API Reference¶

This page provides an auto-generated API reference for the key components of the hydrodatasource library.

Data Resolver (Unified Data Interface)¶

`hydrodatasource.configs.data_resolver.open_dataset(dataset_id, *, source=None, ctx=None, **reader_kwargs)` ¶

Resolve a dataset id and return an instantiated reader object.

Thin wrapper around hydrodataset's open_dataset that injects hydrodatasource-specific datasets (HDS_DATASETS) and reader aliases (_HDS_READER_ALIASES) into the resolution context. Both hydrodataset datasets (e.g. 'camels_us') and hydrodatasource datasets (e.g. 'songliao_event') are supported via a single call.

Parameters¶

dataset_id : str Dataset identifier from the registry (e.g. 'camels_us', 'songliao_event'). source : str, optional Storage backend: 'local' or 'cloud'. When None (default), falls back to storage.default_source from the settings. ctx : ResolverContext, optional Resolution context. When None, HDS defaults are used automatically. **reader_kwargs Extra keyword arguments forwarded to the reader constructor (e.g. time_unit=["1D"] for SelfMadeHydroDataset).

Returns¶

object An instance of the reader class registered for dataset_id.

Raises¶

DatasetResolutionError If any resolution step fails.

Examples¶

ds = open_dataset("songliao_event") ds = open_dataset("camels_us", source="cloud") ds = open_dataset("songliao_event", time_unit=["1D"])

`hydrodatasource.configs.data_resolver.resolve_data_path(dataset_id, *, source=None, ctx=None)` ¶

Resolve a dataset id to an absolute data path (URI).

Thin wrapper around hydrodataset's resolve_data_path that injects hydrodatasource-specific datasets (HDS_DATASETS) and reader aliases (_HDS_READER_ALIASES) into the resolution context.

Parameters¶

dataset_id : str Dataset identifier from the registry (e.g. 'camels_us', 'songliao_event'). source : str, optional Storage backend: 'local' or 'cloud'. When None (default), falls back to storage.default_source from the settings. ctx : ResolverContext, optional Resolution context bundling project_root, storage config, registry overrides, and extra aliases. When None (default), a new context is created with HDS_DATASETS and _HDS_READER_ALIASES injected. When provided, caller's extras are merged with HDS defaults.

Returns¶

str Absolute URI pointing to the dataset's data directory.

Raises¶

DatasetResolutionError If any resolution step fails.

Examples¶

Default: zero-boilerplate, reads ~/hydro_setting.yml¶

uri = resolve_data_path("songliao_event")

Custom storage root via ResolverContext¶

ctx = ResolverContext(storage={"local": {"root": "/custom/data"}}) uri = resolve_data_path("songliao_event", ctx=ctx)

READER_ALIASES maps every reader alias to its module/class, and HDS_DATASETS is the in-code registry of hydrodatasource datasets (e.g. songliao_event). ResolverContext and DatasetResolutionError are re-exported from hydrodataset.

Reader¶

HydroData (base class)¶

`hydrodatasource.reader.data_source.HydroData` ¶

Bases: ABC

Abstract base class for reading multi-modal hydrological data sources.

Single construction pattern: HydroData(uri).

Parameters¶

uri : str or Path Absolute path or S3 URI pointing directly to the data directory.

SelfMadeHydroDataset¶

`hydrodatasource.reader.data_source.SelfMadeHydroDataset` ¶

Bases: HydroData

A class for reading hydrodataset, but not really ready-datasets, just some data directorys organized like a HydroDataset.

NOTE: We compile forcing data and attr data into a directory, organized like a ready dataset -- like Caravan. Only two directories are needed: attributes and timeseries

`init(uri=None, time_unit=None, **kwargs)` ¶

Initialize a self-made Caravan-style dataset.

Parameters¶

uri : str, optional Absolute URI pointing directly to the data directory. time_unit : list, optional Time units to process, by default None. kwargs : dict, optional Additional keyword arguments, by default None.

