);
|
|
83
|
+
const maxCount = sorted.length > 0 ? sorted[0][1] : 1;
|
|
84
|
+
const items = sorted.map(([ngram, count]) => {
|
|
85
|
+
const pct = totalNgrams > 0 ? (count / totalNgrams) * 100 : 0;
|
|
86
|
+
const relativePct = (count / maxCount) * 100;
|
|
87
|
+
return {
|
|
88
|
+
ngram,
|
|
89
|
+
count,
|
|
90
|
+
percentage: Math.round(pct * 100) / 100,
|
|
91
|
+
bar: renderBar(relativePct, 15),
|
|
92
|
+
};
|
|
93
|
+
});
|
|
94
|
+
const formattedRows = items.map((it, idx) => ({
|
|
95
|
+
rank: idx + 1,
|
|
96
|
+
ngram: it.ngram,
|
|
97
|
+
count: it.count,
|
|
98
|
+
share: `${it.percentage.toFixed(2)}%`,
|
|
99
|
+
distribution: it.bar,
|
|
100
|
+
}));
|
|
101
|
+
return {
|
|
102
|
+
n,
|
|
103
|
+
column: targetCol,
|
|
104
|
+
total_ngrams: totalNgrams,
|
|
105
|
+
unique_ngrams: freqMap.size,
|
|
106
|
+
items,
|
|
107
|
+
formattedRows,
|
|
108
|
+
};
|
|
109
|
+
}
|
|
110
|
+
//# sourceMappingURL=ngrams.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"ngrams.js","sourceRoot":"","sources":["../../src/analytics/ngrams.ts"],"names":[],"mappings":"AA2BA,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAC;IAChC,GAAG,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK;IAC5E,KAAK,EAAE,KAAK,EAAE,QAAQ,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO;IAC9E,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,UAAU;IAC/E,KAAK,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,QAAQ;IAC5E,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,SAAS,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,QAAQ;IACzE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM;IACzE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,KAAK,EAAE,SAAS,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,GAAG;IACzE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM;IAC7E,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,QAAQ;IACzE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO;IAC7E,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,QAAQ,EAAE,KAAK;IACzE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM;IACvE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,YAAY;IACxE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,SAAS;IAC3E,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI;IAC1E,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM;IACxE,SAAS,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,SAAS,EAAE,OAAO;IAC1E,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;IACzE,UAAU,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,EAAE,OAAO;IACzE,UAAU,EAAE,YAAY;CACzB,CAAC,CAAC;AAEH,SAAS,YAAY,CAAC,IAAY,EAAE,aAAsB,EAAE,eAAwB;IAClF,MAAM,KAAK,GAAG,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;IACxD,MAAM,KAAK,GAAG,KAAK;SAChB,OAAO,CAAC,oBAAoB,EAAE,GAAG,CAAC;SAClC,KAAK,CAAC,KAAK,CAAC;SACZ,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAE/B,IAAI,eAAe,EAAE,CAAC;QACpB,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,iBAAiB,CAAC,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IACtE,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,SAAS,CAAC,GAAW,EAAE,KAAK,GAAG,EAAE;IACxC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IAC7E,MAAM,KAAK,GAAG,KAAK,GAAG,MAAM,CAAC;IAC7B,OAAO,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AAChD,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CACjC,aAAuC,EACvC,UAAwB,EAAE;IAE1B,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACtC,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,IAAI,EAAE,CAAC;IAC9B,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,CAAC,CAAC;IACrC,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,IAAI,KAAK,CAAC;IACnD,MAAM,aAAa,GAAG,OAAO,CAAC,aAAa,IAAI,KAAK,CAAC;IAErD,IAAI,SAAS,GAAG,OAAO,CAAC,GAAG,CAAC;IAC5B,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,IAAI,WAAW,GAAG,CAAC,CAAC;IAEpB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;QACxC,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAEtC,IAAI,CAAC,SAAS,EAAE,CAAC;YACf,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;YAChC