Feat/cost attribution 868

CHANGELOG.md

@@ -9,6 +9,17 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- **Cost attribution by application tag or memo prefix** ([#868](https://github.com/Nanle-code/stellar-dev-dashboard/issues/868)).
+  Attributes transaction fees and asset transfer volumes to developer-defined project tags for project budgeting and threshold alerts.
+  - `src/lib/costThresholdManager.ts` — `CostThresholdManager` implementation for tag definition, memo prefix and regex pattern matching, budget limits, volume tracking, threshold alerts (`ok`, `warning`, `exceeded`), JSON/CSV report exports, and SSR/unsupported environment fallbacks.
+  - `src/lib/analytics.ts` — integrated `costAttribution` calculation into `buildAnalyticsSnapshot()` and re-exported `CostThresholdManager`.
+  - `docs/features/COST_ATTRIBUTION_ANALYTICS.md` — feature documentation, usage examples, compatibility, security, and migration notes.
+  - `src/lib/__tests__/costThresholdManager.test.ts` — 13 unit tests covering primary flow, boundary cases (0 fee, exact threshold %, 100% budget cap), failure paths (invalid regex, NaN/negative fees, null inputs), and SSR/storage error handling.
+- **Cohort retention views for account activity** ([#863](https://github.com/Nanle-code/stellar-dev-dashboard/issues/863)).
+  Provides cohort charts that group accounts by first-seen period (Day, Week, Month) and subsequent activity.
+  - `src/lib/cohortRetention.ts` — domain calculations for cohort retention matrices, period headers, summary stats, CSV/JSON exports, and graceful error handling.
+  - `src/components/dashboard/CohortRetentionView.tsx` — accessible UI view featuring cohort heatmap matrix table, line chart, controls, view mode toggles, and export triggers.
+  - `docs/features/COHORT_RETENTION_ANALYTICS.md` — comprehensive user-facing and developer documentation including compatibility, security, and migration guidance.
 - **Pre-sign risk summary** ([#982](https://github.com/Nanle-code/stellar-dev-dashboard/issues/982)).
   Every transaction is now parsed and described in plain language before it
   reaches a wallet, so an operation that irreversibly changes an account cannot