`cache_attributes_xrdataset()` ¶

Convert all the attributes to a single dataset

Returns¶

None

`cache_timeseries_xrdataset(**kwargs)` ¶

Save all timeseries data in separate NetCDF files for each time unit.

Parameters¶

t_range : list, optional Time range for the data, by default ["1980-01-01", "2023-12-31"] kwargs : dict, optional batchsize -- Number of basins to process per batch, by default 100 time_units -- List of time units to process, by default None start0101_freq -- for freq setting, if the start date is 01-01, set True, by default False offset_to_utc -- whether to offset the time to UTC, by default False start_hour_in_a_day -- the start hour in a day (0-23), by default 2 which means 2-5-8-11-14-17-20-23 UTC. Chinese basins data always use 08:00 with Beijing Time, so we set the default value to 2. Only applicable for sub-daily intervals (currently only "3h" is supported)

`cache_xrdataset(t_range=None, time_units=None)` ¶

Save all data in a netcdf file in the cache directory

`get_attributes_cols()` ¶

the constant cols in this data_source

`get_timeseries_cols()` ¶

the relevant cols in this data_source

`read_area(gage_id_lst=None)` ¶

read area of each basin/unit

`read_attributes(object_ids=None, constant_cols=None, **kwargs)` ¶

2d data (site_num * var_num), non-time-series data

`read_mean_prcp(gage_id_lst=None, unit='mm/d')` ¶

read mean precipitation of each basin default unit is mm/d, but one can chose other units and we will convert the unit to the specified unit

Parameters¶

gage_id_lst : list, optional the list of gage ids, by default None unit : str, optional the unit of precipitation, by default "mm/d"

Returns¶

xr.Dataset the mean precipitation of each basin

`read_timeseries(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

Returns a dictionary containing data with different time scales.

Parameters¶

object_ids : list, optional List of object IDs. Defaults to None. t_range_list : list, optional List of time ranges. Defaults to None. relevant_cols : list, optional List of relevant columns. Defaults to None. **kwargs : dict, optional Additional keyword arguments. time_units : list, optional List of time units to process start0101_freq : bool, optional For freq setting, if the start date is 01-01, set True offset_to_utc : bool, optional Whether to offset the time to UTC start_hour_in_a_day : int, optional The start hour in a day for sub-daily intervals (0-23). Default is 2.

Returns¶

dict A dictionary containing data with different time scales.

`read_ts_xrdataset(gage_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

Read time-series xarray dataset from multiple NetCDF files and organize them by time units.

Parameters:¶

gage_id_lst: list List of gage IDs to select. t_range: list List of two elements [start_time, end_time] to select time range. var_lst: list List of variables to select. **kwargs Additional arguments.

Returns:¶

dict: A dictionary where each key is a time unit and each value is an xarray.Dataset containing the selected gage IDs, time range, and variables.

`hydrodatasource.reader.data_source.LongTermDataset` ¶

Bases: SelfMadeHydroDataset

`cache_global_dataset(object_ids=None, t_range_list=None)` ¶

读取 CSV 数据并将其转换为 xarray 数据集并保存为 NetCDF 文件。

`hydrodatasource.reader.data_source.SelfMadeForecastDataset` ¶

Bases: SelfMadeHydroDataset

For selfmadehydrodataset, we design a new file format for forecast data from GFS et al.

`init(uri=None, time_unit=None, **kwargs)` ¶

Initialize a class for reading forecast data.

Parameters¶

uri : str Absolute path or S3 URI pointing to the forecast data directory. time_unit : list, optional Unit of one time period, by default None.

`cache_forecast_xrdataset(t_range=None, **kwargs)` ¶

Save all forecast data in separate NetCDF files for each batch of basins and time units.

