feat: add --data-set and --new-data-set flags to add command (#296)

* feat: add --data-set and --new-data-set flags to add command

* fix: mutual exclusivity problem

* chore: fix lint and use Number.isNaN

---------

Co-authored-by: Russell Dempsey <1173416+SgtPooki@users.noreply.github.com>

Author: beck-8

src/add/add.ts

@@ -181,6 +181,8 @@ export async function runAdd(options: AddOptions): Promise<AddResult> {
       ...providerOptions,
       dataset: {
         ...(dataSetMetadata && { metadata: dataSetMetadata }),
+        ...(options.dataSetId && { useExisting: options.dataSetId }),
+        ...(options.createNewDataSet && { createNew: true }),
       },
       callbacks: {
         onProviderSelected: (provider) => {

src/add/types.ts

@@ -4,6 +4,10 @@ import type { CLIAuthOptions } from '../utils/cli-auth.js'
 export interface AddOptions extends CLIAuthOptions {
   filePath: string
   bare?: boolean
+  /** ID of the existing data set to use */
+  dataSetId?: number
+  /** Create a new data set instead of using an existing one */
+  createNewDataSet?: boolean
   /** Auto-fund: automatically ensure minimum 30 days of runway */
   autoFund?: boolean
   /** Piece metadata attached to each upload */

src/commands/add.ts

@@ -1,4 +1,4 @@
-import { Command } from 'commander'
+import { Command, Option } from 'commander'
 import { runAdd } from '../add/add.js'
 import type { AddOptions } from '../add/types.js'
 import { MIN_RUNWAY_DAYS } from '../common/constants.js'
@@ -10,30 +10,59 @@ export const addCommand = new Command('add')
   .argument('<path>', 'Path to the file or directory to add')
   .option('--bare', 'Add file without directory wrapper (files only, not supported for directories)')
   .option('--auto-fund', `Automatically ensure minimum ${MIN_RUNWAY_DAYS} days of runway before upload`)
-  .action(async (path: string, options) => {
-    try {
-      const {
-        metadata: _metadata,
-        dataSetMetadata: _dataSetMetadata,
-        datasetMetadata: _datasetMetadata,
-        '8004Type': _erc8004Type,
-        '8004Agent': _erc8004Agent,
-        ...addOptionsFromCli
-      } = options
-      const { pieceMetadata, dataSetMetadata } = resolveMetadataOptions(options, { includeErc8004: true })
-
-      const addOptions: AddOptions = {
-        ...addOptionsFromCli,
-        filePath: path,
-        ...(pieceMetadata && { pieceMetadata }),
-        ...(dataSetMetadata && { dataSetMetadata }),
+
+// Add data set selection options
+addCommand.addOption(
+  new Option('--data-set <id>', 'ID of the existing data set to use').conflicts(['newDataSet', 'new-data-set'])
+)
+
+addCommand.addOption(
+  new Option('--new-data-set', 'Create a new data set instead of using an existing one').conflicts([
+    'dataSet',
+    'data-set',
+  ])
+)
+
+addCommand.action(async (path: string, options: any) => {
+  try {
+    const {
+      metadata: _metadata,
+      dataSetMetadata: _dataSetMetadata,
+      datasetMetadata: _datasetMetadata,
+      '8004Type': _erc8004Type,
+      '8004Agent': _erc8004Agent,
+      dataSet,
+      newDataSet,
+      ...addOptionsFromCli
+    } = options
+    const { pieceMetadata, dataSetMetadata } = resolveMetadataOptions(options, { includeErc8004: true })
+
+    // Normalize dataSet ID
+    const rawDataSetId = dataSet
+    let dataSetId: number | undefined
+
+    if (rawDataSetId) {
+      dataSetId = parseInt(rawDataSetId, 10)
+      if (Number.isNaN(dataSetId) || dataSetId < 0 || dataSetId.toString() !== rawDataSetId) {
+        console.error('Error: Data set ID must be a valid positive integer')
+        process.exit(1)
       }
+    }
 
-      await runAdd(addOptions)
-    } catch {
-      process.exit(1)
+    const addOptions: AddOptions = {
+      ...addOptionsFromCli,
+      filePath: path,
+      ...(dataSetId !== undefined && { dataSetId }),
+      ...(newDataSet && { createNewDataSet: true }),
+      ...(pieceMetadata && { pieceMetadata }),
+      ...(dataSetMetadata && { dataSetMetadata }),
     }
-  })
+
+    await runAdd(addOptions)
+  } catch {
+    process.exit(1)
+  }
+})
 
 addAuthOptions(addCommand)
 addProviderOptions(addCommand)

src/test/unit/add.test.ts

@@ -206,6 +206,7 @@ describe('Add Command', () => {
         dataSetMetadata: { purpose: 'erc8004' },
       })
       const { createStorageContext, initializeSynapse } = await import('../../core/synapse/index.js')
+      const { performUpload } = await import('../../common/upload-flow.js')
 
       expect(vi.mocked(initializeSynapse)).toHaveBeenCalledWith(
         expect.objectContaining({
@@ -224,7 +225,6 @@ describe('Add Command', () => {
         })
       )
 
-      const { performUpload } = await import('../../common/upload-flow.js')
       expect(vi.mocked(performUpload)).toHaveBeenCalledWith(
         expect.anything(),
         expect.anything(),
@@ -235,6 +235,39 @@ describe('Add Command', () => {
       )
     })
 
+    it('passes data set selection options to storage context', async () => {
+      await runAdd({
+        filePath: testFile,
+        privateKey: 'test-private-key',
+        rpcUrl: 'wss://test.rpc.url',
+        dataSetId: 123,
+      })
+      const { createStorageContext } = await import('../../core/synapse/index.js')
+      expect(vi.mocked(createStorageContext)).toHaveBeenCalledWith(
+        expect.anything(),
+        expect.objectContaining({
+          dataset: expect.objectContaining({
+            useExisting: 123,
+          }),
+        })
+      )
+
+      await runAdd({
+        filePath: testFile,
+        privateKey: 'test-private-key',
+        rpcUrl: 'wss://test.rpc.url',
+        createNewDataSet: true,
+      })
+      expect(vi.mocked(createStorageContext)).toHaveBeenCalledWith(
+        expect.anything(),
+        expect.objectContaining({
+          dataset: expect.objectContaining({
+            createNew: true,
+          }),
+        })
+      )
+    })
+
     it('should reject when file does not exist', async () => {
       const mockExit = vi.spyOn(process, 'exit')