Cs improvements

.cursor/rules/always.mdc

@@ -11,3 +11,4 @@ alwaysApply: true
 - Preserve deferred-import patterns on hot paths (`ADR-0002`).
 - Do not invent or harden a plugin API (placeholder in `AGENTS.md` until redesigned).
 - Prefer small, localized diffs; add/update tests for stateful behavior.
+- **Never create a git commit unless the user explicitly asks to commit in that message.** Skills that say “commit when done” (e.g. `/implement`) do **not** override this — finish the work, leave changes uncommitted, and wait for an explicit commit request. Do not stage-and-commit “to finish the task,” “because the skill says so,” or “so code-review has a SHA.”

AGENTS.md

@@ -87,6 +87,7 @@ Agents should not assume Linux-only or a single Python version. Details live in
 - When changing stateful behavior (settings, databases, signals), update or add tests and keep the same signal contracts.
 - Preserve deferred-import and `mod/` patching patterns unless an ADR says otherwise.
 - Do not invent plugin APIs (see placeholder above).
+- **Never create a git commit unless the user explicitly asks to commit.** Slash skills that mention committing (e.g. `/implement`) do not override this — leave changes uncommitted until the user requests a commit.
 
 ## Useful references
 

CONTEXT.md

@@ -40,7 +40,7 @@ The quantified output (or input in case of a waste treatment system) of the prod
 
 ### Calculation setup (CS)
 
-A named set of functional unit(s) and LCIA method(s) used to run LCA / multi-LCA / Monte Carlo. Managed under calculation-setup UI and `app` actions. It can also additionaly include scenarios as a third element in the "Scenario" LCA mode. 
+A named set of functional unit(s) and LCIA method(s) used to run LCA / multi-LCA / Monte Carlo. Managed under calculation-setup UI and `app` actions. It can also additionaly include scenarios as a third element in the "Scenario" LCA mode. The CS page always opens in **Standard** mode; switching to **Scenario** mode may reload persisted scenario files (with a loading indication). Scenario file paths, combine mode, and the included scenario-combinations set may be stored on the CS.
 
 ### LCIA method / impact category
 
@@ -74,6 +74,14 @@ Analysis of how uncertain inputs drive output variance (e.g. SALib-based), based
 
 An LCA calculation that also considers multiple scenarios for inventory data (based on the superstructure approach). See `activity_browser/bwutils/superstructure/`.
 
+### Scenario name
+
+The string identifier of a scenario column in a scenario difference file (or in a combined scenario table). Always a string — file importers coerce numeric-looking headers (e.g. Excel `2025`) with `str(...)`. When scenarios from multiple files are combined (product), the combined scenario name is the file-order join of the parts with ` | ` (e.g. `A` and `X` → `A | X`). _Avoid_: scenario header (as a typed value), scenario label when meaning the column identity.
+
+### Scenario-combinations list
+
+The list of product-combined scenario names from two or more loaded scenario difference files (e.g. `A | X`), each with an include checkbox. Shown beside the per-file scenario lists in the calculation setup when 2+ files are loaded under Combine. Together with per-file scenario checkboxes it supports coarse (row/column/slice) and fine (cell) inclusion. Under **Extend**, there is no combinations list: inclusion is by shared scenario name, and checkboxes for the same name stay in sync across files. _Avoid_: scenario-selection-matrix (deferred 2D editor), SS-matrix, combiner matrix.
+
 ### Metadata store
 
 Cached tabular metadata for fast UI search and display (`app.metadata` / `activity_browser.bwutils.metadata`). Synced via `app.signals.metadata` and related meta signals; tests often wait for the metadata loader.

activity_browser/app/actions/calculation_setup/cs_calculate.py

@@ -49,6 +49,18 @@ def run(cs_name: str, scenario_data: pd.DataFrame = None):
                 f"Calculation setup '{cs_name}' has no active impact categories."
             )
 
+        if scenario_data is not None and (
+            scenario_data.empty or len(scenario_data.columns) == 0
+        ):
+            QtWidgets.QMessageBox.warning(
+                app.main_window,
+                "No scenarios selected",
+                "Select at least one scenario before calculating.\n\n"
+                "Use the checkboxes on the scenario files, or the combinations "
+                "list when combining multiple files.",
+            )
+            return
+
         dialog = CalculationDialog(cs_name, app.main_window)
         dialog.show()
         app.application.thread().eventDispatcher().processEvents(QtCore.QEventLoop.ProcessEventsFlag.AllEvents)

activity_browser/app/actions/calculation_setup/cs_duplicate.py

@@ -1,4 +1,5 @@
 from loguru import logger
+import copy
 
 from qtpy import QtWidgets
 
@@ -44,6 +45,6 @@ def run(cs_name: str):
             )
             return
 
-        bd.calculation_setups[new_name] = bd.calculation_setups[cs_name].copy()
+        bd.calculation_setups[new_name] = copy.deepcopy(bd.calculation_setups[cs_name])
         logger.info(f"Copied calculation setup {cs_name} as {new_name}")
         CSOpen.run(new_name)

activity_browser/app/actions/calculation_setup/cs_rename.py

@@ -1,4 +1,5 @@
 from loguru import logger
+import copy
 
 from qtpy import QtWidgets
 
@@ -44,6 +45,6 @@ def run(cs_name: str, new_name: str = None):
             return
 
         # instruct the CalculationSetupController to rename the CS to the new name
-        bd.calculation_setups[new_name] = bd.calculation_setups[cs_name].copy()
+        bd.calculation_setups[new_name] = copy.deepcopy(bd.calculation_setups[cs_name])
         del bd.calculation_setups[cs_name]
         logger.info(f"Renamed calculation setup from {cs_name} to {new_name}")

activity_browser/app/pages/calculation_setup/calculation_setup.py

@@ -87,6 +87,7 @@ def type_switch(self, calculation_type: str):
             self.scenario_section.hide()
         elif calculation_type == "Scenario":
             self.scenario_section.show()
+            self.scenario_section.load_persisted_scenarios()
         else:
             raise ValueError(f"Unknown calculation type: {calculation_type}")