Parameters¶

t_range : list, optional Time range for the forecast_date, by default None kwargs : dict, optional batchsize -- Number of basins to process per batch, by default 100 variables -- List of variables to process, by default None time_units -- List of time units to process, by default self.time_unit prefix -- Prefix for the NetCDF file names, by default self.dataset_name

`read_forecast(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

Read forecast data (hourly/daily forecasts) from CSV files, where each basin has its own file named after the basin_id. The time range in the parameters refers to the forecast_date (i.e., the target period of the forecast), not the date (the execution time of the forecast).

Parameters¶

object_ids : list List of basin IDs. t_range_list : list Time range for the target forecast period [start_time, end_time]. relevant_cols : list List of variable names to be read. Returns

dict {basin_id: pd.DataFrame}, where each basin has a DataFrame filtered by the time range and variables.

`read_forecast_xrdataset(gage_id_lst, t_range, var_lst, **kwargs)` ¶

read cache nc file

Parameters¶

gage_id_lst : list the list of gage ids t_range : list the start time and end time variables : list variables list

Returns¶

xr.Dataset forecast data

`read_ts_xrdataset(gage_id_lst, t_range, var_lst, **kwargs)` ¶

读取预见期数据

Parameters¶

gage_id_lst : list 流域ID列表 t_range : datetime time range [start_time, end_time] var_lst : list 变量列表

Returns¶

xr.Dataset 预见期数据

`set_data_source_describe()` ¶

set data source description

Returns¶

dict a dict with name and path of the data source

`hydrodatasource.reader.data_source.StationHydroDataset` ¶

Bases: SelfMadeHydroDataset

A class for reading hydrodataset with additional station data.

This class extends SelfMadeHydroDataset to handle datasets that include a stations folder containing individual station data and basin-station relationship information.

Directory structure: - attributes/ - shapes/ - timeseries/ - stations/ - 1D/ # Daily data for all stations - 3h/ # 3h data for all stations - basin_station_info/ # Basin-station relationship info - all_basin_station_mapping.csv - basin_summary.csv - basin_xxx_stations.csv - adjacency_xxx_True.csv

`init(uri=None, time_unit=None, **kwargs)` ¶

Initialize StationHydroDataset.

Parameters¶

uri : str Absolute path or S3 URI pointing to the station dataset directory. time_unit : list, optional Time units for the data, by default None **kwargs : dict Additional keyword arguments passed to parent class

`cache_adjacency_xrdataset()` ¶

Cache adjacency matrices for all basins as NetCDF files.

`cache_all_station_data(**kwargs)` ¶

Cache all station-related data including timeseries, info and adjacency.

`cache_station_info_xrdataset()` ¶

Cache station information and basin-station relationships.

`cache_station_timeseries_xrdataset(**kwargs)` ¶

Cache all station timeseries data in separate NetCDF files.

Parameters¶

**kwargs : dict batchsize -- Number of stations to process per batch, by default 100 time_units -- List of time units to process, by default None start0101_freq -- Whether to use start0101 frequency, by default False

`get_stations_by_basin(basin_id)` ¶

Get all station IDs for a specific basin.

Parameters¶

basin_id : str Basin ID

Returns¶

list List of station IDs in the basin

`read_adjacency_xrdataset(basin_id)` ¶

Read cached adjacency matrix for a specific basin from NetCDF file.

Parameters¶

basin_id : str Basin ID

Returns¶

xr.Dataset Adjacency matrix dataset for the basin

`read_basin_adjacency(basin_id)` ¶

Read adjacency matrix for a specific basin.

Parameters¶

basin_id : str Basin ID

Returns¶

pd.DataFrame Adjacency matrix for stations in the basin

`read_basin_stations(basin_id)` ¶

Read detailed station information for a specific basin.

Parameters¶

basin_id : str Basin ID

Returns¶

pd.DataFrame Station details for the basin

`read_station_info()` ¶

Read basic station information and basin-station mapping.

