@eigenpal/cli 0.6.14 → 0.6.16
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/CHANGELOG.md +18 -0
- package/README.md +1 -1
- package/dist/cli.js +9865 -9748
- package/dist/skill/SKILL.md +13 -13
- package/dist/skill/reference/cli/agents.md +0 -26
- package/dist/skill/reference/cli/init.md +1 -1
- package/dist/skill/reference/cli/rerun.md +20 -0
- package/dist/skill/reference/cli/run.md +23 -0
- package/dist/skill/reference/cli/runs.md +0 -20
- package/dist/skill/reference/cli/surface.md +2 -3
- package/dist/skill/reference/cli/workflow.md +18 -40
- package/dist/skill/reference/debugging.md +2 -2
- package/dist/skill/reference/step-exec.md +2 -2
- package/dist/skill/reference/step-types.md +4 -23
- package/dist/skill/reference/workflow-yaml.md +2 -2
- package/dist/templates/blank/README.md +1 -1
- package/dist/templates/pdf-extraction/README.md +1 -1
- package/dist/templates/text-classification/README.md +1 -1
- package/package.json +1 -1
package/dist/skill/SKILL.md
CHANGED
|
@@ -25,7 +25,7 @@ interactive users; understand them if you see them, but do not prefer them:
|
|
|
25
25
|
|
|
26
26
|
- `runs fb` = `runs feedback`
|
|
27
27
|
- `runs artifact` = `runs artifacts`
|
|
28
|
-
- `
|
|
28
|
+
- `run` = start a workflow or agent run
|
|
29
29
|
- `workflow exp` = `workflow experiment`
|
|
30
30
|
- `ls` = `list` anywhere in the CLI
|
|
31
31
|
- `diff` = `compare` anywhere in the CLI
|
|
@@ -40,7 +40,7 @@ For exact flags and defaults, prefer the generated references:
|
|
|
40
40
|
## Pick The Surface
|
|
41
41
|
|
|
42
42
|
Choose `workflow` when the work is about a workflow definition, dataset,
|
|
43
|
-
evaluator, experiment
|
|
43
|
+
evaluator, or experiment. Choose root `run` to create workflow runs from local examples.
|
|
44
44
|
|
|
45
45
|
Choose `agents` when the work is about an agent workspace, agent file, trigger,
|
|
46
46
|
or starting an agent run.
|
|
@@ -70,7 +70,7 @@ eigenpal workflow validate ./workflow.yaml
|
|
|
70
70
|
eigenpal workflow push --file workflow.yaml
|
|
71
71
|
|
|
72
72
|
# 4. Run one example while iterating.
|
|
73
|
-
eigenpal
|
|
73
|
+
eigenpal run workflows.<workflow-id> --example <example-name>
|
|
74
74
|
eigenpal runs watch <execution-id>
|
|
75
75
|
eigenpal runs get <execution-id> --json
|
|
76
76
|
|
|
@@ -92,7 +92,7 @@ also accept the workflow slug for convenience, but use the id in scripts.
|
|
|
92
92
|
Useful workflow ids are not interchangeable:
|
|
93
93
|
|
|
94
94
|
- `<workflow-id>` identifies a workflow (`wf_...`).
|
|
95
|
-
- `<execution-id>` identifies one workflow
|
|
95
|
+
- `<execution-id>` identifies one workflow run.
|
|
96
96
|
- `<batch-id>` identifies one experiment batch.
|
|
97
97
|
- `<example-id>` identifies one dataset example.
|
|
98
98
|
|
|
@@ -165,8 +165,8 @@ eigenpal agents save -m "Improve <agent> behavior"
|
|
|
165
165
|
|
|
166
166
|
# 5. Test the exact saved commit.
|
|
167
167
|
SOURCE_REF="$(eigenpal git -- rev-parse HEAD)"
|
|
168
|
-
eigenpal
|
|
169
|
-
eigenpal
|
|
168
|
+
eigenpal run agents.<slug>@"$SOURCE_REF" --input-json '{"text":"hello"}' --wait
|
|
169
|
+
eigenpal run agents.<slug>@"$SOURCE_REF" --example example-name --wait
|
|
170
170
|
eigenpal runs get <run-id> --json | jq '.run | {status,schemaValid,error,requestedSourceRef,resolvedGitRef,resolvedGitTag,resolvedCommitSha,cost}'
