fix(tableau-next-semantic-model-generate): correct broken CLI examples, unqualified syntax, aggregation choices, error docs

skills/tableau-next-semantic-model-generate/SKILL.md

@@ -330,8 +330,12 @@ Format: `fieldApiName:tableApiName` (repeat `--additional-dimension` for multipl
 **Creation success is not data success.** Empty results (empty source, filter excludes everything, expression resolves to nothing) → dashboard shows "No results to show." Check with `--verify-only`:
 
 ```bash
-python scripts/create_metric.py --sdm <SDM> --name <Metric_mtc> --verify-only
-python scripts/create_calc_field.py --sdm <SDM> --type measurement --name <Field_clc> --verify-only
+# --verify-only ignores --label/--calculated-field/--time-* but argparse still requires them; pass any values.
+python scripts/create_metric.py --sdm <SDM> --name <Metric_mtc> \
+  --label "verify" --calculated-field <ANY_clc> \
+  --time-field <ANY_TimeField> --time-table <ANY_TimeTable> \
+  --verify-only
+python scripts/create_calc_field.py --sdm <SDM> --type measurement --name <Field_clc> --label "verify" --verify-only
 ```
 
 Non-zero exit + **NOT shippable** = don't proceed to dashboards. Full protocol (why not to re-POST, `query_data.py --count` for raw objects, empty-source guidance) in **[references/concepts.md](references/concepts.md)** and [references/empty-source-handling.md](references/empty-source-handling.md).
@@ -353,7 +357,7 @@ python scripts/update_sdm.py {{SDM_NAME}} \
 python scripts/update_sdm.py {{SDM_NAME}} --agent-enabled --dry-run
 
 # Confirm via discovery (stored state, not CLI success message)
-python scripts/discover_sdm.py {{SDM_NAME}} --json
+python scripts/discover_sdm.py --sdm {{SDM_NAME}} --json
 ```
 
 The update is a PATCH (partial body — server merges), idempotent. `--categories` takes controlled "Semantic Category" values only — the server rejects unknown values with `Invalid Semantic Category`.

skills/tableau-next-semantic-model-generate/references/api-reference.md

@@ -418,21 +418,34 @@ Downstream builders that consume this contract should enforce this pre-POST and
 
 ## Error Responses
 
-All endpoints return structured error responses:
+Salesforce API error responses come in two shapes depending on the endpoint.
+
+**Array form (most common):** a top-level JSON array of one or more error objects.
+
+```json
+[
+  {
+    "errorCode": "INVALID_API_INPUT",
+    "message": "Field 'Amount' is not found in the semantic model.",
+    "fields": ["dataObjectFields"]
+  }
+]
+```
+
+**Single-object form:** a bare JSON object (used by some endpoints and by the platform for auth/session errors).
 
