@barivia/barsom-mcp 0.23.4 → 0.23.8

package/dist/tools/account.js

@@ -38,7 +38,7 @@ export function registerAccountTool(server) {
 | Action | Use when |
 |--------|----------|
 | status | Before large jobs — see plan tier, GPU availability, queue depth, training time estimates, credit balance. When the API is down, returns calm \`{ok:false, retry_after_sec}\` instead of HTML. |
-| history | Viewing recent compute usage and credit spend |
+| history | Viewing recent compute usage and credit spend (gateway-only; local/direct API soft-degrades to "unavailable" — use inventory/jobs list) |
 | inventory | Rebuild a durable markdown overview of datasets + recent jobs (agent-friendly org catalog) |
 | add_funds | Getting instructions to add credits |
 | health | Diagnose API reachability during outages — GET /health + /ready; does **not** probe R2/object storage |
@@ -157,11 +157,32 @@ NOT FOR: Training itself — use train(action=map). This tool only manages the a
             return { content: [{ type: "text", text: lines.join("\n") }] };
         }
         if (action === "history") {
-            const data = await apiCall("GET", `/v1/compute/history?limit=${limit || 10}`);
-            const history = data.history.map((h) => `- ${h.started_at} | ${h.tier} | ${h.duration_minutes} min | $${(h.credits_charged / 100).toFixed(2)}`).join("\n");
-            return {
-                content: [{ type: "text", text: `Credit Balance: $${(data.credit_balance_cents / 100).toFixed(2)}\n\nRecent Usage:\n${history}` }]
-            };
+            // Gateway serves /v1/compute/history (D1 leases). Local direct API has no handler → 404.
+            try {
+                const data = await apiCall("GET", `/v1/compute/history?limit=${limit || 10}`);
+                const rows = Array.isArray(data?.history) ? data.history : [];
+                const history = rows.length === 0
+                    ? "(none)"
+                    : rows.map((h) => `- ${h.started_at} | ${h.tier} | ${h.duration_minutes} min | $${(h.credits_charged / 100).toFixed(2)}`).join("\n");
+                const bal = data?.credit_balance_cents != null
+                    ? `Credit Balance: $${(Number(data.credit_balance_cents) / 100).toFixed(2)}\n\n`
+                    : "";
+                return {
+                    content: [{ type: "text", text: `${bal}Recent Usage:\n${history}` }]
+                };
+            }
+            catch (e) {
+                const status = e?.httpStatus;
+                if (status === 404 || status === 410) {
+                    return {
+                        content: [{
+                                type: "text",
+                                text: "Usage history unavailable on this API endpoint (compute-lease history is gateway-only). Use account(action=inventory) or jobs(action=list) for recent job activity.",
+                            }]
+                    };
+                }
+                throw e;
+            }
         }
         if (action === "inventory") {
             const jobLimit = limit && limit > 0 ? Math.min(Math.floor(limit), 100) : 50;