|
|
171
171
|
eigenpal runs artifacts list <run-id>
|
|
172
172
|
eigenpal runs trace <run-id> --out ./review/<run-id>/trace.jsonl
|
|
@@ -176,18 +176,18 @@ eigenpal agents release patch agents/<slug>
|
|
|
176
176
|
eigenpal agents sync agents.<slug>
|
|
177
177
|
|
|
178
178
|
# 7. Verify the released package.
|
|
179
|
-
eigenpal
|
|
179
|
+
eigenpal run agents.<slug> --example example-name --wait --json
|
|
180
180
|
eigenpal runs get <released-run-id> --json | jq '.run | {status,schemaValid,error,resolvedGitTag,resolvedCommitSha,cost}'
|
|
181
181
|
```
|
|
182
182
|
|
|
183
183
|
File inputs must use schema field names when there is more than one file input:
|
|
184
184
|
|
|
185
185
|
```bash
|
|
186
|
-
eigenpal
|
|
186
|
+
eigenpal run agents.<slug> \
|
|
187
187
|
--input-file start_info_doc=./ZFRP.docx \
|
|
188
188
|
--wait
|
|
189
189
|
|
|
190
|
-
eigenpal
|
|
190
|
+
eigenpal run agents.<slug> \
|
|
191
191
|
--input-file zfzal_doc=./ZFZAL.pdf \
|
|
192
192
|
--input-file lv_doc=./list_vlastnictva.pdf \
|
|
193
193
|
--input-json '{"skip_name_check":false}' \
|
|
@@ -259,7 +259,7 @@ For `expected.json`, include only fields that must be identical on every correct
|
|
|
259
259
|
run. Omit non-deterministic values such as timestamps, random IDs, and free-form
|
|
260
260
|
LLM text. Validate the actual agent output with the output schema; validate the
|
|
261
261
|
expected file as a partial assertion.
|
|
262
|
-
Run a persisted example with `eigenpal
|
|
262
|
+
Run a persisted example with `eigenpal run agents.<slug>@<ref> --example <name> --wait`.
|
|
263
263
|
|
|
264
264
|
### Debug Agent Runs
|
|
265
265
|
|
|
@@ -357,7 +357,7 @@ eigenpal workflow dataset example delete <workflow-id> <example-id> --yes
|
|
|
357
357
|
### Run And Compare Workflow Executions
|
|
358
358
|
|
|
359
359
|
```bash
|
|
360
|
-
eigenpal
|
|
360
|
+
eigenpal run workflows.<workflow-id> --example <example-name>
|
|
361
361
|
eigenpal runs watch <execution-id>
|
|
362
362
|
eigenpal runs get <execution-id> --include input,output,error --json
|
|
363
363
|
eigenpal runs list <workflow-id> --type workflow --status failed --limit 10
|
|
@@ -491,7 +491,7 @@ Workflow schemas:
|
|
|
491
491
|
- [`reference/step-exec.md`](reference/step-exec.md) — single-step execution command
|
|
492
492
|
- [`reference/dataset-format.md`](reference/dataset-format.md) — dataset folder format
|
|
493
493
|
- [`reference/evaluators.md`](reference/evaluators.md) — evaluator configuration
|
|
494
|
-
- [`reference/debugging.md`](reference/debugging.md) — workflow
|
|
494
|
+
- [`reference/debugging.md`](reference/debugging.md) — workflow run debugging
|
|
495
495
|
|
|
496
496
|
CLI command references:
|
|
497
497
|
|
|
@@ -568,7 +568,7 @@ Agent command roles:
|
|
|
568
568
|
- `agents save` — validate, commit if dirty, and push the current branch.
|
|
569
569
|
- `agents release` — create immutable package release tags.
|
|
570
570
|
- `agents sync` — apply latest released agent manifest/triggers to the server.
|
|
571
|
-
- `
|
|
571
|
+
- `run` — start a workflow or agent run from a target ref (`@latest`, version, branch, commit).
|
|
572
572
|
- `runs rerun` — reuse a previous input snapshot; defaults to latest source.
|
|
573
573
|
- `runs list/get/watch/cancel/compare/trace` — inspect and manage executions.
|
|
574
574
|
- `runs artifacts list/fetch` — inventory and download run artifacts by path.