CUSTOM_METRIC_BUILDER_GUIDE.md

@@ -0,0 +1,259 @@
+# Custom Metric Builder Guide (#864)
+
+Compose reusable metrics from Horizon fields and saved arithmetic expressions.
+
+## Overview
+
+The custom metric builder lets you define formulas that combine any numeric
+Horizon API field with basic arithmetic (`+`, `-`, `*`, `/`) and
+parentheses. Formulas are validated at save time, stored locally, and can be
+evaluated on demand against any analytics snapshot.
+
+## Quick Start
+
+```typescript
+import {
+  saveMetricFormula,
+  evaluateSavedFormula,
+  METRIC_FIELD_CATALOGUE,
+} from './src/lib/metricBuilder';
+
+// 1. Browse available fields
+console.log(METRIC_FIELD_CATALOGUE.map(f => `${f.path} — ${f.label}`));
+
+// 2. Save a reusable formula
+saveMetricFormula({
+  id: 'fee-efficiency',
+  name: 'Fee Efficiency Ratio',
+  formula: 'network.baseFee / network.p90Fee',
+  format: 'percentage',
+  tags: ['fees', 'efficiency'],
+});
+
+// 3. Evaluate against a live snapshot
+const snapshot = buildAnalyticsSnapshot({ /* ... Horizon data ... */ });
+const result = evaluateSavedFormula('fee-efficiency', snapshot);
+// → { value: 0.5, formatted: "50.0%", resolvedFields: ["network.baseFee", "network.p90Fee"] }
+```
+
+## Available Horizon Fields
+
+The field catalogue (`METRIC_FIELD_CATALOGUE`) exposes these paths:
+
+### Account fields
+
+| Path | Label | Description |
+|------|-------|-------------|
+| `account.xlmBalance` | XLM Balance | Native XLM balance |
+| `account.trustlineCount` | Trustline Count | Non-native trustlines |
+| `account.totalAssets` | Total Assets | All balance entries including native |
+| `account.nonNativeBalanceCount` | Non-Native Funded | Non-native trustlines with positive balance |
+
+### Transaction fields
+
+| Path | Label | Description |
+|------|-------|-------------|
+| `transactions.totalTransactions` | Total Transactions | Total transaction count |
+| `transactions.successfulTransactions` | Successful Transactions | Count of successful transactions |
+| `transactions.failedTransactions` | Failed Transactions | Count of failed transactions |
+| `transactions.successRate` | Success Rate | Ratio (0–1) of successful to total |
+| `transactions.weeklyActivity` | Weekly Activity | Transaction count in the last 7 days |
+| `transactions.averageOperationsPerTx` | Avg Ops per Tx | Average operations per transaction |
+
+### Network fields
+
+| Path | Label | Description |
+|------|-------|-------------|
+| `network.latestLedgerSequence` | Latest Ledger | Most recent ledger sequence number |
+| `network.baseFee` | Base Fee | Last ledger base fee (stroops) |
+| `network.p90Fee` | P90 Fee | 90th percentile accepted fee (stroops) |
+| `network.txSuccessCount` | Ledger Tx Success Count | Successful txs in latest ledger |
+| `network.txFailedCount` | Ledger Tx Failed Count | Failed txs in latest ledger |
+| `network.operationCount` | Ledger Operation Count | Total operations in latest ledger |
+| `network.averageCloseSeconds` | Avg Close Time | Average ledger close time (seconds) |
+
+## Formula Syntax
+
+Formulas support:
+
+- **Field references**: dot-separated paths (e.g. `account.xlmBalance`)
+- **Numeric literals**: integers and decimals (e.g. `100`, `3.14`)
+- **Operators**: `+`, `-`, `*`, `/` with standard precedence
+- **Parentheses**: `(` `)` for grouping
+- **Unary negation**: `-field.path` or `-(expr)`
+
+### Examples
+
+```
+account.xlmBalance
+transactions.successfulTransactions / transactions.totalTransactions
+(network.baseFee + network.p90Fee) / 2
+account.xlmBalance - (account.trustlineCount * 0.5)
+-transactions.failedTransactions + transactions.successfulTransactions
+```
+
+### Limits
+
+| Limit | Value |
+|-------|-------|
+| Max formula length | 1,024 characters |
+| Max field depth | 5 segments |
+| Max tokens | 256 |
+| Max saved formulas | 100 |
+
+## Display Formats
+
+Each saved formula can specify a `format`:
+
+| Format | Example output | Use case |
+|--------|---------------|----------|
+| `number` | `1,234.56` | General numeric values |
+| `percentage` | `90.0%` | Ratios (value × 100) |
+| `stroops` | `100 stroops` | Raw Stellar fee units |
+| `xlm` | `1.0000000 XLM` | XLM amounts (stroops ÷ 10⁷) |
+
+## CRUD API
+
+```typescript
+import {
+  saveMetricFormula,
+  getMetricFormula,
+  updateMetricFormula,
+  deleteMetricFormula,
+  listMetricFormulas,
+  clearAllMetricFormulas,
+} from './src/lib/metricBuilder';
+
+// Save
+saveMetricFormula({ id: 'my-metric', name: 'My Metric', formula: '...' });