`read_station_info_xrdataset(**kwargs)` ¶

Read station information from cached NetCDF files.

Returns¶

tuple (basin_station_mapping_dataset, basin_summary_dataset)

`read_station_object_ids()` ¶

Get all station IDs.

`read_station_timeseries(station_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

Read timeseries data for stations.

Parameters¶

station_ids : list, optional List of station IDs, by default None t_range_list : list, optional Time range [start_time, end_time], by default None relevant_cols : list, optional List of relevant columns, by default None **kwargs : dict Additional keyword arguments

Returns¶

dict Dictionary containing data with different time scales

`read_station_ts_xrdataset(station_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

Read station timeseries data from cached NetCDF files.

Parameters¶

station_id_lst : list, optional List of station IDs to select, by default None t_range : list, optional Time range [start_time, end_time], by default None var_lst : list, optional List of variables to select, by default None **kwargs : dict Additional arguments

Returns¶

dict Dictionary with time units as keys and xarray.Dataset as values

`set_data_source_describe()` ¶

Set data source description including stations directory.

`hydrodatasource.reader.data_source.TgHydroDatasource` ¶

Bases: SelfMadeHydroDataset

TG流域数据集 - 继承自SelfMadeHydroDataset，添加LSTM预测数据和图网络结构支持

该类在标准的水文数据集基础上，增加了对LSTM预测数据和图网络结构的支持，主要用于图神经网络相关的水文建模任务。 catch中应包含LSTM预测结果NC文件，文件名格式为lstmpred.nc

数据目录结构: - attributes/ # 控制流域属性数据 - timeseries/ # 控制流域时间序列数据 - shapes/ # 控制流域形状文件 - intermediate/ # 区间流域数据 ├── attributes/ # 区间流域属性数据，包含拓扑关系 ├── timeseries/ # 区间流域时间序列数据 └── shapes/ # 区间流域形状文件

`init(uri=None, time_unit=None, **kwargs)` ¶

Initialize TG basin dataset.

Parameters¶

uri : str Absolute path or S3 URI pointing to the TG dataset directory. time_unit : list, optional Time unit list, by default None. **kwargs : dict Other parameters.

`cache_intermediate_attributes_xrdataset()` ¶

Convert intermediate attributes to a single dataset and cache as NetCDF.

`cache_intermediate_timeseries_xrdataset(**kwargs)` ¶

Cache intermediate timeseries into NetCDF files per time unit, similar to parent cache_timeseries_xrdataset.

`classify_nodes(basin_names=None)` ¶

分类节点为源头节点和汇流节点。

默认在全局图 self.dg 上运行；
若 basin_names 为 None，则使用图中的所有节点，按数据集顺序（与 read_object_ids 对齐）。

Returns¶

tuple (source_nodes, confluence_nodes, node_mask)

`get_name()` ¶

返回数据源名称

`read_attr_xrdataset(gage_id_lst=None, var_lst=None, **kwargs)` ¶

读取属性数据，兼容基础目录与 intermediate 目录的缓存。

默认沿用父类逻辑读取基础缓存；
当设置 prefer_intermediate=True 或 combine=True 时，读取/生成 intermediate 缓存；
intermediate 缓存以 dataset_name_intermediate 作为前缀独立保存。

`read_graph_data(object_ids=None, as_numpy=False, include_optional=True)` ¶

统一返回下游所需的图拓扑结构，满足最大兼容性约定。

参数¶

object_ids : list, optional 指定参与图构建的节点集合；不传则使用数据集可用节点与图节点的交集 as_numpy : bool, default False 将 edge_index 返回为 numpy.int64 的 ndarray；默认返回 torch.long 的 Tensor include_optional : bool, default True 是否包含可选扩展字段（目前支持 upstream_indices）

返回¶

dict { 'edge_index': [2, E] 的 torch.LongTensor 或 numpy.ndarray, 'node_ids': List[str]，长度为 num_nodes，顺序与索引一致, 'node_mask': List[bool]，长度为 num_nodes（汇流 True / 源头 False）, 可选：'upstream_indices': List[List[int]] }

`read_intermediate_timeseries(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

Read timeseries from the intermediate/timeseries directory.