|
|
@@ -15,7 +15,6 @@ Manage Eigenpal agents: Git source, datasets, runs, experiments, sessions, and r
|
|
|
15
15
|
- [Env](#env)
|
|
16
16
|
- [Secrets](#secrets)
|
|
17
17
|
- [Details](#details)
|
|
18
|
-
- [`eigenpal agents run [options] <target>`](#eigenpal-agents-run-options-target)
|
|
19
18
|
- [`eigenpal agents list|ls [options]`](#eigenpal-agents-listls-options)
|
|
20
19
|
- [`eigenpal agents validate [options] [dir]`](#eigenpal-agents-validate-options-dir)
|
|
21
20
|
- [`eigenpal agents clone [options]`](#eigenpal-agents-clone-options)
|
|
@@ -62,7 +61,6 @@ Manage Eigenpal agents: Git source, datasets, runs, experiments, sessions, and r
|
|
|
62
61
|
|
|
63
62
|
```
|
|
64
63
|
agents
|
|
65
|
-
├── run <target>
|
|
66
64
|
├── list|ls
|
|
67
65
|
├── file
|
|
68
66
|
│ ├── list|ls <agent-id-or-slug>
|
|
@@ -119,7 +117,6 @@ agents
|
|
|
119
117
|
|
|
120
118
|
| Command | Description |
|
|
121
119
|
| --------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
|
|
122
|
-
| `eigenpal agents run [options] <target>` | Run an agent target, e.g. agents.invoice-agent@latest. |
|
|
123
120
|
| `eigenpal agents list\|ls [options]` | List agents. |
|
|
124
121
|
| `eigenpal agents validate [options] [dir]` | Validate a local agent package (layout, manifest, schemas, and Git source rules). |
|
|
125
122
|
| `eigenpal agents clone [options]` | Clone the organization source repository. |
|
|
@@ -199,29 +196,6 @@ agents
|
|
|
199
196
|
|
|
200
197
|
## Details
|
|
201
198
|
|
|
202
|
-
### `eigenpal agents run [options] <target>`
|
|
203
|
-
|
|
204
|
-
Run an agent target, e.g. agents.invoice-agent@latest.
|
|
205
|
-
|
|
206
|
-
### Arguments
|
|
207
|
-
|
|
208
|
-
| Name | Required | Variadic | Description |
|
|
209
|
-
| -------- | -------- | -------- | ----------- |
|
|
210
|
-
| `target` | yes | no | |
|
|
211
|
-
|
|
212
|
-
### Options
|
|
213
|
-
|
|
214
|
-
| Flag | Required | Default | Description |
|
|
215
|
-
| --------------------------- | -------- | ------- | ---------------------------------------------------------------------------------------------------- |
|
|
216
|
-
| `--base-url <url>` | no | | Server base URL |
|
|
217
|
-
| `--json` | no | | Output the raw server response as JSON |
|
|
218
|
-
| `--input-json <json>` | no | | JSON input object |
|
|
219
|
-
| `--input-file <field=path>` | no | `[]` | Input file to upload as multipart form-data. Repeat for multiple files; bare paths use field "file". |
|
|
220
|
-
| `--example <name>` | no | | Run a persisted dataset example by name |
|
|
221
|
-
| `--wait` | no | | Poll until the run reaches a terminal status |
|
|
222
|
-
| `--interval <seconds>` | no | `2` | Polling interval in seconds |
|
|
223
|
-
| `--max-wait <seconds>` | no | `1800` | Maximum wait before exiting 2 |
|
|
224
|
-
|
|
225
199
|
### `eigenpal agents list|ls [options]`
|
|
226
200
|
|
|
227
201
|
List agents.
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
# eigenpal init
|
|
2
2
|
|
|
3
|
-
Scaffold a new workflow project. Without `[name]`, scaffolds into the current directory using the cwd basename as the workflow name. With `[name]`, creates `./<name>/` and uses that as the slug. The flat layout matches what `
|
|
3
|
+
Scaffold a new workflow project. Without `[name]`, scaffolds into the current directory using the cwd basename as the workflow name. With `[name]`, creates `./<name>/` and uses that as the slug. The flat layout matches what `run workflows.<slug>` already discovers — no manual file moves.
|
|
4
4
|
|
|
5
5
|
## Contents
|
|
6
6
|
|
|
@@ -0,0 +1,20 @@
|
|
|
1
|
+
# eigenpal rerun
|
|
2
|
+
|
|
3
|
+
Create a new run from a previous run's stored input snapshot.
|
|
4
|
+
|
|
5
|
+
### Arguments
|
|
6
|
+
|
|
7
|
+
| Name | Required | Variadic | Description |
|
|
8
|
+
| -------- | -------- | -------- | ----------- |
|
|
9
|
+
| `run-id` | yes | no | |
|
|
10
|
+
|
|
11
|
+
### Options
|
|
12
|
+
|
|
13
|
+
| Flag | Required | Default | Description |
|
|
14
|
+
| ---------------------- | -------- | ---------- | ------------------------------------------------------------------------ |
|
|
15
|
+
| `--base-url <url>` | no | | Server base URL |
|
|
16
|
+
| `--json` | no | | Output the raw server response as JSON |
|
|
17
|
+
| `--version <version>` | no | `"latest"` | Version/source ref for the new run: latest, original, or an explicit ref |
|
|
18
|
+
| `--wait` | no | | Poll until the rerun reaches a terminal status |
|
|
19
|
+
| `--interval <seconds>` | no | `2` | Polling interval in seconds |
|
|
20
|
+
| `--max-wait <seconds>` | no | `1800` | Maximum wait before exiting 2 |
|
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
# eigenpal run
|
|
2
|
+
|
|
3
|
+
Start a workflow or agent run, e.g. workflows.extract-invoice.