package/dist/tools/datasets.js

@@ -4,7 +4,7 @@ import { gzipSync } from "node:zlib";
 import { createHash } from "node:crypto";
 import { z } from "zod";
 import { registerAuditedTool } from "../audit.js";
-import { apiCall, getWorkspaceRootAsync, resolveFilePathForUpload, textResult, structuredTextResult, pollUntilComplete, POLL_DERIVE_MAX_MS, POLL_ANALYZE_MAX_MS, UPLOAD_DATASET_TIMEOUT_MS, LARGE_UPLOAD_BYTES, PRESIGNED_PUT_TIMEOUT_MS, POLL_STAGE_MAX_MS, streamFileSha256, putPresignedStream, resolveUploadContentType, suggestDatasetPreview, } from "../shared.js";
+import { apiCall, getWorkspaceRootAsync, resolveFilePathForUpload, textResult, structuredTextResult, pollUntilComplete, POLL_DERIVE_MAX_MS, POLL_ANALYZE_MAX_MS, UPLOAD_DATASET_TIMEOUT_MS, LARGE_UPLOAD_BYTES, PRESIGNED_PUT_TIMEOUT_MS, POLL_STAGE_MAX_MS, awaitDatasetStageIfNeeded, streamFileSha256, putPresignedStream, resolveUploadContentType, suggestDatasetPreview, } from "../shared.js";
 import { GZIP_UPLOAD_HINT } from "../job_status_format.js";
 import { formatDatasetInventoryLine, formatTags } from "../inventory_format.js";
 import { PERIODICITY_ROW_ORDER_NOTE, SMALL_N_SOFT, buildCpuScaleGuidanceLines, buildSmallNGuardrailLines, } from "../training_scale_guidance.js";
@@ -165,7 +165,7 @@ Step 1 (after upload): always datasets(action=preview) to verify column types an
 action=preview: Show columns, stats, sample rows, cyclic/datetime detections. ALWAYS preview before train(action=map) on an unfamiliar dataset.
 action=analyze: Pre-training analysis on numeric columns — Pearson correlation, autocorrelation periodicity scores, and column recommendations (train, consider_dropping, project_later, low_variance). Use to choose which columns to include in training and which to project onto the map after training.
 action=list: List all datasets belonging to the organisation with id, name, rows, cols, created_at, description, tags (use tags/description or id to distinguish datasets with the same name).
-action=get: Fetch one dataset by dataset_id — returns status, staged_prefix, staged_version, ingest_error, description/tags, and stage_job_id when staging is pending.
+action=get: Fetch one dataset by dataset_id — returns status, train_ready (true only when status=ready AND staged_prefix is set), staged_prefix, staged_version, ingest_error, description/tags, and stage_job_id when staging is pending. Train only when train_ready=true (do not treat bare status=ready as enough).
 action=update: PATCH name/description/tags on an existing dataset.
 action=upload: Optional description and tags help later inventory (also settable via update).
 DO NOT AUTO-SUBSET: Call action=subset only when the user explicitly requests a filter, slice, or random sample. Default path is upload → preview/analyze → train on the full dataset_id (GPU handles scale). Never downsample to "save time" or "be helpful".
@@ -230,7 +230,7 @@ ESCALATION: If upload fails with column errors, open the file locally and verify
         })
             .optional()
             .describe("Deprecated — use filters instead. Single filter condition."),
-        expression: z.string().optional().describe("action=add_expression: math expression referencing column names (e.g. revenue/cost, log(price), rolling_mean(vol, 20))"),
+        expression: z.string().optional().describe("action=add_expression: math expression referencing column names (e.g. revenue/cost, log(price), rolling_mean(vol, 20)). Awkward names (spaces/parens): quote with backticks, e.g. `P1 (bar)` / 2"),
         options: z
             .object({
             missing: z.enum(["skip", "zero", "interpolate"]).optional(),
@@ -377,7 +377,7 @@ ESCALATION: If upload fails with column errors, open the file locally and verify
                         gzHeaders["X-Dataset-Description"] = description;
                     if (tags !== undefined)
                         gzHeaders["X-Dataset-Tags"] = JSON.stringify(tags);
-                    const data = (await apiCall("POST", "/v1/datasets", gzBytes, gzHeaders, UPLOAD_DATASET_TIMEOUT_MS));
+                    const data = await awaitDatasetStageIfNeeded((await apiCall("POST", "/v1/datasets", gzBytes, gzHeaders, UPLOAD_DATASET_TIMEOUT_MS)));
                     const gid = data.id ?? data.dataset_id;
                     if (gid != null)
                         data.suggested_next_step = `${suggestDatasetPreview(String(gid))} to inspect columns before training.`;
@@ -410,7 +410,7 @@ ESCALATION: If upload fails with column errors, open the file locally and verify
                 uploadBody = gzipSync(Buffer.from(body, "utf-8"));
                 uploadHeaders["Content-Encoding"] = "gzip";
             }
-            const data = (await apiCall("POST", "/v1/datasets", uploadBody, uploadHeaders, UPLOAD_DATASET_TIMEOUT_MS));
+            const data = await awaitDatasetStageIfNeeded((await apiCall("POST", "/v1/datasets", uploadBody, uploadHeaders, UPLOAD_DATASET_TIMEOUT_MS)));
             const id = data.id ?? data.dataset_id;
             if (id != null)
                 data.suggested_next_step = `${suggestDatasetPreview(String(id))} to inspect columns before training.`;
@@ -732,16 +732,20 @@ ESCALATION: If upload fails with column errors, open the file locally and verify
                 throw new Error("datasets(get) requires dataset_id");
             const ds = (await apiCall("GET", `/v1/datasets/${dataset_id}`));
             const tagStr = formatTags(ds.tags);
+            const trainReady = ds.train_ready === true;
             const lines = [
                 `Dataset: ${ds.name ?? "?"} (${ds.id ?? dataset_id})`,
-                `Status: ${ds.status ?? "ready"}`,
+                `Status: ${ds.status ?? "unknown"}`,
+                `Train ready: ${trainReady ? "yes" : "no"}`,
                 `Rows × cols: ${ds.rows ?? "?"} × ${ds.cols ?? "?"}`,
                 ds.size_bytes != null ? `Size: ${Number(ds.size_bytes).toLocaleString()} bytes` : "",
                 ds.description != null && String(ds.description).trim() !== "" ? `Description: ${String(ds.description)}` : "",
                 tagStr ? `Tags: ${tagStr}` : "",
                 ds.staged_prefix != null ? `Staged prefix: ${String(ds.staged_prefix)}` : "",
                 ds.staged_version != null ? `Staged version: ${String(ds.staged_version)}` : "",
-                ds.stage_job_id != null ? `Stage job: ${String(ds.stage_job_id)} (poll jobs(action=status))` : "",
+                !trainReady && ds.stage_job_id != null
+                    ? `Stage job: ${String(ds.stage_job_id)} (poll jobs(action=status) until train_ready)`
+                    : "",
                 cleanNullable(ds.ingest_error) ? `Ingest error: ${cleanNullable(ds.ingest_error)}` : "",
                 ds.created_at != null ? `Created: ${String(ds.created_at)}` : "",
             ].filter(Boolean);

package/dist/tools/explore_map.js

@@ -184,7 +184,7 @@ async function handleResultsExplorer(job_id) {
 export function registerExploreMapTool(server) {
     const toolConfig = {
         title: "Results Explorer",
-        description: "PREFERRED way to browse a completed map's figures after a first results(get) glance. Interactive explorer for a completed map job: a responsive view of key metrics and every available figure (combined, U-matrix, hit histogram, component planes). Embeds as an MCP App or falls back to a standalone localhost page. Offer it to the user once a train job completes; results(get) remains the headless/metrics path.",
+        description: "PREFERRED way to browse a completed map's figures after a first results(get) glance. Interactive explorer with a grouped figure dropdown above the plot (Summary → Diagnostics → Components → Other): combined, U-matrix, hit histogram, component planes, plus sidebar metrics/highlights. Embeds as an MCP App or falls back to a standalone localhost page. Offer it to the user once a train job completes; results(get) remains the headless/metrics path.",
         inputSchema: {
             job_id: z.string().describe("Job ID of a completed map training job"),
         },

package/dist/tools/inference.js

@@ -79,7 +79,7 @@ action=predict: Score rows against the trained map.
     - "compact"   → predictions.csv (row_id, bmu_x, bmu_y, bmu_node_index, cluster_id [, quantization_error, potential_anomaly]).
     - "annotated" → annotated.csv (full source CSV with bmu_x, bmu_y, bmu_node_index, cluster_id appended). Requires a dataset (no inline rows).
   Regime (auto): if the resolved dataset is the parent training dataset, regime="training" and QE / qe_p95 / potential_anomaly are omitted (QE on training data is fitting error, not generalisation — use a held-out dataset for quality). Otherwise regime="new" and full QE columns are returned.
-  Schema rules: dataset / batch rows must match the trained feature set including cyclic-expanded columns (e.g. key_cos, key_sin, not raw key). For a single inline row, the proxy uses the stateless model endpoint — raw categorical strings are allowed for baseline categorical_features models, and raw cyclic inputs are expanded from training config.
+  Schema rules: prefer dataset_id for scoring. Inline rows may use raw cyclic keys (e.g. hour) — the API/worker expands them from training config to hour_cos/hour_sin; pre-expanded cos/sin columns also work. For a single inline row, the proxy uses the stateless model endpoint (same cyclic expansion). Raw categorical strings are allowed for baseline categorical_features models on that single-row path.
   Routing: prefer dataset_id for many rows or whenever the map uses irregular SIOM / GeneralTopology layouts — the async worker path is the supported batch scorer. Single-row rows take a fast stateless path that may return invalid_inference_input on some topologies; if so, retry with dataset_id (a one-row dataset is fine). FLooP-SIOM: use dataset_id predict first.
   When the scored set has at most ${PREDICT_PREVIEW_ROW_CAP} rows, completed responses include a short per-line preview in the tool text for chat agents.
 