Parameters¶

object_ids : list, optional Basin IDs to read t_range_list : list, optional [start_time, end_time] relevant_cols : list, optional Variables to read **kwargs : dict time_units, start0101_freq, offset_to_utc, start_hour_in_a_day

Returns¶

dict

`read_ts_xrdataset(gage_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

读取时间序列数据，兼容基础目录与 intermediate 目录的缓存读取。

默认沿用父类逻辑读取基础缓存；
当设置 prefer_intermediate=True 或存在 combine=True 时，读取/生成 intermediate 缓存；
intermediate 缓存以 dataset_name_intermediate 作为前缀独立保存。

`set_data_source_describe()` ¶

设置数据源描述，在父类基础上添加TGGNN相关路径

Returns¶

collections.OrderedDict 包含所有数据路径的有序字典

`hydrodatasource.reader.floodevent.FloodEventDatasource` ¶

Bases: SelfMadeHydroDataset

Flood event dataset processing class

Inherits from SelfMadeHydroDataset, specifically designed for processing individual flood event data, including event extraction functions.

`init(uri=None, time_unit=None, *, rain_key='rain', pet_key='ES', net_rain_key='net_rain', obs_flow_key='inflow', warmup_length=0, **kwargs)` ¶

Initialize the flood event dataset.

Parameters¶

uri : str Absolute path or S3 URI pointing to the data directory. time_unit : list of str, optional List of time units, default is ["3h"]. rain_key : str, optional Key name for rain data, default is "rain". net_rain_key : str, optional Key name for net rain data, default is "net_rain". obs_flow_key : str, optional Key name for observed flow data, default is "inflow". warmup_length : int, optional Number of time steps to include before flood event starts as warmup period, default is 0. **kwargs Additional keyword arguments passed to the parent class.

`adjust_warmup_time_to_augmented_year(warmup_df, augmented_start_time)` ¶

调整预热期数据的年份到增强数据的年份

Parameters¶

warmup_df : pd.DataFrame 预热期数据 augmented_start_time : str 增强数据的起始时间 (YYYYMMDDHH格式)

Returns¶

pd.DataFrame 调整年份后的预热期数据

`check_event_data_nan(all_event_data, exclude_warmup=True)` ¶

Check for NaN values in rainfall and runoff data for all flood events.

This method leverages the class's attributes (net_rain_key, obs_flow_key, warmup_length) and can optionally exclude warmup period from NaN checking.

Parameters¶

all_event_data : list of dict List of event dictionaries, each containing net rainfall, runoff, filepath, etc. exclude_warmup : bool, optional Whether to exclude warmup period data from NaN checking, by default True. When True, only checks data points where flood_event_markers > 0 or excludes the first warmup_length data points if markers are not available.

Raises¶

ValueError If any NaN values are found in the non-warmup data, raises an exception and prints detailed information.

Notes¶

This method uses the class attributes: - self.net_rain_key: Key name for net rainfall data - self.obs_flow_key: Key name for observed flow data - self.warmup_length: Number of warmup time steps

When exclude_warmup=True: 1. If 'flood_event_markers' exists in event data, only checks where markers > 0 2. Otherwise, skips the first self.warmup_length data points

`concatenate_warmup_and_augmented_data(warmup_df, augmented_file_path, use_score_as_weight=False, score_threshold=60)` ¶

拼接预热期数据和增强场次数据

Parameters¶

warmup_df : pd.DataFrame 预热期数据（仅在增强CSV未包含预热期时使用） augmented_file_path : str 增强文件路径

