CLI reference¶
Every command below is generated directly from the argparse definition in
science_adk/cli.py, so this page cannot describe a flag that
does not exist.
Finding the project
Commands operate on the research project found by searching upward from
the current directory. Pass --path DIR to point at one explicitly.
Global options¶
| Option | Description |
|---|---|
--version |
Print the installed version and exit. |
--path DIR |
Project directory (default: search upward from the cwd). |
-h, --help |
Show help for any command or subcommand. |
Commands at a glance¶
science-adk init¶
| Argument | Description | Required |
|---|---|---|
directory (optional) |
||
-q, --question |
The research question | |
--example NAME |
Start from a bundled example: kepler | |
--force |
Overwrite existing files |
Example
science-adk goal¶
| Argument | Description | Required |
|---|---|---|
--set-question |
Set the research question | |
--metric |
Name of the target metric | |
--target TARGET |
Value the metric must reach | |
--json |
Example
science-adk goal
science-adk goal \
--set-question "Does a GNN embedding predict aqueous solubility better than Morgan fingerprints?" \
--metric rmse \
--target 0.8
science-adk experiment¶
science-adk experiment new¶
| Argument | Description | Required |
|---|---|---|
hypothesis |
The testable claim | yes |
--rationale |
Why this is worth testing | |
--parent |
Experiment whose findings motivated this one |
Example
science-adk experiment new \
"GNN embeddings outperform Morgan fingerprints on ESOL" \
--rationale "Learned 3D representations should capture solvation geometry that topological fingerprints discard."
science-adk experiment list¶
| Argument | Description | Required |
|---|---|---|
--json |
Example
science-adk validate¶
| Argument | Description | Required |
|---|---|---|
experiment (optional) |
||
--json |
Example
science-adk run¶
| Argument | Description | Required |
|---|---|---|
experiment (optional) |
||
--only |
Run only these nodes and their dependents | |
--reuse |
Reuse unchanged outputs from a previous run (default: latest) | |
-q, --quiet |
Example
science-adk run
science-adk run --reuse # re-run only what changed
science-adk run --only analyze evaluate
science-adk score¶
| Argument | Description | Required |
|---|---|---|
experiment (optional) |
||
--run |
Run id (default: latest) | |
--json |
Example
science-adk audit¶
science-adk audit [-h] [--run RUN] --scientific-value 0..1
--method-fidelity 0..1 --implementation-quality 0..1
--rationale RATIONALE
[experiment]
| Argument | Description | Required |
|---|---|---|
experiment (optional) |
||
--run |
Run id (default: latest) | |
--scientific-value 0..1 |
yes | |
--method-fidelity 0..1 |
yes | |
--implementation-quality 0..1 |
yes | |
--rationale |
Why these ratings | yes |
Example
science-adk audit \
--scientific-value 0.85 \
--method-fidelity 0.90 \
--implementation-quality 0.95 \
--rationale "GNN embedding reaches RMSE 0.72 on held-out ESOL against 0.95 for Morgan fingerprints."
science-adk report¶
| Argument | Description | Required |
|---|---|---|
experiment (optional) |
||
--run |
Run id (default: latest) |
Example
science-adk campaign¶
Example
science-adk learn¶
| Argument | Description | Required |
|---|---|---|
insight |
yes | |
--evidence |
What supports this | |
--experiment |
||
--run |
Example
science-adk tools¶
| Argument | Description | Required |
|---|---|---|
--json |
Example
science-adk graph¶
| Argument | Description | Required |
|---|---|---|
experiment (optional) |
Example
science-adk status¶
Example