 ```json
 {
-  "error": {
-    "code": "INVALID_FIELD",
-    "message": "Field 'Amount' not found in semantic model 'Sales_Cloud12_backward'",
-    "details": {
-      "fieldName": "Amount",
-      "availableFields": ["Total_Amount", "Close_Date", "Stage"]
-    }
-  }
+  "errorCode": "INVALID_SESSION_ID",
+  "message": "Session expired or invalid.",
+  "fields": []
 }
 ```
 
+A `localizedMessage` key may appear in place of (or alongside) `message` in platform-level errors.
+
+The `fields` array is present when the error is scoped to specific request fields; it is omitted or empty otherwise.
+
 **Common Error Codes:**
 - `INVALID_TOKEN`: Authentication token is invalid or expired
 - `INSUFFICIENT_PERMISSIONS`: User lacks required permissions

skills/tableau-next-semantic-model-generate/references/metric-design.md

@@ -26,7 +26,7 @@ python scripts/create_calc_field.py \
   --type measurement \
   --name Total_Revenue_clc \
   --label "Total Revenue" \
-  --expression "SUM([Amount])" \
+  --expression "SUM([Opportunity_TAB_Sales_Cloud].[Amount])" \
   --aggregation Sum
 
 # Step 2: Create metric
@@ -141,7 +141,7 @@ python scripts/create_calc_field.py \
   --type measurement \
   --name Win_Rate_clc \
   --label "Win Rate" \
-  --expression "SUM([Won_Count]) / SUM([Total_Count])" \
+  --expression "SUM([Opportunity_TAB_Sales_Cloud].[Won_Count]) / SUM([Opportunity_TAB_Sales_Cloud].[Total_Count])" \
   --aggregation UserAgg
 
 # Step 2: Create metric
@@ -169,7 +169,7 @@ python scripts/create_calc_field.py \
   --type measurement \
   --name Weighted_Pipeline_clc \
   --label "Weighted Pipeline Value" \
-  --expression "SUM([Amount] * [Probability])" \
+  --expression "SUM([Opportunity_TAB_Sales_Cloud].[Amount] * [Opportunity_TAB_Sales_Cloud].[Probability])" \
   --aggregation Sum
 
 # Step 2: Create metric
@@ -197,7 +197,7 @@ python scripts/create_calc_field.py \
   --type measurement \
   --name Avg_Sales_Cycle_clc \
   --label "Average Sales Cycle (Days)" \
-  --expression "AVG(DATEDIFF('day', [Created_Date], [Close_Date]))" \
+  --expression "AVG(DATEDIFF('day', [Opportunity_TAB_Sales_Cloud].[Created_Date], [Opportunity_TAB_Sales_Cloud].[Close_Date]))" \
   --aggregation Avg
 
 # Step 2: Create metric
@@ -375,12 +375,12 @@ After creating a metric, test it by:
 
 | Pattern | Calc Field Expression | Aggregation | Use Case |
 |---------|----------------------|-------------|----------|
-| Simple sum | `SUM([Field])` | Sum | Total revenue, total count |
-| Average | `AVG([Field])` | Avg | Average deal size, average duration |
-| Ratio | `SUM([A]) / SUM([B])` | UserAgg | Win rate, conversion rate |
-| Weighted | `SUM([A] * [B])` | Sum | Weighted pipeline, risk-adjusted |
-| Time calc | `AVG(DATEDIFF('day', [Start], [End]))` | Avg | Sales cycle, time to close |
-| Conditional sum | `SUM(IF condition THEN [Field] ELSE 0 END)` | Sum | Won revenue, qualified leads |
+| Simple sum | `SUM([Table].[Field])` | Sum | Total revenue, total count |
+| Average | `AVG([Table].[Field])` | Avg | Average deal size, average duration |
+| Ratio | `SUM([Table].[A]) / SUM([Table].[B])` | UserAgg | Win rate, conversion rate |
+| Weighted | `SUM([Table].[A] * [Table].[B])` | Sum | Weighted pipeline, risk-adjusted |
+| Time calc | `AVG(DATEDIFF('day', [Table].[Start], [Table].[End]))` | Avg | Sales cycle, time to close |
+| Conditional sum | `SUM(IF condition THEN [Table].[Field] ELSE 0 END)` | Sum | Won revenue, qualified leads |
 
 ## Metric Lifecycle
 

skills/tableau-next-semantic-model-generate/scripts/_shared/calc_field_templates.py

@@ -119,7 +119,7 @@ def build_calculated_measurement(
         api_name: API name (must end with _clc)
         label: Display label
         expression: Tableau formula expression
-        aggregation_type: Aggregation type (Sum, Avg, Count, Min, Max, UserAgg, Median)
+        aggregation_type: Aggregation type (Sum, Avg, Count, CountDistinct, Min, Max, UserAgg)
         data_type: Data type (Number, Text, Boolean, DateTime)
         decimal_place: Decimal places for Number fields
         description: Field description
@@ -155,11 +155,11 @@ def build_calculated_measurement(
         final_agg_type = aggregation_type
         level = "Row"
         # Validate explicit aggregation types are allowed
-        if final_agg_type not in ("Sum", "Avg", "Count", "Min", "Max", "Median", "UserAgg"):
+        if final_agg_type not in ("Sum", "Avg", "Count", "CountDistinct", "Min", "Max", "UserAgg"):
             import warnings
             warnings.warn(
                 f"Explicit aggregation type '{final_agg_type}' may not be supported. "
-                f"Common types: Sum, Avg, Count, Min, Max, Median"
+                f"Common types: Sum, Avg, Count, CountDistinct, Min, Max"
             )
     
     payload = {
@@ -262,7 +262,7 @@ def validate_calc_field(
 
     # Check aggregation type
     if field_type == "measurement" and aggregation_type:
-        valid_agg = ["Sum", "Avg", "Count", "Min", "Max", "UserAgg", "Median"]
+        valid_agg = ["Sum", "Avg", "Count", "CountDistinct", "Min", "Max", "UserAgg"]
         if aggregation_type not in valid_agg:
             errors.append(f"aggregationType must be one of: {', '.join(valid_agg)}")
 

skills/tableau-next-semantic-model-generate/scripts/create_calc_field.py

@@ -139,7 +139,7 @@ def main() -> int:
 
     # Measurement-specific
     parser.add_argument("--aggregation", default="UserAgg",
-                        choices=["Sum", "Avg", "Count", "Min", "Max", "UserAgg", "Median"],
+                        choices=["Sum", "Avg", "Count", "CountDistinct", "Min", "Max", "UserAgg"],
                         help="Aggregation type (for measurements, default: UserAgg)")
     parser.add_argument("--data-type", choices=["Number", "Text", "Boolean", "DateTime"],
                         help="Data type (default: Number for measurements, inferred for dimensions)")