@@ -348,32 +348,64 @@ action=report: Returns a report manifest for the given job_id (job must be compl
         const download_urls = (results.download_urls ?? {});
         const figure_manifest = (results.figure_manifest ?? []);
         const cluster_summary = results.cluster_summary;
+        const summary = (results.summary ?? {});
         const expires_in = results.expires_in ?? 900;
+        const downloadKeys = Object.keys(download_urls);
+        const clusterSnippet = Array.isArray(cluster_summary) && cluster_summary.length > 0
+            ? cluster_summary.slice(0, 3).map((c) => {
+                const id = c.cluster_id ?? c.id ?? "?";
+                const n = c.n_nodes ?? c.size ?? "?";
+                return { cluster_id: id, n_nodes: n };
+            })
+            : [];
         const lines = [
             `Report manifest — job: ${job_id}`,
             `Use these artifacts to build your own report (Quarto, Jupyter, script). URLs expire in ${expires_in}s.`,
             "",
             "Figures (figure_manifest → download_urls):",
-            ...figure_manifest.slice(0, 12).map((e) => `  ${e.logical_name} → ${e.filename}`),
+            ...(figure_manifest.length
+                ? figure_manifest.slice(0, 12).map((e) => `  ${e.logical_name} → ${e.filename}`)
+                : ["  (none listed — small grids may omit PNGs; use cluster_summary / summary metrics below)"]),
             figure_manifest.length > 12 ? `  ... and ${figure_manifest.length - 12} more` : "",
             "",
             "Key download URLs (use get_result_image or download_urls from results(action=get)):",
-            ...Object.keys(download_urls)
-                .filter((k) => !k.endsWith(".json") || k === "cluster_summary.json")
-                .slice(0, 8)
-                .map((k) => `  ${k}: ${download_urls[k]?.slice(0, 60)}...`),
+            ...(downloadKeys.length
+                ? downloadKeys
+                    .filter((k) => !k.endsWith(".json") || k === "cluster_summary.json")
+                    .slice(0, 8)
+                    .map((k) => `  ${k}: ${download_urls[k]?.slice(0, 60)}...`)
+                : ["  (no presigned URLs — artifacts may still be in summary / cluster_summary inline)"]),
             "",
             cluster_summary?.length
-                ? `Cluster summary: ${cluster_summary.length} clusters (in response or GET .../cluster_summary).`
+                ? `Cluster summary: ${cluster_summary.length} clusters (inline snippet below; full via GET .../cluster_summary or download_urls.cluster_summary.json).`
                 : "Cluster summary: not available for this job (train with current worker to get it).",
+            ...(clusterSnippet.length
+                ? ["Inline cluster_summary (first 3):", ...clusterSnippet.map((c) => `  cluster ${c.cluster_id}: n_nodes=${c.n_nodes}`)]
+                : []),
             "",
             "Metrics: in summary (quantization_error, topographic_error, explained_variance, silhouette, etc.) or GET .../quality-report.",
+            summary.quantization_error != null ? `  QE=${summary.quantization_error}` : "",
+            summary.topographic_error != null ? `  TE=${summary.topographic_error}` : "",
         ];
         const manifestText = lines.filter(Boolean).join("\n");
         return {
             content: [
                 { type: "text", text: manifestText },
-                { type: "text", text: "Structured manifest (for automation): " + JSON.stringify({ job_id, figure_manifest, has_cluster_summary: !!cluster_summary?.length, download_url_keys: Object.keys(download_urls) }) },
+                {
+                    type: "text",
+                    text: "Structured manifest (for automation): " +
+                        JSON.stringify({
+                            job_id,
+                            figure_manifest,
+                            has_cluster_summary: !!cluster_summary?.length,
+                            cluster_summary_snippet: clusterSnippet,
+                            download_url_keys: downloadKeys,
+                            summary_metrics: {
+                                quantization_error: summary.quantization_error ?? null,
+                                topographic_error: summary.topographic_error ?? null,
+                            },
+                        }),
+                },
             ],
         };
     });

package/dist/tools/results.js

@@ -2,8 +2,7 @@ import path from "node:path";
 import fs from "node:fs/promises";
 import { z } from "zod";
 import { registerAuditedTool } from "../audit.js";
-import { apiCall, apiRawCall, getWorkspaceRootAsync, sandboxPath, pollUntilComplete, tryAttachImage, resetInlineAttachBudget, attachResultsImages, resolveOutputDpi, POLL_RECOLOR_MAX_MS, fmtDensityDiffNode, getCaptionForImage, shouldFetchAllRemainingFigures, mimeForFilename, structuredTextResult, } from "../shared.js";
-import { formatSiomOccupationSummary } from "../siom_occupation_format.js";
+import { apiCall, apiRawCall, getWorkspaceRootAsync, sandboxPath, pollUntilComplete, tryAttachImage, resetInlineAttachBudget, attachResultsImages, resolveOutputDpi, POLL_RECOLOR_MAX_MS, fmtDensityDiffNode, getCaptionForImage, shouldFetchAllRemainingFigures, mimeForFilename, structuredTextResult, coerceFiguresArg, } from "../shared.js";
 export function registerResultsTool(server) {
     registerAuditedTool(server, "results", `Retrieve, recolor, download, or export figures and metrics for a completed map job. Presentation and artifact access only — operations on the frozen map (predict, compare, project, impute_column, transition_flow) live in **inference**.
 
@@ -13,6 +12,7 @@ export function registerResultsTool(server) {
 | export | Learning curve, raw weights, or per-node stats | instant |
 | download | Saving figures to a local folder | instant |
 | recolor | Changing colormap or output format without retraining | async (~10–30s) |
+| feature_divergences | Refresh KL matrices / optional Wasserstein on component planes | async |
 
 Temporal state-transition arrows moved to **inference(action=transition_flow)** (use the new result job_id with results(get/download) to read the figure).
 
