feat(tasks)!: remove n2o1 task + ConfigEntsoe inherits ConfigMulti

_quarto.yml

@@ -421,8 +421,6 @@ website:
             contents:
               - text: "task_safe_demo"
                 file: docs/reference/tasks.task_safe_demo.qmd
-              - text: "task_safe_n_to_1_with_covariates_and_dataframe"
-                file: docs/reference/tasks.task_safe_n_to_1_with_covariates_and_dataframe.qmd
 
       - section: "Processing Guides"
         contents:
@@ -791,4 +789,3 @@ quartodoc:
         Executable tasks for demonstration and production pipelines.
       contents:
         - tasks.task_safe_demo
-        - tasks.task_safe_n_to_1_with_covariates_and_dataframe

docs/reference/index.qmd

@@ -292,5 +292,4 @@ Executable tasks for demonstration and production pipelines.
 
 | | |
 | --- | --- |
-| [tasks.task_safe_demo](tasks.task_safe_demo.qmd#spotforecast2_safe.tasks.task_safe_demo) | Task demo: compare baseline, covariate, and custom LightGBM forecasts against ground truth. |
-| [tasks.task_safe_n_to_1_with_covariates_and_dataframe](tasks.task_safe_n_to_1_with_covariates_and_dataframe.qmd#spotforecast2_safe.tasks.task_safe_n_to_1_with_covariates_and_dataframe) | Thin ConfigMulti-driven entry point for N-to-1 forecasting. |
+| [tasks.task_safe_demo](tasks.task_safe_demo.qmd#spotforecast2_safe.tasks.task_safe_demo) | Task demo: compare baseline, covariate, and custom LightGBM forecasts against ground truth. |

docs/safe/spotforecast2-safe.qmd

@@ -11,7 +11,7 @@ In safety-critical environments, reducing the "dead code" and unnecessary depend
 ---
 
 ## Positive List (Retained Components)
-The following files are essential for the execution of the primary workflows: `task_safe_demo.py` and `task_safe_n_to_1_with_covariates_and_dataframe.py`.
+The following files are essential for the execution of the primary workflow `task_safe_demo.py`. (The `task_safe_n_to_1_with_covariates_and_dataframe.py` console task was removed in 20.0.0; the `multitask` pipeline it wrapped — `BaseTask`/`MultiTask`/`runner.run` — is retained as public API and is exercised by the downstream `spotforecast2` package.)
 
 ### Orchestration & Pipelines
 - `src/spotforecast2_safe/processing/n2n_predict_with_covariates.py`
@@ -74,13 +74,13 @@ To maintain a green build and avoid import errors, the following non-essential t
 The resulting `spotforecast2_safe` project is a hardened version of the original, with $0$ unreachable code paths for the specified tasks and $100\%$ test coverage on the remaining logic.
 
 ## Essential Classes and Functions (Positive List)
-The following classes and functions (including internal helpers) are strictly required for the execution of `task_safe_demo.py` and `task_safe_n_to_1_with_covariates_and_dataframe.py`:
+The following classes and functions (including internal helpers) are strictly required for the execution of `task_safe_demo.py` and the public `multitask` pipeline:
 
 ### Orchestration & Processing
 - `agg_predict` (Function)
 - `n2n_predict` (Function)
 - `n2n_predict_with_covariates` (Function) — used by `task_safe_demo`
-- `run_pipeline` (Function, `task_safe_n_to_1_with_covariates_and_dataframe`) — delegates to `multitask.runner.run`
+- `run` (Function, `multitask.runner`) — public single-call pipeline entry point
 
 ### Multitask Pipeline (n-to-1 task)
 - `BaseTask` (Class) — shared prepare/detect/impute/exog steps

docs/tasks/tasks.qmd

@@ -7,7 +7,6 @@
 | Command | Description |
 |---------|-------------|
 | `spotforecast-safe-demo` | Demo task comparing baseline, covariate, and custom LightGBM forecasts |
-| `spotforecast-safe-n2o1-cov-df` | N-to-1 forecasting with exogenous covariates and DataFrame input |
 
