bakaano.neuralnet.simulate

Simulation and inference utilities for streamflow prediction.

Role: Prepare simulation inputs and run trained model inference.

bakaano.neuralnet.simulate._load_pysheds_grid()

Import pysheds lazily to avoid import-time backend failures.

bakaano.neuralnet.simulate._open_dataset_with_fallback(nc_path)

Open NetCDF with backend fallback for Colab/Drive compatibility.

bakaano.neuralnet.simulate.split_predictions_by_station(flat_predictions, station_window_counts)

Split flat model outputs into station-aligned chunks.

bakaano.neuralnet.simulate._preview_items(items, limit=8)

Return a compact preview string for notebook progress messages.

bakaano.neuralnet.simulate.coerce_prediction_array(predicted_streamflow)

Convert model outputs to a NumPy float array safe for NumPy ops.

bakaano.neuralnet.simulate.clip_negative_predictions(predicted_streamflow)

Clamp negative predictions to zero after normalizing dtype.

bakaano.neuralnet.simulate.convert_area_normalized_flow(predicted_streamflow, catch_area)

Convert area-normalized depth (mm/day) back to discharge (m3/s).

bakaano.neuralnet.simulate.inverse_log1p_predictions(predicted_streamflow)

Convert model outputs from log1p target space back to linear values.

bakaano.neuralnet.simulate.build_prediction_frame(predicted_streamflow, sim_start)

Return a dated prediction frame after the 365-day model lookback.

bakaano.neuralnet.simulate._plot_grdc_streamflow(observed_streamflow, predicted_streamflow, val_start)

Plot observed vs predicted streamflow for one interactive evaluation run.

bakaano.neuralnet.simulate.evaluate_streamflow_model_interactively(working_dir, study_area, model_path, val_start, val_end, grdc_netcdf=None, routing_method='mfd', catchment_size_threshold=1000, area_normalize=True, log_transform=True, csv_dir=None, lookup_csv=None, id_col='id', lat_col='latitude', lon_col='longitude', date_col='date', discharge_col='discharge', file_pattern='{id}.csv', runoff_output_dir=None)

Interactively evaluate a trained model against one station.

bakaano.neuralnet.simulate.simulate_streamflow(working_dir, study_area, model_path, sim_start, sim_end, latlist, lonlist, routing_method='mfd', area_normalize=True, log_transform=True, runoff_output_dir=None)

Simulate streamflow at arbitrary latitude/longitude points.

bakaano.neuralnet.simulate.simulate_grdc_csv_stations(working_dir, study_area, model_path, sim_start, sim_end, grdc_netcdf=None, routing_method='mfd', csv_dir=None, lookup_csv=None, id_col='id', lat_col='latitude', lon_col='longitude', date_col='date', discharge_col='discharge', file_pattern='{id}.csv', area_normalize=True, log_transform=True, runoff_output_dir=None)

Simulate streamflow for GRDC or CSV-defined station sets.

class bakaano.neuralnet.simulate.PredictDataPreprocessor(working_dir, study_area, sim_start, sim_end, routing_method, grdc_streamflow_nc_file=None, catchment_size_threshold=None, runoff_output_dir=None)

Bases: object

_load_runoff_entries_for_period()

Load routed runoff entries and require exact daily coverage for simulation.

_load_optional_rainfall_entries_for_period()

Load routed rainfall entries if available and fully aligned.

_extract_station_rowcol(lat, lon)

Extract the row and column indices for a given latitude and longitude from given raster file.

Parameters:

• lat (float) – The latitude of the station.

• lon (float) – The longitude of the station.

Returns:

• row (int) – The row index corresponding to the given latitude and longitude.

• col (int) – The column index corresponding to the given latitude and longitude.

_snap_coordinates(lat, lon)

Snap the given latitude and longitude to the nearest river segment based on a river grid.

Parameters:

• lat (float) – The latitude to be snapped.

• lon (float) – The longitude to be snapped.

Returns:

• snapped_lat (float) – The latitude of the nearest river segment.

• snapped_lon (float) – The longitude of the nearest river segment.

_check_point_in_region(olat, olon)

Check whether a single (olat, olon) point lies within a study-area shapefile.

• If NOT inside: raise SystemExit with a formatted, user-facing message

• If inside: print confirmation and do nothing

load_observed_streamflow(grdc_streamflow_nc_file)

Load and filter observed GRDC streamflow data in a schema-robust way. Works for single- and multi-station NetCDFs.

Parameters:

grdc_streamflow_nc_file (str) – Path to GRDC NetCDF file.

Returns:

Filtered GRDC subset for the study area.

Return type:

xarray.Dataset

load_observed_streamflow_from_csv_dir(csv_dir, lookup_csv, id_col='id', lat_col='latitude', lon_col='longitude', date_col='date', discharge_col='discharge', file_pattern='{id}.csv')

Load observed streamflow from per-station CSV files using a lookup table.

The lookup table must include station identifiers and coordinates. The method filters stations to the study area, then loads per-station CSVs by ID.

Parameters:

• csv_dir (str) – Directory containing per-station CSV files.

• lookup_csv (str) – CSV file with station ids and coordinates.

• id_col (str) – Station id column in lookup CSV.

• lat_col (str) – Latitude column in lookup CSV.

• lon_col (str) – Longitude column in lookup CSV.

• date_col (str) – Date column in station CSVs.

• discharge_col (str) – Discharge column in station CSVs.

• file_pattern (str) – Pattern for station CSV filenames (e.g., "{id}.csv").

Returns:

Mapping of station_id to observed discharge DataFrame.

Return type:

dict

get_data()

Extract and preprocess predictor and response variables for each station based on its coordinates.

Returns:

A list containing two elements: - self.data_list: A list of tuples, each containing predictors (DataFrame) and response (DataFrame). - self.catchment: A list of tuples, each containing catchment data (accumulation and slope values).

Return type:

list

get_data_latlng(latlist, lonlist)

Prepare predictors for arbitrary latitude/longitude points.

Parameters:

• latlist (list[float]) – Latitudes to simulate.

• lonlist (list[float]) – Longitudes to simulate.

Returns:

[data_list, catchment, latlist, lonlist].

Return type:

list

class bakaano.neuralnet.simulate.PredictStreamflow(working_dir, area_normalize=True, log_transform=True)

Bases: object

predict(batch_size=None)

Run model inference using the prepared predictor tensors.

predict_station_series(batch_size=None, area_normalize=None)

Run inference and return one prediction array per valid station/point.

load_model_config(model_path)

Load saved model-scale options before preparing inference tensors.

load_default_model_config()

Load default working-dir model config when available.

_load_config_file(config_path)

prepare_data(data_list)

prepare_data_latlng(data_list)

print_prediction_summary(point_count=None)

Print a compact summary of prepared simulation tensors.

load_model(path)

Load a trained regional model from disk.

Parameters:

path (str) – Path to the saved Keras model.

Returns:

Loaded model instance.

Return type:

tensorflow.keras.Model

_align_temporal_feature_count_to_model()

Match prepared temporal feature channels to the loaded model.

summary()

Print a summary of the loaded model.