Returns¶

pd.DataFrame 拼接后的完整数据

`create_xarray_dataset_from_augdf(df, station_id, time_unit='3h')` ¶

将时间序列DataFrame转换为xarray Dataset格式

Parameters¶

df : pd.DataFrame 时间序列数据 station_id : str 站点ID time_unit : str, optional 时间单位，默认"3h"

Returns¶

xr.Dataset xarray格式的数据集

`discover_augmented_files(augmented_files_dir, source_event=None, modified_by=None, time_range=None, latest_only=False)` ¶

发现增强数据文件的智能接口

Parameters:

Name	Type	Description	Default
`augmented_files_dir`	`str`	增强数据文件目录	required
`source_event`	`Optional[str]`	源事件名过滤 (如 "event_1994081520_1994081805")	`None`
`modified_by`	`Optional[List[str]]`	修改者列表过滤	`None`
`time_range`	`Optional[Tuple[str, str]]`	修改时间范围过滤 ("2025-01-01", "2025-12-31")	`None`
`latest_only`	`bool`	是否只返回每个源事件的最新版本	`False`

Returns:

Type	Description
`List[Dict]`	List[Dict]: 文件信息列表，包含文件路径、元数据等

`extract_flood_events(df, include_peak_obs=True)` ¶

提取洪水事件并转换为标准格式字典

Parameters¶

df : pd.DataFrame DataFrame containing station data. include_peak_obs : bool, default=True 是否包含洪峰观测值

Returns¶

List[Dict] 标准格式的事件字典列表，与现有算法完全兼容

`get_constants()` ¶

Get the constant values used by this datasource.

Returns¶

dict Dictionary containing constant values with keys: - 'net_rain_key': Key name for net rain data - 'obs_flow_key': Key name for observed flow data - 'delta_t_hours': Time step in hours (derived from time_unit) - 'delta_t_seconds': Time step in seconds - 'warmup_length': Number of warmup time steps

`get_user_contributions_summary(augmented_files_dir)` ¶

获取用户贡献统计

`get_warmup_period_data(original_start_time, original_end_time, station_id, warmup_hours=240)` ¶

获取原始场次前面的预热期数据

Parameters¶

original_start_time : str 原始场次起始时间 (YYYYMMDDHH格式) original_end_time : str 原始场次结束时间 (YYYYMMDDHH格式) station_id : str 站点ID warmup_hours : int, optional 预热期小时数，默认240小时(10天)

Returns¶

Optional[pd.DataFrame] 预热期数据，包含time, rain, gen_discharge, obs_discharge列

`load_1basin_flood_events(station_id=None, flow_unit='mm/3h', include_peak_obs=True, verbose=True, **kwargs)` ¶

加载洪水事件数据

Parameters¶

station_id: 指定站点ID，如果为None则处理所有站点 flow_unit Unit of streamflow, default is "mm/3h". include_peak_obs: 是否包含洪峰观测值 verbose: 是否打印详细信息

Returns¶

1	`List[Dict]: 标准格式的事件字典列表，与现有算法完全兼容`

`parse_augmented_file_metadata(augmented_file_path)` ¶

解析增强文件的元信息

Parameters¶

augmented_file_path : str 增强文件的路径

Returns¶

Dict 包含源场次信息的字典，包括起始时间、结束时间、源文件名等

`process_augmented_files_by_discovery(station_ids, augmented_files_dir, source_event=None, modified_by=None, time_range=None, latest_only=True, warmup_hours=240, time_unit='3h', score_threshold=60, use_score_as_weight=False)` ¶

Process augmented data files based on file discovery.

