Feature/allow full compute asset

README.md

@@ -211,6 +211,17 @@ e.g.: `'[{"id":"cpu","amount":3},{"id":"ram","amount":16772672536},{"id":"disk",
 -  `--accept` option can be set to `true` or `false`. If it is set to `false` a prompt will be displayed to the user for manual accepting the payment before starting a compute job. If it is set to `true`, the compute job starts automatically, without user input.
 - `output` is an optional stringified JSON object specifying a remote storage backend where job results will be uploaded. Supported types include S3, FTP, URL, Arweave, and IPFS. If omitted, results are stored on the node's local disk. e.g.: `'{"remoteStorage":{"type":"s3","s3Access":{"endpoint":"https://s3.amazonaws.com","region":"us-east-1","bucket":"my-results","objectKey":"jobs/result.tar","accessKeyId":"AKIA...","secretAccessKey":"..."}}}'`
 
+**Raw (unpublished) datasets and algorithms:**
+
+Instead of a DID, you can pass a full `ComputeAsset` (datasets) or `ComputeAlgorithm` (algorithm) JSON object with a `fileObject`, to run compute on raw data/algorithms that are not published as assets. Raw entries have no DID and are not ordered (no datatoken). DID-based and raw entries can be mixed within the datasets argument (the value must then be valid JSON, with DIDs quoted). JSON must be single-quoted on the shell.
+
+- Raw algorithm against a published dataset that allows raw algorithms (`allowRawAlgorithm: true`):  
+  `npm run cli startCompute -- did:op:dataset '{"fileObject":{"type":"url","url":"https://example.com/algo.py","method":"GET"},"meta":{"container":{"entrypoint":"python $ALGO","image":"oceanprotocol/algo_dockers","tag":"python-branin","checksum":"sha256:..."}}}' env1 900 paymentToken resources --accept true`
+- Raw dataset(s) against a published algorithm:  
+  `npm run cli startCompute -- '[{"fileObject":{"type":"url","url":"https://example.com/data.csv","method":"GET"}}]' did:op:algo env1 900 paymentToken resources --accept true`
+- Mixed datasets (a published DID and a raw file):  
+  `npm run cli startCompute -- '["did:op:dataset",{"fileObject":{"type":"ipfs","hash":"Qm..."}}]' did:op:algo env1 900 paymentToken resources --accept true`
+
 ---
 
 **Start Free Compute:**
@@ -223,6 +234,8 @@ e.g.: `'[{"id":"cpu","amount":3},{"id":"ram","amount":16772672536},{"id":"disk",
   (Options can be provided in any order.)
 
 - `output` is an optional stringified JSON object specifying a remote storage backend where job results will be uploaded. Same format as `startCompute`.
+- Like `startCompute`, the datasets and algorithm arguments accept raw `ComputeAsset`/`ComputeAlgorithm` JSON objects with a `fileObject` (no DID), and mixed DID + raw datasets. e.g.:  
+  `npm run cli startFreeCompute did:op:dataset '{"fileObject":{"type":"url","url":"https://example.com/algo.py","method":"GET"},"meta":{"container":{"entrypoint":"python $ALGO","image":"oceanprotocol/algo_dockers","tag":"python-branin","checksum":"sha256:..."}}}' env1`
 
 ---
 

src/cli.ts

@@ -264,9 +264,12 @@ export async function createCLI() {
     .description("Starts a compute job")
     .argument(
       "<datasetDids>",
-      "Dataset DIDs (comma-separated) OR (empty array for none)"
+      "Dataset DIDs (comma-separated), an empty array for none, OR a JSON ComputeAsset object/array with a fileObject (raw datasets, no DID). Mixed input must be valid JSON, e.g. '[\"did:op:abc\",{\"fileObject\":{...}}]'"
+    )
+    .argument(
+      "<algoDid>",
+      "Algorithm DID, OR a JSON ComputeAlgorithm object with a fileObject and meta (raw algorithm, no DID)"
     )
-    .argument("<algoDid>", "Algorithm DID")
     .argument("<computeEnvId>", "Compute environment ID")
     .argument("<maxJobDuration>", "maxJobDuration for compute job")
     .argument("<paymentToken>", "Payment token for compute")
@@ -277,9 +280,12 @@ export async function createCLI() {
     )
     .option(
       "-d, --datasets <datasetDids>",
-      "Dataset DIDs (comma-separated) OR (empty array for none)"
+      "Dataset DIDs (comma-separated), an empty array for none, OR a JSON ComputeAsset object/array with a fileObject (raw datasets, no DID)"
+    )
+    .option(
+      "-a, --algo <algoDid>",
+      "Algorithm DID, OR a JSON ComputeAlgorithm object with a fileObject and meta (raw algorithm, no DID)"
     )
-    .option("-a, --algo <algoDid>", "Algorithm DID")
     .option("-e, --env <computeEnvId>", "Compute environment ID")
     .option("--maxJobDuration <maxJobDuration>", "Compute maxJobDuration")
     .option("-t, --token <paymentToken>", "Compute payment token")
@@ -384,19 +390,25 @@ export async function createCLI() {
     .description("Starts a FREE compute job")
     .argument(
       "<datasetDids>",
-      "Dataset DIDs (comma-separated) OR (empty array for none)"
+      "Dataset DIDs (comma-separated), an empty array for none, OR a JSON ComputeAsset object/array with a fileObject (raw datasets, no DID). Mixed input must be valid JSON, e.g. '[\"did:op:abc\",{\"fileObject\":{...}}]'"
+    )
+    .argument(
+      "<algoDid>",
+      "Algorithm DID, OR a JSON ComputeAlgorithm object with a fileObject and meta (raw algorithm, no DID)"
     )
-    .argument("<algoDid>", "Algorithm DID")
     .argument("<computeEnvId>", "Compute environment ID")
     .argument(
       "[output]",
       "Output backend to save job results to. Supported types include S3, FTP, URL, Arweave, etc. Defaults to node local disk if omitted."
     )
     .option(
       "-d, --datasets <datasetDids>",
-      "Dataset DIDs (comma-separated) OR (empty array for none)"
+      "Dataset DIDs (comma-separated), an empty array for none, OR a JSON ComputeAsset object/array with a fileObject (raw datasets, no DID)"
+    )
+    .option(
+      "-a, --algo <algoDid>",
+      "Algorithm DID, OR a JSON ComputeAlgorithm object with a fileObject and meta (raw algorithm, no DID)"
     )
-    .option("-a, --algo <algoDid>", "Algorithm DID")
     .option("-e, --env <computeEnvId>", "Compute environment ID")
     .option(
       "-o, --output [output]",