|
|
4
|
+
|
|
5
|
+
### Arguments
|
|
6
|
+
|
|
7
|
+
| Name | Required | Variadic | Description |
|
|
8
|
+
| -------- | -------- | -------- | ----------- |
|
|
9
|
+
| `target` | yes | no | |
|
|
10
|
+
|
|
11
|
+
### Options
|
|
12
|
+
|
|
13
|
+
| Flag | Required | Default | Description |
|
|
14
|
+
| --------------------------- | -------- | ------- | ---------------------------------------------------------------------------------------------------- |
|
|
15
|
+
| `--base-url <url>` | no | | Server base URL |
|
|
16
|
+
| `--json` | no | | Output the raw server response as JSON |
|
|
17
|
+
| `--input-json <json>` | no | | JSON input object |
|
|
18
|
+
| `--input-file <field=path>` | no | `[]` | Input file to upload as multipart form-data. Repeat for multiple files; bare paths use field "file". |
|
|
19
|
+
| `--example <name>` | no | | Run one persisted example by name |
|
|
20
|
+
| `--dir <dir>` | no | | Local eigenpal directory for workflow examples |
|
|
21
|
+
| `--wait` | no | | Poll until the run reaches a terminal status |
|
|
22
|
+
| `--interval <seconds>` | no | `2` | Polling interval in seconds |
|
|
23
|
+
| `--max-wait <seconds>` | no | `1800` | Maximum wait before exiting 2 |
|
|
@@ -17,7 +17,6 @@ Inspect, watch, and manage workflow, agent, and eval runs.
|
|
|
17
17
|
- [`eigenpal runs rerun [options] <run-id>`](#eigenpal-runs-rerun-options-run-id)
|
|
18
18
|
- [`eigenpal runs trace [options] <run-id>`](#eigenpal-runs-trace-options-run-id)
|
|
19
19
|
- [`eigenpal runs watch [options] <run-id>`](#eigenpal-runs-watch-options-run-id)
|
|
20
|
-
- [`eigenpal runs resume [options] <run-id>`](#eigenpal-runs-resume-options-run-id)
|
|
21
20
|
- [`eigenpal runs cancel [options] <run-id>`](#eigenpal-runs-cancel-options-run-id)
|
|
22
21
|
- [`eigenpal runs artifacts|artifact list|ls [options] <run-id>`](#eigenpal-runs-artifactsartifact-listls-options-run-id)
|
|
23
22
|
- [`eigenpal runs artifacts|artifact fetch [options] <run-id>`](#eigenpal-runs-artifactsartifact-fetch-options-run-id)
|
|
@@ -55,7 +54,6 @@ runs
|
|
|
55
54
|
│ ├── rename <run-id> <old-name> <new-name>
|
|
56
55
|
│ └── delete <run-id> <name>
|
|
57
56
|
├── watch <run-id>
|
|
58
|
-
├── resume <run-id>
|
|
59
57
|
└── cancel <run-id>
|
|
60
58
|
```
|
|
61
59
|
|
|
@@ -71,7 +69,6 @@ runs
|
|
|
71
69
|
| `eigenpal runs rerun [options] <run-id>` | Create a new run from a previous run's stored input snapshot. |
|
|
72
70
|
| `eigenpal runs trace [options] <run-id>` | Print raw trace.jsonl for a run, or write it with --out. |
|
|
73
71
|
| `eigenpal runs watch [options] <run-id>` | Watch a run until it reaches a terminal status. |
|
|
74
|
-
| `eigenpal runs resume [options] <run-id>` | Resume a workflow run that is waiting for approval. |
|
|
75
72
|
| `eigenpal runs cancel [options] <run-id>` | Cancel a run. |
|
|
76
73
|
|
|
77
74
|
### Artifacts
|
|
@@ -235,23 +232,6 @@ Watch a run until it reaches a terminal status.
|
|
|
235
232
|
| `--interval <seconds>` | no | `2` | Polling interval in seconds |
|
|
236
233
|
| `--max-wait <seconds>` | no | `1800` | Maximum wait before exiting 2 |
|
|
237
234
|
|
|
238
|
-
### `eigenpal runs resume [options] <run-id>`
|
|
239
|
-
|
|
240
|
-
Resume a workflow run that is waiting for approval.