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YACnC,SAAS;gBACP,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,oEAAoE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;oBAC9F,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,QAAQ,CAAC,CAAC,CAAC,KAAK,QAAQ,CAAC;oBACjD,IAAI,CAAC,CAAC,CAAC,CAAC;QACZ,CAAC;QAED,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;YAC7B,MAAM,MAAM,GAAG,GAAG,CAAC,SAAU,CAAC,CAAC;YAC/B,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,EAAE;gBAAE,SAAS;YACvE,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;YAE5B,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,EAAE,aAAa,EAAE,eAAe,CAAC,CAAC;YAClE,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;gBAAE,SAAS;YAEhC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC5C,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC9C,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAChD,WAAW,EAAE,CAAC;YAChB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,CAAC,SAAS,EAAE,CAAC;QACf,MAAM,IAAI,KAAK,CAAC,6CAA6C,CAAC,CAAC;IACjE,CAAC;IAED,8BAA8B;IAC9B,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,CAAC;SACzC,MAAM,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,KAAK,IAAI,OAAO,CAAC;SACvC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;SAC3B,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAEjB,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAEvD,MAAM,KAAK,GAAyB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,CAAC,EAAE,EAAE;QAChE,MAAM,GAAG,GAAG,WAAW,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,WAAW,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9D,MAAM,WAAW,GAAG,CAAC,KAAK,GAAG,QAAQ,CAAC,GAAG,GAAG,CAAC;QAC7C,OAAO;YACL,KAAK;YACL,KAAK;YACL,UAAU,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,GAAG,CAAC,GAAG,GAAG;YACvC,GAAG,EAAE,SAAS,CAAC,WAAW,EAAE,EAAE,CAAC;SAChC,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,MAAM,aAAa,GAAU,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC;QACnD,IAAI,EAAE,GAAG,GAAG,CAAC;QACb,KAAK,EAAE,EAAE,CAAC,KAAK;QACf,KAAK,EAAE,EAAE,CAAC,KAAK;QACf,KAAK,EAAE,GAAG,EAAE,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG;QACrC,YAAY,EAAE,EAAE,CAAC,GAAG;KACrB,CAAC,CAAC,CAAC;IAEJ,OAAO;QACL,CAAC;QACD,MAAM,EAAE,SAAS;QACjB,YAAY,EAAE,WAAW;QACzB,aAAa,EAAE,OAAO,CAAC,IAAI;QAC3B,KAAK;QACL,aAAa;KACd,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,25 @@
|
|
|
1
|
+
import type { TransformFunction } from "../core/types.js";
|
|
2
|
+
export type OutlierMethod = "zscore" | "iqr" | "mad";
|
|
3
|
+
export interface OutlierOptions {
|
|
4
|
+
column: string;
|
|
5
|
+
method?: OutlierMethod;
|
|
6
|
+
threshold?: number;
|
|
7
|
+
onlyOutliers?: boolean;
|
|
8
|
+
addColumns?: boolean;
|
|
9
|
+
invert?: boolean;
|
|
10
|
+
}
|
|
11
|
+
export interface OutlierStats {
|
|
12
|
+
column: string;
|
|
13
|
+
method: OutlierMethod;
|
|
14
|
+
threshold: number;
|
|
15
|
+
totalRows: number;
|
|
16
|
+
outlierCount: number;
|
|
17
|
+
outlierPercent: number;
|
|
18
|
+
lowerBound?: number;
|
|
19
|
+
upperBound?: number;
|
|
20
|
+
}
|
|
21
|
+
/**
|
|
22
|
+
* Streaming Outlier & Anomaly Detection Transform.
|
|
23
|
+
*/
|
|
24
|
+
export declare function outliersTransform(options: OutlierOptions): TransformFunction;
|
|
25
|
+
//# sourceMappingURL=outliers.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"outliers.d.ts","sourceRoot":"","sources":["../../src/analytics/outliers.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAmB,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAE3E,MAAM,MAAM,aAAa,GAAG,QAAQ,GAAG,KAAK,GAAG,KAAK,CAAC;AAErD,MAAM,WAAW,cAAc;IAC7B,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,CAAC,EAAE,aAAa,CAAC;IACvB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,MAAM,CAAC,EAAE,OAAO,CAAC;CAClB;AAED,MAAM,WAAW,YAAY;IAC3B,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,aAAa,CAAC;IACtB,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;IAClB,YAAY,EAAE,MAAM,CAAC;IACrB,cAAc,EAAE,MAAM,CAAC;IACvB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,cAAc,GAAG,iBAAiB,CAmH5E"}
|