@@ -20,35 +20,37 @@ ONLY call this after jobs(action=status) returns "completed".
 ESCALATION: If job not found, verify job_id. If "job not complete", poll with jobs(action=status).
 
 action=get: Returns text summary with quality metrics and inline images.
-  - figures: omit = combined only. "all" = all published plots (excludes unpublished cyclic cos/sin expansions unless train used viz_upload_cyclic_components=true). "none" = metrics text only, no images (recommended for sweeps and LLM clients to keep tool payloads small). Array = specific logical names (combined, umatrix, hit_histogram, learning_curve, correlation, component_1..N). Request unpublished cyclic planes via results(recolor, figures=[component_N, ...]).
+  - figures: omit = combined only. Prefer string "all" / "none" (not an array). "all" = all published plots (excludes unpublished cyclic cos/sin expansions unless train used viz_upload_cyclic_components=true). "none" = metrics text only, no images (recommended for sweeps and LLM clients to keep tool payloads small). Array = specific logical names (combined, umatrix, hit_histogram, learning_curve, correlation, divergence_kl, component_1..N) or download filenames. Hosts that stringify arrays are coerced. Request unpublished cyclic planes via results(recolor, figures=[component_N, ...]).
   - include_individual: if true (and figures omitted), inlines every component plane.
   - After recolor or project, use the job_id returned by that operation to get the new artifacts, not the original training job_id.
   - After showing results, guide the user: QE interpretation, whether to retrain, which features to explore.
   - METRIC INTERPRETATION: QE lower=better (<1.5 good) | TE<0.1 good | Explained variance>0.7 good | Silhouette higher=better | Davies-Bouldin lower=better.
 
-action=export: Structured data exports. Use export_type= to choose what to export.
-  - export_type=training_log: learning curve sparklines + plot. Diagnose convergence/plateau/divergence.
+action=export: Structured data exports (not the same as download). Use export_type= to choose what to export.
+  - export_type=training_log: learning curve sparklines + plot. Diagnose convergence/plateau/divergence. Prefer this over download for the training log.
   - export_type=weights: full weight matrix + normalization stats. For external analysis or custom viz.
   - export_type=nodes: per-node hit count + feature stats. Profile clusters and operating modes.
 
-action=download: Save figures to disk. Use so user can open, share, or version files locally.
+action=download: Save figures/artifacts to disk (deduped file list). Use so user can open, share, or version files locally.
   - folder: e.g. "." or "./results". Relative to the client's working directory (or MCP workspace). Each job lands in a dedicated subfolder under folder: {job_type}_{label} when label was set at train time, otherwise {job_type}_{job_id_prefix} — so parallel downloads never overwrite summary.json or figure names.
-  - figures: "all" (default) or array of filenames.
-  - include_json: when true, also saves summary.json (and adds it to the download list if figures omitted).
+  - figures: "all" (default) downloads published image files; OR a string array of exact artifact filenames from the job (e.g. ["predictions.csv"], ["combined.png","umatrix.png"]). Enum tokens like "none" are for action=get, not download.
+  - include_json: when true, also saves summary.json and other JSON artifacts (use this when you need CSVs listed in summary.files without naming each figure).
 
 action=recolor: Change colormap or output format — no retraining. Returns a new job_id; auto-polls 60s.
   AFTER: use results(action=get, job_id=NEW_JOB_ID).
   Colormaps: viridis, plasma, inferno, magma, cividis, turbo, coolwarm, balance, hsv, twilight, RdBu, Spectral. Also: thermal, hot, grays, RdYlBu, delta, curl, phase, rainbow, tol_bright, tol_muted.
 