|
|
241
|
-
|
|
242
|
-
### Arguments
|
|
243
|
-
|
|
244
|
-
| Name | Required | Variadic | Description |
|
|
245
|
-
| -------- | -------- | -------- | ----------- |
|
|
246
|
-
| `run-id` | yes | no | |
|
|
247
|
-
|
|
248
|
-
### Options
|
|
249
|
-
|
|
250
|
-
| Flag | Required | Default | Description |
|
|
251
|
-
| ------------------ | -------- | ------- | -------------------------------------- |
|
|
252
|
-
| `--base-url <url>` | no | | Server base URL |
|
|
253
|
-
| `--json` | no | | Output the raw server response as JSON |
|
|
254
|
-
|
|
255
235
|
### `eigenpal runs cancel [options] <run-id>`
|
|
256
236
|
|
|
257
237
|
Cancel a run.
|
|
@@ -39,7 +39,6 @@ eigenpal
|
|
|
39
39
|
│ │ ├── results <workflow-id> [batchId]
|
|
40
40
|
│ │ ├── compare|diff <batchIdA> <batchIdB>
|
|
41
41
|
│ │ └── watch <workflow-id> <batchId>
|
|
42
|
-
│ ├── run <workflow-id> [examples...]
|
|
43
42
|
│ ├── versions
|
|
44
43
|
│ │ ├── list|ls <workflow-id>
|
|
45
44
|
│ │ └── restore <workflow-id> <versionId>
|
|
@@ -54,7 +53,6 @@ eigenpal
|
|
|
54
53
|
│ └── step
|
|
55
54
|
│ └── exec <type>
|
|
56
55
|
├── agents
|
|
57
|
-
│ ├── run <target>
|
|
58
56
|
│ ├── list|ls
|
|
59
57
|
│ ├── file
|
|
60
58
|
│ │ ├── list|ls <agent-id-or-slug>
|
|
@@ -103,6 +101,8 @@ eigenpal
|
|
|
103
101
|
│ │ └── pull [target]
|
|
104
102
|
│ └── secrets
|
|
105
103
|
│ └── export [target]
|
|
104
|
+
├── run <target>
|
|
105
|
+
├── rerun <run-id>
|
|
106
106
|
├── runs
|
|
107
107
|
│ ├── list|ls [source]
|
|
108
108
|
│ ├── get <run-id>
|
|
@@ -124,7 +124,6 @@ eigenpal
|
|
|
124
124
|
│ │ ├── rename <run-id> <old-name> <new-name>
|
|
125
125
|
│ │ └── delete <run-id> <name>
|
|
126
126
|
│ ├── watch <run-id>
|
|
127
|
-
│ ├── resume <run-id>
|
|
128
127
|
│ └── cancel <run-id>
|
|
129
128
|
├── git [gitArgs...]
|
|
130
129
|
├── completion <shell>
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
# eigenpal workflow
|
|
2
2
|
|
|
3
|
-
Manage workflows: push, pull,
|
|
3
|
+
Manage workflows: push, pull, and evaluate.
|
|
4
4
|
|
|
5
5
|
## Contents
|
|
6
6
|
|
|
@@ -19,7 +19,6 @@ Manage workflows: push, pull, run, evaluate.
|
|
|
19
19
|
- [`eigenpal workflow pull [options] <workflow-id>`](#eigenpal-workflow-pull-options-workflow-id)
|
|
20
20
|
- [`eigenpal workflow push [options]`](#eigenpal-workflow-push-options)
|
|
21
21
|
- [`eigenpal workflow move [options] <workflow-id>`](#eigenpal-workflow-move-options-workflow-id)
|
|
22
|
-
- [`eigenpal workflow run [options] <workflow-id> [examples...]`](#eigenpal-workflow-run-options-workflow-id-examples)
|
|
23
22
|
- [`eigenpal workflow validate [options] [path]`](#eigenpal-workflow-validate-options-path)
|
|
24
23
|
- [`eigenpal workflow clear-local [options] [examples...]`](#eigenpal-workflow-clear-local-options-examples)
|
|
25
24
|
- [`eigenpal workflow evaluators pull [options] <workflow-id>`](#eigenpal-workflow-evaluators-pull-options-workflow-id)
|
|
@@ -78,7 +77,6 @@ workflow
|
|
|
78
77
|
│ ├── results <workflow-id> [batchId]
|
|
79
78
|
│ ├── compare|diff <batchIdA> <batchIdB>
|
|
80
79
|
│ └── watch <workflow-id> <batchId>
|
|
81
|
-
├── run <workflow-id> [examples...]
