CLI Reference

scitex-ml exposes the stateless analysis surface from the terminal. Run scitex-ml --help for the live command list, or scitex-ml --help-recursive to dump help for every subcommand at once.

scitex-ml --version            # scitex-ml/X.Y.Z
scitex-ml --help-recursive     # all commands + subcommands

Commands

Category	Command	Description
Analysis	`scitex-ml compute-metrics`	Balanced accuracy, MCC, confusion matrix, ROC/PR-AUC from a predictions table
Analysis	`scitex-ml generate-report`	Full ClassificationReporter report (summary + plots) into a directory
Analysis	`scitex-ml reduce-dimensions`	PCA / UMAP 2-D projection figure from a feature matrix
Introspection	`scitex-ml list-python-apis`	List the public Python API (`-v` signatures, `-vv` docstrings, `--json`)
Integration	`scitex-ml mcp`	MCP server management (`start`, `doctor`, `list-tools`, `install`)
Integration	`scitex-ml skills`	List / get / install the bundled agent-facing skill pages
Utility	`scitex-ml install-shell-completion`	Wire up `<TAB>` completion in the user’s shell rc

Analysis verbs

Each verb is file-in → JSON/artifact-out and mirrors an MCP tool exactly (same name, same JSON — see MCP Server).

# Metrics from a saved predictions table (columns: y_true, y_pred[, y_proba])
scitex-ml compute-metrics preds.csv
scitex-ml compute-metrics preds.csv --label neg --label pos --json

# Full report (summary.json + plots) under ./report
scitex-ml generate-report preds.csv -o ./report

# PCA / UMAP figure coloured by a label column
scitex-ml reduce-dimensions features.csv -o pca.png --label-col target
scitex-ml reduce-dimensions features.csv -o umap.png --method umap --supervised

Introspection

scitex-ml list-python-apis          # names
scitex-ml list-python-apis -vv      # + signatures + docstrings
scitex-ml list-python-apis --json

Shell completion

scitex-ml install-shell-completion --shell bash
eval "$(scitex-ml print-shell-completion --shell bash)"