+action=feature_divergences: Recompute component-plane divergences (min-shift sum-normalized denormalized planes). Finalize already writes symmetric KL + rankings; use this for Wasserstein or a refresh. metrics default ["symmetric_kl","directed_kl"]; add "wasserstein" for W1. Auto-polls then summarizes top pairs vs Pearson top_correlations.
+
 NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or temporal flow (use inference(action=transition_flow)).`, {
         action: z
-            .enum(["get", "recolor", "download", "export"])
-            .describe("get: inline results + metrics; recolor: new colormap/format (async); download: save to disk; export: structured data (training_log/weights/nodes). Temporal flow moved to inference(action=transition_flow)."),
+            .enum(["get", "recolor", "download", "export", "feature_divergences"])
+            .describe("get: inline results + metrics; recolor: new colormap/format (async); download: save to disk; export: structured data; feature_divergences: KL/W1 on component planes (async)."),
         job_id: z.string().describe("Job ID of a completed job"),
-        figures: z
+        figures: z.preprocess(coerceFiguresArg, z
             .union([z.enum(["default", "combined_only", "all", "images", "none"]), z.array(z.string())])
             .optional()
-            .describe("action=get: omit=combined only; 'all'=all plots; 'none'=metrics text only (no images, ideal for sweeps and LLM clients); array=specific (combined,umatrix,hit_histogram,learning_curve,correlation,component_1..N). action=download: 'all'=all image files; array=specific filenames."),
+            .describe("Prefer string 'all'/'none'. action=get: omit=combined only; 'all'=all plots; 'none'=metrics text only; array=logical names (combined,umatrix,…). action=download: 'all'=image files; array=exact filenames (e.g. combined.png). Stringified JSON arrays from hosts are accepted.")),
         include_individual: z
             .boolean()
             .optional()
@@ -80,12 +82,20 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
         recolor_figures: z
             .array(z.string())
             .optional()
-            .describe("action=recolor: which figures to re-render (default: [combined]). Options: combined, umatrix, hit_histogram, correlation, component_1..N"),
+            .describe("action=recolor: which figures to re-render (default: [combined]). Options: combined, umatrix, hit_histogram, correlation, divergence_kl, component_1..N"),
         export_type: z
             .enum(["training_log", "weights", "nodes", "clusters", "transition_matrix"])
             .optional()
             .describe("action=export: training_log=learning curve; weights=weight matrix; nodes=per-node stats; clusters=regime boundaries; transition_matrix=Markov transition matrix"),
-    }, async ({ action, job_id, figures, include_individual, include_json, folder, colormap, output_format, output_dpi, recolor_figures, export_type: exportType }) => {
+        metrics: z
+            .array(z.enum(["symmetric_kl", "directed_kl", "wasserstein"]))
+            .optional()
+            .describe("action=feature_divergences: which matrices to compute (default symmetric_kl+directed_kl). Add wasserstein for W1 (slower)."),
+        wasserstein_method: z
+            .enum(["sinkhorn", "exact"])
+            .optional()
+            .describe("action=feature_divergences: OT solver when metrics includes wasserstein (default sinkhorn)."),
+    }, async ({ action, job_id, figures, include_individual, include_json, folder, colormap, output_format, output_dpi, recolor_figures, export_type: exportType, metrics: divergenceMetrics, wasserstein_method: wassersteinMethod }) => {
         const figuresArg = figures;
         if (action === "get") {
             resetInlineAttachBudget();
@@ -226,10 +236,7 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
                     hitStats.active_node_fraction !== undefined
                         ? `Grid BMU occupancy: ${(Number(hitStats.active_node_fraction) * 100).toFixed(1)}% active nodes on the map grid`
                         : "",
-                    ...formatSiomOccupationSummary(siom, {
-                        files: summary.files ?? [],
-                        compact: true,
-                    }),
+                    ...(siom ? [`SIOM layer coverage (distinct from grid dead nodes) — utilization: ${fmt(siom.utilization)} | dead_fraction: ${fmt(siom.dead_fraction)} | siom_qe: ${fmt(siom.siom_qe)}`] : []),
                     hitStats.grid_suggestion ? `Grid hint: ${String(hitStats.grid_suggestion)}` : "",
                     ...(policy.auto_rules_applied && Array.isArray(policy.auto_rules_applied) && policy.auto_rules_applied.length > 0
                         ? [`Auto policy: ${policy.auto_rules_applied.join(", ")} (${policy.viz_mode ?? "viz"}, metrics=${policy.quality_metrics ?? "?"})`]
@@ -300,21 +307,37 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
                     `Coverage Assessment:`,
                     `  Status:        ${coverageStatus}${isDegenerate ? " *** NON-COMPETITIVE ***" : isWarning ? " * CAUTION *" : ""}`,
                     `  Diagnosis:     ${String(coverage.diagnosis ?? "No plain-language diagnosis available.")}`,
-                    ...formatSiomOccupationSummary(siom, { files }),
+                    `  Utilization:   ${siom.utilization !== undefined ? Number(siom.utilization).toFixed(4) : "N/A"}`,
+                    `  SIOM dead fraction (layer coverage, not grid BMU hits): ${siom.dead_fraction !== undefined ? Number(siom.dead_fraction).toFixed(4) : "N/A"}`,
                     `  Occupied nodes:${hitStats.occupied_nodes !== undefined ? ` ${hitStats.occupied_nodes}` : " N/A"}`,
                     `  Top hit share: ${hitStats.top_hit_share !== undefined ? Number(hitStats.top_hit_share).toFixed(4) : "N/A"}`,
+                    `  Gini:          ${siom.gini !== undefined ? Number(siom.gini).toFixed(4) : "N/A"}`,
+                    `  Entropy:       ${siom.entropy !== undefined ? Number(siom.entropy).toFixed(4) : "N/A"}`,
+                    `  SIOM gamma:    ${siom.gamma !== undefined ? Number(siom.gamma).toFixed(3) : "N/A"} | decay: ${siom.decay !== undefined ? Number(siom.decay).toFixed(4) : "N/A"}`,
                     ...(summary.elastic ? (() => {
                         const el = summary.elastic;
                         return [`  Elastic:       λ=${el.lambda !== undefined ? Number(el.lambda).toFixed(3) : "0"} μ=${el.mu !== undefined ? Number(el.mu).toFixed(3) : "0"} anchor=${el.anchor !== undefined ? Number(el.anchor).toFixed(3) : "0"}`];
                     })() : []),
                     ``,
-                    `Layout tuning: if Voronoi cells look crowded in weight space (high top_hit_share), reduce max_nodes and/or raise gamma/siom_decay before growth_interval. See training_guidance keys floop_voronoi_crowding, floop_siom_params, floop_elastic_params, siom_gamma.`,
+                    `Layout tuning: if Voronoi cells look crowded in weight space (high top_hit_share), reduce max_nodes and/or raise gamma/siom_decay before growth_interval. See training_guidance keys floop_voronoi_crowding, floop_siom_params, floop_elastic_params.`,
                     ``,
                     `Quality Metrics:`,
                     `  Quantization Error: ${summary.quantization_error !== undefined ? Number(summary.quantization_error).toFixed(4) : "N/A"}${qeNote}`,
                     `  Topographic Error:  ${summary.topographic_error !== undefined ? Number(summary.topographic_error).toFixed(4) : "N/A"}${teNote}`,
                     "",
                     "Choosing a map for presentation: Lower QE/TE often favors larger, denser graphs. For slides and interpretation, sparser graphs (fewer occupied nodes, lower top_hit_share) are often easier to read even when metrics rank them lower. Compare figures side-by-side before picking a winner from metrics alone.",
+                    ...(() => {
+                        const topDivs = summary.top_divergences;
+                        const klComp = summary.feature_kl_complexity;
+                        const out = [];
+                        if (topDivs?.length) {
+                            out.push(`Top component-plane divergences (symmetric KL): ${topDivs.map((t) => `${t.f1}&${t.f2}=${typeof t.symmetric_kl === "number" ? Number(t.symmetric_kl).toFixed(3) : t.symmetric_kl}`).join(", ")}`);
+                        }
+                        if (klComp?.length) {
