fix: lazy flows-DB engine in simulator (v2.0.5)

CHANGELOG.md

@@ -5,6 +5,19 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [2.0.5] - 2026-05-07
+
+### Fixed
+
+- `skypro simulate` now skips constructing a SQLAlchemy engine for the
+  flows database when all imbalance data sources are `csvTimeseries`.
+  Previously the engine was created eagerly and unconditionally, so a
+  placeholder or unparseable `env_config["flows"]["dbUrl"]` raised
+  before any simulation logic ran — blocking standalone CSV-only
+  configuration bundles. When at least one source is `flowsMarketData`
+  the engine is still constructed, with a clearer error if `dbUrl` is
+  missing.
+
 ## [2.0.0] - Unreleased
 
 ### Breaking Changes

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "skypro"
-version = "2.0.4"
+version = "2.0.5"
 description = "Skyprospector by Cepro"
 authors = ["damonrand <damon@cepro.energy>"]
 license = "MIT"

src/skypro/commands/simulator/main.py

@@ -510,6 +510,33 @@ def _get_rates_from_config(
     and a dataframe containing live and final imbalance data.
     """
 
+    # Build the flows DB engine only if at least one imbalance source is
+    # `flowsMarketData`. CSV-only configs (`csvTimeseries` for all four
+    # sources) don't need the engine — `get_timeseries` accepts
+    # `db_engine=None` and only forwards it down the flowsMarketData
+    # branch. Mirrors the rates-DB gating pattern used below for
+    # `rates_config.{live,final}.rates_db`.
+    imbalance_sources = [
+        rates_config.final.imbalance_data_source.price,
+        rates_config.final.imbalance_data_source.volume,
+        rates_config.live.imbalance_data_source.price,
+        rates_config.live.imbalance_data_source.volume,
+    ]
+    needs_flows_db = any(
+        s.flows_market_data_source is not None for s in imbalance_sources
+    )
+    if needs_flows_db:
+        try:
+            flows_db_url = env_config["flows"]["dbUrl"]
+        except KeyError:
+            raise SystemExit(
+                "simulate.yaml references a flowsMarketData imbalance source, but "
+                "env_config['flows']['dbUrl'] is not set in the environment file."
+            )
+        db_engine = sqlalchemy.create_engine(flows_db_url)
+    else:
+        db_engine = None
+
     def read_imbalance_data(source: TimeseriesDataSource, context: str):
         """
         Convenience function for reading imbalance data
@@ -519,7 +546,7 @@ def read_imbalance_data(source: TimeseriesDataSource, context: str):
             start=time_index[0],
             end=time_index[-1],
             file_path_resolver_func=file_path_resolver_func,
-            db_engine=sqlalchemy.create_engine(env_config["flows"]["dbUrl"]),
+            db_engine=db_engine,
             context=context
         )
         for notice in notices:

src/tests/integration/fixtures/env_no_flows_db.json

@@ -0,0 +1,15 @@
+{
+  "_doc": "Variant of env.json with an unparseable flows.dbUrl. Used to verify that skypro simulate does NOT eagerly construct a SQLAlchemy engine for the flows DB when all imbalance sources are csvTimeseries (the case for standalone CSV-only handoff bundles). See test_integration_simulator.py::test_integration_csv_only_no_flows_db.",
+  "vars": {
+  },
+  "flows": {
+    "dbUrl": "REPLACE_OR_REMOVE_NOT_USED_BY_SIMULATE"
+  },
+  "flux": {
+    "dbUrl": "postgresql://username:password@someserver.com:5687/database",
+    "schema": "flows"
+  },
+  "rates": {
+    "dbUrl": "postgresql://username:password@someserver.com:5687/database"
+  }
+}

src/tests/integration/test_integration_simulator.py

@@ -131,3 +131,49 @@ class SubTest:
                     0,
                     msg=f"Summary value out of tolerance:\n{error.transpose()}"
                 )
+
+    def test_integration_csv_only_no_flows_db(self):
+        """
+        Regression test for the eager-flows-DB bug fixed in v2.0.5.
+
+        The integration fixture has only csvTimeseries imbalance sources
+        and file-based rates — no flowsMarketData, no rates_db. This
+        means env_config["flows"]["dbUrl"] is never queried at runtime.
+        Before v2.0.5, the simulator constructed a SQLAlchemy engine for
+        the flows DB unconditionally, so an unparseable placeholder URL
+        in flows.dbUrl would crash the run before any simulation logic
+        executed — blocking standalone CSV-only handoff bundles.
+
+        env_no_flows_db.json sets flows.dbUrl to the literal placeholder
+        string "REPLACE_OR_REMOVE_NOT_USED_BY_SIMULATE" — SQLAlchemy
+        can't parse it as a URL. With the v2.0.5 fix, the simulator
+        skips engine construction when no source is flowsMarketData and
+        the run completes successfully.
+        """
+        print(
+            "\n\n\n\nSTARTING SIMULATION INTEGRATION TEST 'csv_only_no_flows_db' "
+            "- - - - - - - - - - - - - - - - - - - - -"
+        )
+
+        res = subprocess.run([
+            'python3',
+            './src/skypro/main.py',
+            'simulate',
+            '--env',
+            './src/tests/integration/fixtures/env_no_flows_db.json',
+            '-y',
+            '--config',
+            './src/tests/integration/fixtures/simulation/config.yaml',
+            '--sim',
+            'integrationTestPriceCurve',
+        ])
+
+        self.assertEqual(
+            res.returncode,
+            0,
+            msg=(
+                "Simulator exited non-zero with an unparseable flows.dbUrl "
+                "in the env file, despite all imbalance sources being "
+                "csvTimeseries. Pre-v2.0.5 regression."
+            ),
+        )