+
+// Get by id
+const formula = getMetricFormula('my-metric');
+
+// List (optionally filtered by tags)
+const all = listMetricFormulas();
+const feeMetrics = listMetricFormulas(['fees']);
+
+// Update
+updateMetricFormula('my-metric', { name: 'Updated Name', formula: '...' });
+
+// Delete
+deleteMetricFormula('my-metric');
+
+// Clear all (testing / reset)
+clearAllMetricFormulas();
+```
+
+## Batch Evaluation
+
+Evaluate multiple saved formulas against the same data snapshot. Failures
+are captured individually so one bad formula does not block the others:
+
+```typescript
+import { evaluateBatch } from './src/lib/metricBuilder';
+
+const results = evaluateBatch(
+  ['fee-ratio', 'success-pct', 'missing-formula'],
+  snapshot,
+);
+
+for (const [id, result] of results) {
+  if ('error' in result) {
+    console.warn(`${id} failed: ${result.error}`);
+  } else {
+    console.log(`${id} = ${result.formatted}`);
+  }
+}
+```
+
+## Import / Export
+
+Back up or share saved formulas across environments:
+
+```typescript
+import { exportMetricFormulas, importMetricFormulas } from './src/lib/metricBuilder';
+
+// Export
+const json = exportMetricFormulas();
+// → portable JSON array string
+
+// Import (merges with existing; overwrites on id collision)
+const { imported, skipped } = importMetricFormulas(json);
+```
+
+## Error Handling
+
+All public functions throw `MetricBuilderError` with a typed `code` property:
+
+| Code | When |
+|------|------|
+| `invalid_formula` | Expression cannot be parsed or is empty/too long |
+| `invalid_input` | Non-string formula, null data, bad id/name, non-numeric field |
+| `field_not_found` | A field path does not exist in the data snapshot |
+| `division_by_zero` | Division by zero encountered during evaluation |
+| `duplicate_id` | Attempting to save with an id that already exists |
+| `not_found` | Attempting to update/evaluate a formula that does not exist |
+| `limit_exceeded` | Exceeding the 100-formula storage limit |
+| `storage_unavailable` | Reserved for environments where localStorage is missing |
+
+```typescript
+import { MetricBuilderError } from './src/lib/metricBuilder';
+
+try {
+  evaluateFormula('account.missing / 0', snapshot);
+} catch (err) {
+  if (err instanceof MetricBuilderError) {
+    switch (err.code) {
+      case 'field_not_found': /* guide user to pick a valid field */ break;
+      case 'division_by_zero': /* show a user-friendly message */ break;
+      default: console.error(err.message);
+    }
+  }
+}
+```
+
+## Security Notes
+
+- **No `eval()`**: Formulas are parsed with a hand-written recursive-descent
+  parser. Arbitrary code execution is not possible.
+- **Field path validation**: Only alphanumeric characters, underscores, and
+  dots are allowed in identifiers. SQL injection, script injection, and
+  prototype pollution via field paths are blocked by the tokenizer.
+- **Storage isolation**: Saved formulas are persisted under a dedicated
+  localStorage key (`stellar-dev-dashboard-saved-metric-formulas`) and
+  scoped to the origin.
+- **Complexity limits**: Token count, formula length, and field depth are
+  capped to prevent denial-of-service via pathological expressions.
+
+## Compatibility & Migration
+
+- **Browser requirements**: Full functionality with any browser that
+  supports `localStorage`. Falls back to in-memory storage in SSR, Web
+  Workers, or privacy mode — formulas work but are not persisted across
+  page reloads.
+- **Data source compatibility**: Works with `buildAnalyticsSnapshot()` from
+  `analytics.ts` and `ReportDataSet` from `customReports.ts`.
+- **No breaking changes**: This is a new module. No existing API surfaces
+  or data schemas are modified.
+- **TypeScript**: Fully typed with exported types for `SavedMetricFormula`,
+  `MetricEvaluationResult`, `MetricBuilderError`, and `MetricField`.
+
+## Integration with Custom Reports
+
+The metric builder extends the `customReports.ts` infrastructure. You can
+use evaluated formulas as additional metric entries in report templates:
+
+```typescript
+import { evaluateFormula } from './src/lib/metricBuilder';
+import { transformReportData } from './src/lib/customReports';
+
+const reportData = transformReportData('account-activity', analyticsSnapshot);
+
+// Add a custom metric to the report
+const custom = evaluateFormula(
+  'transactions.successRate * 100',
+  analyticsSnapshot,
+  'number',
+);
+reportData.metrics.push({ label: 'Custom Success %', value: custom.formatted });
+```