+                            out.push(`Feature KL complexity: ${klComp.map((t) => `${t.feature}=${typeof t.mean_outgoing_kl === "number" ? Number(t.mean_outgoing_kl).toFixed(3) : t.mean_outgoing_kl}`).join(", ")}`);
+                        }
+                        return out.length ? ["", ...out] : [];
+                    })(),
                     ...(Array.isArray(coverage.recommended_actions) && coverage.recommended_actions.length > 0
                         ? ["", "Recommended next steps:", ...coverage.recommended_actions.map((step) => `  - ${String(step)}`)]
                         : []),
@@ -357,7 +380,15 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
                     : "";
                 const topCorrs = summary.top_correlations;
                 const topCorrLine = topCorrs && topCorrs.length > 0
-                    ? `Strongest feature correlations: ${topCorrs.map((t) => `${t.f1 ?? "?"} & ${t.f2 ?? "?"} (${typeof t.r === "number" ? Number(t.r).toFixed(2) : t.r})`).join(", ")}`
+                    ? `Strongest feature correlations (Pearson, linear): ${topCorrs.map((t) => `${t.f1 ?? "?"} & ${t.f2 ?? "?"} (${typeof t.r === "number" ? Number(t.r).toFixed(2) : t.r})`).join(", ")}`
+                    : "";
+                const topDivs = summary.top_divergences;
+                const topDivLine = topDivs && topDivs.length > 0
+                    ? `Strongest component-plane divergences (symmetric KL, nonlinear): ${topDivs.map((t) => `${t.f1 ?? "?"} & ${t.f2 ?? "?"} (${typeof t.symmetric_kl === "number" ? Number(t.symmetric_kl).toFixed(3) : t.symmetric_kl})`).join(", ")}`
+                    : "";
+                const klComplexity = summary.feature_kl_complexity;
+                const klComplexityLine = klComplexity && klComplexity.length > 0
+                    ? `Feature KL complexity (mean outgoing directed KL): ${klComplexity.map((t) => `${t.feature ?? "?"} (${typeof t.mean_outgoing_kl === "number" ? Number(t.mean_outgoing_kl).toFixed(3) : t.mean_outgoing_kl})`).join(", ")}`
                     : "";
                 const siom = summary.siom;
                 const textSummary = [
@@ -386,11 +417,14 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
                     ...(compLines ? [`Component value ranges:\n${compLines}`] : []),
                     ...(clusterLines ? ["", "Cluster map (defining features):", clusterLines] : []),
                     ...(topCorrLine ? ["", topCorrLine] : []),
-                    ...(siom
-                        ? ["", ...formatSiomOccupationSummary(siom, {
-                                files: summary.files ?? [],
-                            })]
-                        : []),
+                    ...(topDivLine ? [topDivLine] : []),
+                    ...(klComplexityLine ? [klComplexityLine] : []),
+                    ...(topDivLine || klComplexityLine ? ["(Compare Pearson vs KL: high |r| ≠ low KL; use results(action=feature_divergences, metrics=[\"wasserstein\"]) for W1.)"] : []),
+                    ...(siom ? ["", "SIOM layer coverage (distinct from grid dead nodes):",
+                        `  Utilization: ${siom.utilization !== undefined ? Number(siom.utilization).toFixed(4) : "N/A"}`,
+                        `  SIOM dead fraction: ${siom.dead_fraction !== undefined ? Number(siom.dead_fraction).toFixed(4) : "N/A"}`,
+                        `  Gini: ${siom.gini !== undefined ? Number(siom.gini).toFixed(4) : "N/A"}`,
+                        `  Entropy: ${siom.entropy !== undefined ? Number(siom.entropy).toFixed(4) : "N/A"}`] : []),
                     ``, `Features: ${features.join(", ")}`,
                     summary.selected_columns ? `Selected columns: ${summary.selected_columns.join(", ")}` : "",
                     summary.transforms ? `Transforms: ${Object.entries(summary.transforms).map(([k, v]) => `${k}=${v}`).join(", ")}` : "",
@@ -434,7 +468,7 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
             const showAvailable = files.length > 0 && !(figures === "all" || figures === "images" || figures === "none");
             if (showAvailable) {
                 const logicalNames = jobType2 === "train_som" || jobType2 === "train_siom" || jobType2 === "render_variant"
-                    ? `Logical names: combined, umatrix, hit_histogram, correlation, ${featuresForLog.map((_, i) => `component_${i + 1}`).join(", ")}. `
+                    ? `Logical names: combined, umatrix, hit_histogram, correlation, divergence_kl, ${featuresForLog.map((_, i) => `component_${i + 1}`).join(", ")}. `
                     : "";
                 content.push({ type: "text", text: `Available: ${files.join(", ")}. ${logicalNames}Use results(action=get, figures=[...]) for specific plots, figures=all for everything, or figures="none" for metrics-only.` });
             }
@@ -597,6 +631,7 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
             else {
                 toDownload = files.filter(isImage);
             }
+            toDownload = [...new Set(toDownload)];
             let resolvedDir = sandboxPath(folder, await getWorkspaceRootAsync(server));
             // Always namespace each job's files into its own subfolder so that
             // downloading multiple jobs (or job types) into the same folder never
@@ -664,6 +699,53 @@ NOT FOR: Jobs that haven't completed (use jobs(action=status) first), or tempora
             }
             return { content: [{ type: "text", text: `Recolor job ${newJobId} submitted. Poll with jobs(action=status, job_id="${newJobId}"), then results(action=get, job_id="${newJobId}").` }] };
         }
+        if (action === "feature_divergences") {
+            const body = {
+                metrics: divergenceMetrics ?? ["symmetric_kl", "directed_kl"],
+                output_format: output_format ?? "png",
+                output_dpi: resolveOutputDpi(output_dpi),
+            };
+            if (wassersteinMethod != null)
+                body.wasserstein_method = wassersteinMethod;
+            const data = (await apiCall("POST", `/v1/results/${job_id}/feature_divergences`, body));
+            const newJobId = data.id;
+            const poll = await pollUntilComplete(newJobId, POLL_RECOLOR_MAX_MS);
+            if (poll.status !== "completed") {
+                return {
+                    content: [{
+                            type: "text",
+                            text: `Feature divergences job ${newJobId} submitted (parent ${job_id}). Poll with jobs(action=status, job_id="${newJobId}"), then results(action=get, job_id="${newJobId}").`,
+                        }],
+                };
+            }
+            const res = (await apiCall("GET", `/v1/results/${newJobId}`));
+            const sum = (res.summary ?? {});
+            const topDivs = sum.top_divergences ?? [];
+            const klComp = sum.feature_kl_complexity ?? [];
+            const topW1 = sum.top_wasserstein ?? null;
+            const lines = [
+                `Feature divergences complete — job_id: ${newJobId} (parent training: ${job_id})`,
+                `Method: ${sum.divergence_method ?? "symmetric_kl_min_shift_sum_norm_denormalized_component_planes"}`,
+                `Weight space: ${sum.weight_space ?? "denormalized"}`,
+                `Metrics: ${JSON.stringify(sum.metrics ?? body.metrics)}`,
+                topDivs.length
+                    ? `Top symmetric KL pairs: ${topDivs.map((t) => `${t.f1}&${t.f2}=${typeof t.symmetric_kl === "number" ? Number(t.symmetric_kl).toFixed(3) : t.symmetric_kl}`).join(", ")}`
+                    : "",
+                klComp.length
+                    ? `KL complexity ranking: ${klComp.map((t) => `${t.feature}=${typeof t.mean_outgoing_kl === "number" ? Number(t.mean_outgoing_kl).toFixed(3) : t.mean_outgoing_kl}`).join(", ")}`
+                    : "",
+                topW1 && topW1.length
+                    ? `Top Wasserstein pairs: ${topW1.map((t) => `${t.f1}&${t.f2}=${typeof t.wasserstein === "number" ? Number(t.wasserstein).toFixed(3) : t.wasserstein}`).join(", ")}`
+                    : "",
+                `Artifacts refreshed on parent result (divergence_kl / feature_divergences.json). Compare with Pearson top_correlations from results(get) on the training job.`,
+            ].filter(Boolean);
+            const content = [{ type: "text", text: lines.join("\n") }];
+            const files = sum.files ?? [];
+            for (const f of files.filter((x) => /\.(png|svg)$/i.test(x))) {
+                await tryAttachImage(content, newJobId, f);
+            }
+            return { content };
+        }
         throw new Error("Invalid action");
     });
 }