|
|
82
80
|
├── versions
|
|
83
81
|
│ ├── list|ls <workflow-id>
|
|
84
82
|
│ └── restore <workflow-id> <versionId>
|
|
@@ -98,15 +96,14 @@ workflow
|
|
|
98
96
|
|
|
99
97
|
### Core
|
|
100
98
|
|
|
101
|
-
| Command
|
|
102
|
-
|
|
|
103
|
-
| `eigenpal workflow list\|ls [options]`
|
|
104
|
-
| `eigenpal workflow pull [options] <workflow-id>`
|
|
105
|
-
| `eigenpal workflow push [options]`
|
|
106
|
-
| `eigenpal workflow move [options] <workflow-id>`
|
|
107
|
-
| `eigenpal workflow
|
|
108
|
-
| `eigenpal workflow
|
|
109
|
-
| `eigenpal workflow clear-local [options] [examples...]` | Delete local execution artifacts under ./dataset/examples/. Keeps the latest run per example by default. |
|
|
99
|
+
| Command | Description |
|
|
100
|
+
| ------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
|
101
|
+
| `eigenpal workflow list\|ls [options]` | List workflows the caller can read. |
|
|
102
|
+
| `eigenpal workflow pull [options] <workflow-id>` | Download the YAML definition of the workflow at its current version. |
|
|
103
|
+
| `eigenpal workflow push [options]` | Create or update a workflow from a YAML file. |
|
|
104
|
+
| `eigenpal workflow move [options] <workflow-id>` | Move a workflow to a folder path, creating folders as needed |
|
|
105
|
+
| `eigenpal workflow validate [options] [path]` | Local-only validation. Without [path]: runs the templated three-way check (./workflow.yaml + ./evaluators.yaml + ./dataset/). With [path] pointing at a YAML file: validates just that workflow.yaml. For per-noun targeting use `evaluators validate` or `dataset validate`. |
|
|
106
|
+
| `eigenpal workflow clear-local [options] [examples...]` | Delete local execution artifacts under ./dataset/examples/. Keeps the latest run per example by default. |
|
|
110
107
|
|
|
111
108
|
### Evaluators
|
|
112
109
|
|
|
@@ -164,9 +161,9 @@ workflow
|
|
|
164
161
|
|
|
165
162
|
### Step
|
|
166
163
|
|
|
167
|
-
| Command | Description
|
|
168
|
-
| ---------------------------------------------- |
|
|
169
|
-
| `eigenpal workflow step exec [options] <type>` | DISABLED — local mimic runners removed pending server-side redesign (EIG-104). Use `
|
|
164
|
+
| Command | Description |
|
|
165
|
+
| ---------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |
|
|
166
|
+
| `eigenpal workflow step exec [options] <type>` | DISABLED — local mimic runners removed pending server-side redesign (EIG-104). Use `run` or `workflow experiment run` instead. |
|
|
170
167
|
|
|
171
168
|
## Details
|
|
172
169
|
|
|
@@ -234,26 +231,6 @@ Move a workflow to a folder path, creating folders as needed
|
|
|
234
231
|
| `--base-url <url>` | no | | Server base URL |
|
|
235
232
|
| `--json` | no | | Output the raw server response as JSON |
|
|
236
233
|
|
|
237
|
-
### `eigenpal workflow run [options] <workflow-id> [examples...]`
|
|
238
|
-
|
|
239
|
-
Run a saved workflow against local dataset examples.
|
|
240
|
-
|
|
241
|
-
### Arguments
|
|
242
|
-
|
|
243
|
-
| Name | Required | Variadic | Description |
|
|
244
|
-
| ------------- | -------- | -------- | ----------- |
|
|
245
|
-
| `workflow-id` | yes | no | |
|
|
246
|
-
| `examples` | no | yes | |
|
|
247
|
-
|
|
248
|
-
### Options
|
|
249
|
-
|
|
250
|
-
| Flag | Required | Default | Description |
|
|
251
|
-
| ------------------- | -------- | ------- | -------------------------------------------- |
|
|
252
|
-
| `--dir <dir>` | no | | Local eigenpal directory |
|
|
253
|
-
| `--concurrency <n>` | no | | Max examples to run in parallel (default: 3) |
|
|
254
|
-
| `--base-url <url>` | no | | Server base URL |
|
|
255
|
-
| `--json` | no | | Output the raw server response as JSON |
|
|
256
|
-
|
|
257
234
|
### `eigenpal workflow validate [options] [path]`
|
|
258
235
|
|
|
259
236
|
Local-only validation. Without [path]: runs the templated three-way check (./workflow.yaml + ./evaluators.yaml + ./dataset/). With [path] pointing at a YAML file: validates just that workflow.yaml. For per-noun targeting use `evaluators validate` or `dataset validate`.
|
|
@@ -363,10 +340,11 @@ Download the workflow's dataset as a ZIP archive.