PR_DESCRIPTION.md

@@ -0,0 +1,29 @@
+## Summary
+
+Implements cost attribution by application tag or memo prefix for project budgeting (#868). Attributes transaction fees (in Stroops and XLM) and payment transfer volumes to developer-defined project tags (e.g. `billing`, `auth`, `nft-drop`, `defi-swap`) via literal memo prefixes (e.g., `[APP:billing]`, `BILL:`) or custom regex patterns.
+
+- **Domain Library (`src/lib/costThresholdManager.ts`)**: `CostThresholdManager` implementation for tag definition, memo prefix and regex pattern matching, budget limits, volume tracking, threshold alerts (`ok`, `warning` >= 80%, `exceeded` >= 100%, `unbudgeted`), auto-discovery of embedded tags (`[APP:tag]`), JSON/CSV report exports, and graceful SSR/storage error handling.
+- **Analytics Integration (`src/lib/analytics.ts`)**: Embedded cost attribution calculations into `buildAnalyticsSnapshot()` and re-exported `CostThresholdManager`.
+- **Documentation**: User-facing & developer guide in [`docs/features/COST_ATTRIBUTION_ANALYTICS.md`](docs/features/COST_ATTRIBUTION_ANALYTICS.md) and changelog entry in `CHANGELOG.md`.
+
+Closes #868
+
+## How was this tested?
+
+Ran unit tests:
+`pnpm run test:unit src/lib/__tests__/costThresholdManager.test.ts` (13 passed)
+
+- **Primary flow**: Verified adding/updating rules, attributing fees and payment volumes to configured tags by memo prefix, auto-discovering unbudgeted tag prefixes, calculating fee sums in Stroops and XLM, average fee calculations, and JSON/CSV report exporting.
+- **Boundary cases**: Verified empty transaction list handling, exact warning threshold calculation (80.0%), exact 100% budget cap exceedance, zero-fee transactions, and case-insensitive memo matching.
+- **Failure cases**: Verified `null` / `undefined` / non-array transaction inputs, malformed transaction objects with invalid fee strings (`"invalid_number"`, negative fees), malformed regex rules (unclosed brackets), throwing `localStorage` security/quota errors in restricted SSR/iframe sandboxes, and missing required rule properties.
+
+## Merge requirements
+
+A PR is merged only when **every** box below is true. See
+[Merge requirements](https://github.com/Nanle-code/stellar-dev-dashboard/blob/master/docs/contributing.md#merge-requirements) for the full policy.
+
+- [x] All required CI checks pass on the latest commit (not just an earlier push).
+- [x] No required checks are failing, pending, or skipped — re-run or fix them; do not ask for a merge while any are outstanding.
+- [x] The branch has no merge conflicts with the target branch (rebase or merge `master` if GitHub shows "This branch has conflicts").
+- [x] Tests were added or updated for the change (primary flow, a boundary case, and a failure case).
+- [x] Docs were updated where behaviour, configuration, or security posture changed.

README.md

@@ -260,6 +260,7 @@ This keeps user-specific and operational endpoints behind explicit authenticatio
 - **Scheduled report delivery via webhooks (#869)** — authenticated HMAC/bearer delivery of analytics summaries with retries: [docs/features/report-webhook-delivery.md](docs/features/report-webhook-delivery.md).
 - **Transaction Builder i18n (#878)** — complete locale coverage of builder strings across all nine languages: [docs/features/builder-i18n.md](docs/features/builder-i18n.md).
 - **Mutation testing gate for fee math (#895)** — Stryker score gate on stroop conversion and fee estimation: [docs/features/mutation-testing-gate.md](docs/features/mutation-testing-gate.md). Run locally with `pnpm run test:mutation:feemath`.
+- **Custom metric builder (#864)** — compose reusable metrics from Horizon fields and saved arithmetic formulas: [CUSTOM_METRIC_BUILDER_GUIDE.md](CUSTOM_METRIC_BUILDER_GUIDE.md).
 
 ## Canary Deployment Health Probes