 ---
 
@@ -49,63 +48,6 @@ uv run spotforecast-safe-demo --logging true
 
 ---
 
-## N-to-1 with Covariates and DataFrame
-
-The `spotforecast-safe-n2o1-cov-df` command runs the `ConfigMulti`-driven
-multitask pipeline.  It delegates to `spotforecast2_safe.multitask.runner.run`
-with `task="lazy"`.  All pipeline parameters are centralised in a `ConfigMulti`
-object; aggregation weights come from `ConfigMulti.agg_weights` (replacing the
-former hard-coded `DEFAULT_WEIGHTS` constant).
-
-See the `run_pipeline` and `_build_config_from_cli` API reference for a full
-description of the pipeline stages and `ConfigMulti` field mappings.
-
-### Features
-
-- `ConfigMulti` as single source of truth — horizon, training window, outlier
-  policy, feature flags, and aggregation weights in one object.
-- Multitask pipeline stages: prepare_data, detect_outliers, impute,
-  build_exogenous_features, run(task="lazy").
-- Model persistence: fitted models are saved under `cache_home/models/` and
-  can be reloaded with `task="predict"`.
-
-### Usage
-
-```bash
-# Run with custom settings (a ConfigMulti and explicit data path are required)
-uv run spotforecast-safe-n2o1-cov-df
-
-# Custom forecast horizon
-uv run spotforecast-safe-n2o1-cov-df --forecast_horizon 48
-
-# Enable holiday features and verbose output
-uv run spotforecast-safe-n2o1-cov-df --include_holiday_features true --verbose true
-
-# Custom aggregation weights
-uv run spotforecast-safe-n2o1-cov-df --weights 1.0 1.0 -1.0
-
-# Specify cache directory for models and logs
-uv run spotforecast-safe-n2o1-cov-df --log_dir ~/my_cache
-```
-
-### Parameters
-
-| Parameter | Default | Description |
-|-----------|---------|-------------|
-| `--forecast_horizon` | 24 | Number of steps ahead to forecast (`predict_size`) |
-| `--lags` | 24 | Lag depth N; expands to `lags_consider=range(1, N+1)` |
-| `--train_ratio` | 0.8 | Fraction of data used for training |
-| `--contamination` | 0.01 | Outlier contamination parameter |
-| `--window_size` | 72 | Rolling window size for imputation (hours) |
-| `--include_holiday_features` | false | Enable holiday indicator features |
-| `--include_holiday_adjacency_features` | false | Enable Brückentag features |
-| `--poly_features_degree` | 1 | Polynomial interaction degree (1 = off) |
-| `--weights` | None | Space-separated aggregation weights (`agg_weights`) |
-| `--verbose` | false | Enable detailed output |
-| `--log_dir` | None | Cache directory for models and logs (`cache_home`) |
-
----
-
 ## Configuration
 
 All tasks use sensible defaults but can be customized via:
@@ -117,7 +59,6 @@ All tasks use sensible defaults but can be customized via:
 ```bash
 # View available options for any command
 uv run spotforecast-safe-demo --help
-uv run spotforecast-safe-n2o1-cov-df --help
 ```
 
 ---
@@ -142,9 +83,6 @@ Safety-critical tasks support comprehensive logging:
 ```bash
 # Enable logging to default directory
 uv run spotforecast-safe-demo --logging true
-
-# Specify custom log directory
-uv run spotforecast-safe-n2o1-cov-df --log_dir /var/log/spotforecast
 ```
 
 Log files include:

docs/tutorials/n2n_predict_with_covariates_explained.qmd

@@ -714,11 +714,11 @@ convention is used with the eleven-element vector
 aggregation in which the first, second, fifth, seventh, eighth, ninth, and
 eleventh columns are added and the remaining columns are subtracted.
 
-The n2n pipeline remains available and is used by `task_safe_demo`.  The n-to-1
-task (`tasks/task_safe_n_to_1_with_covariates_and_dataframe.py`) now runs on
-`spotforecast2_safe.multitask`; see the
-[API reference](../reference/tasks.task_safe_n_to_1_with_covariates_and_dataframe.qmd)
-for its `ConfigMulti`-driven workflow and `agg_weights` configuration.
+The n2n pipeline remains available and is used by `task_safe_demo`.  The
+`ConfigMulti`-driven, `agg_weights`-configurable n-to-1 workflow it describes now
+runs directly on `spotforecast2_safe.multitask` (`MultiTask` /
+`multitask.runner.run`); the former `spotforecast-safe-n2o1-cov-df` console task
+was removed in 20.0.0.
 
 ```{python}
 from spotforecast2_safe.processing.agg_predict import agg_predict

pyproject.toml

@@ -80,7 +80,6 @@ dev = [
 
 [project.scripts]
 spotforecast-safe-demo = "spotforecast2_safe.tasks.task_safe_demo:main"
-spotforecast-safe-n2o1-cov-df = "spotforecast2_safe.tasks.task_safe_n_to_1_with_covariates_and_dataframe:main"
 
 [tool.uv]
 # Force uv lock to resolve dependencies for every platform we ship on,