|
|
|
363
340
|
|
|
364
341
|
### Options
|
|
365
342
|
|
|
366
|
-
| Flag
|
|
367
|
-
|
|
|
368
|
-
| `--out <zip>`
|
|
369
|
-
| `--
|
|
343
|
+
| Flag | Required | Default | Description |
|
|
344
|
+
| ------------------- | -------- | ------- | ----------------------------------------------------------------------------------- |
|
|
345
|
+
| `--out <zip>` | no | | Write the dataset ZIP to this path. When omitted, the binary streams to stdout. |
|
|
346
|
+
| `--example-id <id>` | no | `[]` | Export only this example (repeatable). When omitted, the whole dataset is exported. |
|
|
347
|
+
| `--base-url <url>` | no | | Server base URL |
|
|
370
348
|
|
|
371
349
|
### `eigenpal workflow dataset push [options] <workflow-id>`
|
|
372
350
|
|
|
@@ -725,7 +703,7 @@ Fetch the JSON Schema for one evaluator type. Pipe through `jq` to inspect speci
|
|
|
725
703
|
|
|
726
704
|
### `eigenpal workflow step exec [options] <type>`
|
|
727
705
|
|
|
728
|
-
DISABLED — local mimic runners removed pending server-side redesign (EIG-104). Use `
|
|
706
|
+
DISABLED — local mimic runners removed pending server-side redesign (EIG-104). Use `run` or `workflow experiment run` instead.
|
|
729
707
|
|
|
730
708
|
### Arguments
|
|
731
709
|
|
|
@@ -11,12 +11,12 @@ eigenpal runs compare <a> <b> # side-by-side diff of two runs
|
|
|
11
11
|
## 1. Reproduce the failure
|
|
12
12
|
|
|
13
13
|
```bash
|
|
14
|
-
eigenpal
|
|
14
|
+
eigenpal run workflows.<workflow-id> --example <example-name>
|
|
15
15
|
```
|
|
16
16
|
|
|
17
17
|
`<workflow-id>` is either a `wf_…` id or a slug (the YAML's `name:`
|
|
18
18
|
field). The workflow must be pushed (`eigenpal workflow push`) before
|
|
19
|
-
`
|
|
19
|
+
`eigenpal run` works — only saved workflows execute. Streams a live
|
|
20
20
|
step list as the run progresses. Bails to the prompt the moment a step
|
|
21
21
|
transitions to `failed`.
|
|
22
22
|
|
|
@@ -20,7 +20,7 @@ Until EIG-104 lands, single-step iteration goes through the queue:
|
|
|
20
20
|
|
|
21
21
|
```bash
|
|
22
22
|
# Run the whole workflow once with a specific example.
|
|
23
|
-
eigenpal
|
|
23
|
+
eigenpal run workflows.<workflow-id>
|
|
24
24
|
|
|
25
25
|
# Or iterate on one example via the experiment surface.
|
|
26
26
|
eigenpal workflow experiment run <workflow-id> --example-id <name>
|
|
@@ -47,7 +47,7 @@ names (`input` / `output` / `config`).
|
|
|
47
47
|
|
|
48
48
|
| Code | When |
|
|
49
49
|
| --- | --- |
|
|
50
|
-
| 2 | Always — the command is disabled and redirects to `
|
|
50
|
+
| 2 | Always — the command is disabled and redirects to `run` / `workflow experiment run`. |
|
|
51
51
|
|
|
52
52
|
## When this comes back
|
|
53
53
|
|
|
@@ -25,7 +25,7 @@ implement at runtime. Server-side validation errors point you here.
|
|
|
25
25
|
- `transform.*` — deterministic data transforms (set, remove, combine, split, merge,
|
|
26
26
|
script, template, pdf-embed, xlsx-to-json). WASM sandboxed where applicable.
|
|
27
27
|
- `action.*` — external side effects (HTTP, invoke another workflow, website reader).
|
|
28
|
-
- `control.*` — flow control (if, foreach, parallel, parallel_map, wait,
|
|
28
|
+
- `control.*` — flow control (if, foreach, parallel, parallel_map, wait, block, fail).
|
|
29
29
|
|
|
30
30
|
The full per-type catalog with field tables is auto-generated below from
|
|
31
31
|
`STEP_SCHEMAS`. The high-level map above tells you which family you want;
|
|
@@ -47,7 +47,6 @@ the catalog tells you what fields it takes.