package/dist/tools/train.js

@@ -1,26 +1,25 @@
 import { z } from "zod";
 import { registerAuditedTool } from "../audit.js";
-import { apiCall } from "../shared.js";
+import { apiCall, ensureDatasetTrainReady } from "../shared.js";
 import { buildTrainSubmitExtras } from "../train_submit_message.js";
 import { TRAINING_MONITOR_URI, buildTrainingMonitorResult, } from "./training_monitor.js";
 import { TRAINING_INPUT_SCHEMA, buildTrainMapParams, buildFloopParams, fetchTrainingPresets, assertValidNormalizationArgs, } from "./training_core.js";
-import { SMALL_N_SOFT, buildSmallNGuardrailLines, suggestedAutoGridSide, } from "../training_scale_guidance.js";
 export const TRAIN_DESCRIPTION = `Submit a self-organizing map training job. Returns a job_id; poll with jobs(action=status) and then results(action=get). All actions are async.
 
 | Action | Use when |
 |--------|----------|
 | map | Standard SOM on a fixed hex grid — complete-case numeric data (no NaNs in training columns). |
 | siom_map | Self-interacting map — same grid flow plus SIOM coverage control (gamma, siom_decay, penalty) to avoid dead nodes. |
-| impute | Sparse training data: trains a missing-tolerant map (accelerated missSOM; SIOM missSOM when model=auto on som_siom+ or model=SIOM) AND returns dense imputed.csv in one job. Plain numeric columns only. Optional SIOM knobs: gamma, siom_decay, siom_penalty, penalty_alpha, reset_per_epoch. |
+| impute | Sparse training data (requires SIOM entitlement): trains a missing-tolerant map (SIOM missSOM when model=auto/SIOM; plain missSOM when model=SOM) AND returns dense imputed.csv in one job. Plain numeric columns only. Optional SIOM knobs: gamma, siom_decay, siom_penalty, penalty_alpha, reset_per_epoch. |
 | floop_siom | FLooP-SIOM — a growing node-budget manifold instead of a fixed grid; default topology=free (CHL). Requires a plan with all_algorithms (Premium/Enterprise). |
 
-First quick map (no separate baseline action): use \`train(action=map, preset=quick)\` when n is large enough (hundreds+ rows); for n<1000 prefer omitting grid_x/grid_y for auto (~sqrt(5·√n)) or an explicit small grid — see **training_guidance** / analyze next_steps. On CPU, preset=quick 15×15 is typically very fast up to ~100k rows — do not subsample first.
+First quick map (no separate baseline action): use \`train(action=map, preset=quick)\` (or \`normalize=mad\` with a grid sized from the preview row count); see the \`baseline_explore\` preset in **training_guidance**.
 
 Parameter details (grid, epochs, batch, model, presets, normalize, categorical_features, backend, FLooP max_nodes/effort, SIOM geometry) live in the **training_guidance** tool / **prepare_training** prompt — call them instead of guessing. Backend strings are "cpu" | "gpu" | "gpu_graphs" (no colon); on CPU-only hosts pass backend=cpu.
 
-ASYNC: every action returns a job_id immediately. Poll jobs(action=status, job_id=...) every 10–15s (do NOT poll faster). Map training typical times: preset=quick 15×15 often tens of seconds even at ~30k–100k rows on CPU | 20×20 ~3–5 min | 40×40 ~15–30 min — do not assume failure before 3 minutes on CPU. When status is completed, call results(action=get, job_id=...).
+ASYNC: every action returns a job_id immediately. Poll jobs(action=status, job_id=...) every 10–15s (do NOT poll faster). Map training typical times: 10×10 ~30s | 20×20 ~3–5 min | 40×40 ~15–30 min — do not assume failure before 3 minutes on CPU. When status is completed, call results(action=get, job_id=...).
 