|
|
|
47
47
|
| Map over an input array | `forEach:` on the step or `control.foreach` |
|
|
48
48
|
| Concurrent map over an array | `control.parallel_map` |
|
|
49
49
|
| Independent parallel branches | `control.parallel` |
|
|
50
|
-
| Pause for human approval | `control.approval` |
|
|
51
50
|
| External HTTP call | `action.http` |
|
|
52
51
|
| Call another workflow | `action.invoke-workflow` |
|
|
53
52
|
| Fetch a webpage as markdown | `action.website-reader` |
|
|
@@ -681,6 +680,9 @@ Execute another workflow and return its output
|
|
|
681
680
|
| --- | --- | --- | --- | --- |
|
|
682
681
|
| `workflowId` | string | yes | | ID of workflow to invoke |
|
|
683
682
|
| `input` | record<string, unknown> | no | | Input to pass to workflow |
|
|
683
|
+
| `wait` | boolean | no | | If true, wait for the invoked workflow to complete (default: false) |
|
|
684
|
+
| `timeout` | number | no | | Max wait time in ms when wait=true (default: 300000) |
|
|
685
|
+
| `pollInterval` | number | no | | How often to poll status in ms when wait=true (default: 1000) |
|
|
684
686
|
|
|
685
687
|
**Output:** `record<string, unknown>`
|
|
686
688
|
|
|
@@ -799,27 +801,6 @@ Pause workflow execution for a specified duration
|
|
|
799
801
|
| --- | --- | --- | --- | --- |
|
|
800
802
|
| `waited` | number | yes | | Actual milliseconds waited |
|
|
801
803
|
|
|
802
|
-
#### `control.approval` — Approval
|
|
803
|
-
|
|
804
|
-
Pause workflow for human approval before continuing
|
|
805
|
-
|
|
806
|
-
**Config** (at step level):
|
|
807
|
-
|
|
808
|
-
| Field | Type | Required | Default | Description |
|
|
809
|
-
| --- | --- | --- | --- | --- |
|
|
810
|
-
| `message` | string | no | | Message to display for approval |
|
|
811
|
-
| `timeoutMinutes` | number | no | `1440` | Auto-reject after timeout |
|
|
812
|
-
| `notifyEmail` | string | no | | Email to notify |
|
|
813
|
-
|
|
814
|
-
**Output:** `object`
|
|
815
|
-
|
|
816
|
-
| Field | Type | Required | Default | Description |
|
|
817
|
-
| --- | --- | --- | --- | --- |
|
|
818
|
-
| `approved` | boolean | yes | | Whether the request was approved |
|
|
819
|
-
| `approvedBy` | string | no | | Who approved |
|
|
820
|
-
| `approvedAt` | string | no | | When approved |
|
|
821
|
-
| `comment` | string | no | | Approver comment |
|
|
822
|
-
|
|
823
804
|
#### `control.block` — Block
|
|
824
805
|
|
|
825
806
|
Execute a reusable block workflow inline with input/output mapping
|
|
@@ -38,11 +38,11 @@ folder.
|
|
|
38
38
|
|
|
39
39
|
Required. Each entry is an object that adds an invocation surface:
|
|
40
40
|
|
|
41
|
-
- `{ type: api }` — the workflow can be called via REST or the CLI's `
|
|
41
|
+
- `{ type: api }` — the workflow can be called via REST or the CLI's `run`
|
|
42
42
|
- `{ type: manual }` — the workflow can be started manually
|
|
43
43
|
- `{ type: email, whitelist?: ... }` — inbound email
|
|
44
44
|
|
|
45
|
-
If `{ type: api }` is missing, `eigenpal
|
|
45
|
+
If `{ type: api }` is missing, `eigenpal run workflows.<slug>` rejects with a
|
|
46
46
|
403 / "Workflow does not have API trigger enabled."
|
|
47
47
|
|
|
48
48
|
### `inputs`
|
|
@@ -8,7 +8,7 @@ then run:
|
|
|
8
8
|
eigenpal workflow validate
|
|
9
9
|
|
|
10
10
|
# Run an example end-to-end against the configured base URL
|
|
11
|
-
eigenpal
|
|
11
|
+
eigenpal run workflows.__NAME__ --example sample
|
|
12
12
|
```
|
|
13
13
|
|
|
14
14
|
Add more examples under `dataset/examples/<name>/`. Each example needs an
|
|
@@ -9,7 +9,7 @@ fixed JSON schema.
|
|
|
9
9
|
|
|
10
10
|
1. Drop a sample PDF at `dataset/examples/sample-invoice/input/document/<filename>`.
|
|
11
11
|
2. Adjust the `extract.with.schema` to the fields your documents actually carry.
|
|
12
|
-
3. Run an example: `eigenpal
|
|
12
|
+
3. Run an example: `eigenpal run workflows.__NAME__ --example sample-invoice`.
|
|
13
13
|
4. When the output looks right, push: `eigenpal workflow push --file workflow.yaml`.
|
|
14
14
|
|
|
15
15
|
## Extending
|
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
# __NAME__
|
|
2
2
|
|
|
3
3
|
A text classification workflow scaffolded by `eigenpal init workflow`. Two examples
|
|
4
|
-
are pre-seeded so you can immediately run `eigenpal
|
|
4
|
+
are pre-seeded so you can immediately run `eigenpal run workflows.__NAME__ --example sample` and see
|
|
5
5
|
end-to-end behavior.
|
|
6
6
|
|
|
7
7
|
The eval config asserts on the `label` path only — confidence/rationale are
|