-action=impute: use when many cells are missing across training columns; use inference(impute_column) instead when you already have a complete-case map and need to fill one held-out column. Supports transforms (log/sqrt/etc., not rank), temporal_features, cyclic_features, optional cyclic_pairs=[[cos,sin],...] (circular prototype refresh + atomic chord BMU; those columns get circular quality metrics only — never quote Euclidean R² on cos/sin), and normalization_methods (zscore/mad/sigmoidal/none — not sepd). Does not support auto_log_transforms, rank transforms, or normalize=sepd. Defaults model=auto (SIOM missSOM on som_siom+ plans, else plain missSOM), cv_folds=5 → quality.csv (linear: kfold_holdout_pool + resubstitution_optimistic; cyclic pairs: kfold_holdout_circular). Aggregate MAE/RMSE excludes cyclic columns. Do not treat k-fold alone as true missing-cell ground truth. Optional SIOM knobs: gamma, gamma_f, siom_decay, siom_penalty, penalty_alpha, reset_per_epoch. Artifacts: imputed.csv, imputation_mask.csv, optional imputation_uncertainty.csv (pool_std = neighbourhood dispersion). A completed impute job_id is a valid parent for inference(impute_column).
+action=impute: use when many cells are missing across training columns; use inference(impute_column) instead when you already have a complete-case map and need to fill one held-out column. Supports transforms (log/sqrt/etc., not rank), temporal_features, cyclic_features, optional cyclic_pairs=[[cos,sin],...] (circular prototype refresh + atomic chord BMU; those columns get circular quality metrics only — never quote Euclidean R² on cos/sin), and normalization_methods (zscore/mad/sigmoidal/none — not sepd). Does not support auto_log_transforms, rank transforms, or normalize=sepd. Requires SIOM entitlement (train_impute). Defaults model=auto (SIOM missSOM; model=SOM for plain missSOM), cv_folds=5 → quality.csv (linear: kfold_holdout_pool + resubstitution_optimistic; cyclic pairs: kfold_holdout_circular). Aggregate MAE/RMSE excludes cyclic columns. Do not treat k-fold alone as true missing-cell ground truth. Optional SIOM knobs: gamma, gamma_f, siom_decay, siom_penalty, penalty_alpha, reset_per_epoch. Artifacts: imputed.csv, imputation_mask.csv, optional imputation_uncertainty.csv (pool_std = neighbourhood dispersion). A completed impute job_id is a valid parent for inference(impute_column).
 action=floop_siom: max_nodes is a TOTAL node budget (not a grid width/area); if coverage collapses, reduce max_nodes before increasing effort.
 NOT FOR: lifecycle (status/list/compare/cancel/delete) → use jobs; scoring data → use inference.
 
@@ -53,6 +52,7 @@ function assertImputeTrainingArgs(args) {
 }
 /** Shared POST /v1/jobs + train submit message extras (map/siom/impute and floop). */
 async function submitTrainJob(opts) {
+    await ensureDatasetTrainReady(opts.dataset_id);
     const submitBody = { dataset_id: opts.dataset_id, params: opts.params };
     if (opts.label && opts.label.trim() !== "")
         submitBody.label = opts.label;
@@ -67,11 +67,6 @@ async function submitTrainJob(opts) {
     if (opts.gridLargeForRows) {
         data.grid_auto_warning = true;
     }
-    const guardrails = opts.guardrailLines ?? [];
-    if (guardrails.length > 0) {
-        data.training_guardrails = guardrails;
-        data.small_n_warning = opts.totalRows > 0 && opts.totalRows < SMALL_N_SOFT;
-    }
     const extras = buildTrainSubmitExtras({
         newJobId,
         totalRows: opts.totalRows,
@@ -95,12 +90,8 @@ async function submitTrainJob(opts) {
     catch {
         /* ignore queue enrichment */
     }
-    if (guardrails.length > 0) {
-        msg += ` ${guardrails.join(" ")}`;
-    }
-    else if (opts.gridLargeForRows) {
-        const autoSide = suggestedAutoGridSide(opts.totalRows);
-        msg += ` Note: Grid may be large for ${opts.totalRows} rows — prefer omitting grid_x/grid_y for auto (~${autoSide}×${autoSide}) to reduce dead nodes.`;
+    if (opts.gridLargeForRows) {
+        msg += ` Note: Grid may be large for ${opts.totalRows} rows (consider grid=auto for fewer dead nodes).`;
     }
     data.message = msg;
     data.suggested_next_step = extras.suggested_next_step;
@@ -205,16 +196,6 @@ export async function runTrain(action, args) {
             typeof transforms === "object" &&
             Object.keys(transforms).length > 0;
         const gridLargeForRows = effectiveGrid && totalRows > 0 && effectiveGrid[0] * effectiveGrid[1] > totalRows * 0.75;
-        const nFeatures = Array.isArray(columns) ? columns.length
-            : Array.isArray(params.columns) ? params.columns.length
-                : undefined;
-        const guardrailLines = buildSmallNGuardrailLines({
-            n: totalRows,
-            nFeatures,
-            preset: typeof preset === "string" ? preset : undefined,
-            gridX: effectiveGrid?.[0],
-            gridY: effectiveGrid?.[1],
-        });
         const resultsHint = action === "impute" ? "view the map and imputed.csv"
             : "view the map and metrics";
         const variantPrefix = action === "siom_map" ? "variant=siom"
@@ -233,7 +214,6 @@ export async function runTrain(action, args) {
             resultsHint,
             impute: action === "impute",
             gridLargeForRows: !!gridLargeForRows,
-            guardrailLines,
         });
     }
     if (action === "floop_siom") {

package/dist/train_submit_message.js

@@ -2,9 +2,10 @@ import { buildVizStandaloneUrl } from "./viz_links.js";
 export function buildTrainSubmitExtras(opts) {
     const { newJobId, totalRows, hasTransforms, prepareJobId, variantPrefix, paramSummary, impute, resultsHint } = opts;
     const monitorUrl = buildVizStandaloneUrl("training-monitor", newJobId);
+    const prepareKind = impute ? "prepare_impute_matrix" : "prepare_training_matrix";
     let msg = `Job submitted (${variantPrefix}, ${paramSummary}). `;
     if (prepareJobId) {
-        msg += `Preprocessing job prepare_training_matrix (${prepareJobId}) runs on worker-io first; train job ${newJobId} starts after it completes. Poll jobs(action=status, job_id="${prepareJobId}") until completed, then poll the train job. `;
+        msg += `Preprocessing job ${prepareKind} (${prepareJobId}) runs on worker-io first; train job ${newJobId} starts after it completes. Poll jobs(action=status, job_id="${prepareJobId}") or jobs(action=status, job_id="${newJobId}") (shows prepare_job_id + waiting_for_prepare) until prepare completes, then poll the train job. `;
     }
     msg += `Training monitor opened for job ${newJobId}`;
     if (monitorUrl)

package/dist/training_monitor_curve.js

@@ -25,7 +25,8 @@ export function teCurveLabel(base) {
  * epochs=[N,0], so convergence_errors is empty while ordering has many batch points.
  */
 export function isFloopJob(data) {
-    const phase = String(data.training_progress_phase ?? data.progress_phase ?? "").toLowerCase();
+    const rawPhase = String(data.training_progress_phase ?? data.progress_phase ?? "").toLowerCase();
+    const phase = rawPhase === "missing" ? "" : rawPhase;
     if (phase === "floop")
         return true;
     const jt = String(data.job_type ?? "").toLowerCase();