Parameters¶

station_ids : Union[str, List[str]] Station ID or list of station IDs. augmented_files_dir : str Directory containing augmented data files. source_event : Optional[str], optional Filter by source event name. modified_by : Optional[List[str]], optional Filter by list of modifiers. time_range : Optional[Tuple[str, str]], optional Filter by modification time range. latest_only : bool, optional Whether to process only the latest version for each source event. warmup_hours : int, optional Number of warmup hours. time_unit : str, optional Time unit. score_threshold : int, optional Score threshold (default: 60). Behavior depends on use_score_as_weight: - If use_score_as_weight=False: Only events with Score >= score_threshold are included (filtering mode) - If use_score_as_weight=True: Events with Score < score_threshold get flood_event=1, events with Score >= score_threshold get flood_event=score (weighting mode) use_score_as_weight : bool, optional If True, use score-based weighting mode where flood_event stores actual scores (1-100) instead of binary values (0/1). Default is False (threshold filtering mode). - False (filtering): flood_event=0/1, only score>=threshold included - True (weighting): flood_event=-1/0 (non-flood), 1 (low-quality), or score (high-quality)

Returns¶

Optional[str] Path to the cache file, or None if processing fails.

`read_ts_xrdataset_augmented(gage_id_lst=None, t_range=None, var_lst=None, time_unit='3h', **kwargs)` ¶

读取增强的时间序列数据，优先从dataaugment缓存文件读取

Parameters¶

gage_id_lst : Optional[List[str]], optional 站点ID列表 t_range : Optional[List[str]], optional 时间范围 var_lst : Optional[List[str]], optional 变量列表 time_unit : str, optional 时间单位，默认"3h" **kwargs 其他参数

Returns¶

Dict 包含增强数据的字典，格式与read_ts_xrdataset一致

`rename_dataframe_columns(df, custom_mapping=None)` ¶

重命名数据框的列名，包括默认的映射和自定义映射

Parameters¶

df : pd.DataFrame 需要重命名列的数据框 custom_mapping : dict, optional 自定义的列名映射字典，例如 {'old_name': 'new_name'}

Returns¶

pd.DataFrame 列名重命名后的数据框

`save_augmented_timeseries_to_cache(ds, station_ids, time_unit='3h')` ¶

将增强时间序列数据保存到cache目录

Parameters¶

ds : xr.Dataset 要保存的数据集 station_ids : Union[str, List[str]] 站点ID或站点ID列表 time_unit : str, optional 时间单位，默认"3h"

Returns¶

str 保存的文件路径

`hydrodatasource.reader.gages.Gages` ¶

Bases: HydroData

`cache_attributes_xrdataset()` ¶

Convert all the attributes to a single dataset

Returns¶

None

`cache_timeseries_xrdataset(trange4cache=None, **kwargs)` ¶

Save all timeseries data in separate NetCDF files for each time unit.

Parameters¶

trange4cache : list, optional Time range for caching data, by default ["1980-01-01", "2023-12-31"] kwargs : dict, optional batchsize -- Number of basins to process per batch, by default 100 time_units -- List of time units to process, by default None start0101_freq -- for freq setting, if the start date is 01-01, set True, by default False

`cache_xrdataset(t_range=None, time_units=None)` ¶

Save all data in a netcdf file in the cache directory

`get_constant_cols()` ¶

all readable attrs in GAGES-II

`read_attr_all(gages_ids)` ¶

read all attr data for some sites in GAGES-II TODO: now it is not same as functions in CAMELS where read_attr_all has no "gages_ids" parameter

Parameters¶

gages_ids : Union[list, np.ndarray] gages sites' ids

Returns¶

ndarray all attr data for gages_ids

`read_attr_origin(gages_ids, attr_lst)` ¶

this function read the attrs data in GAGES-II but not transform them to int when they are str

Parameters¶

gages_ids : [type] [description] attr_lst : [type] [description]

Returns¶

np.ndarray the first dim is types of attrs, and the second one is sites

`read_constant_cols(object_ids=None, constant_cols=None, **kwargs)` ¶

read some attrs of some sites

Parameters¶

object_ids : [type], optional sites_ids, by default None constant_cols : list, optional attrs' names, by default None

Returns¶

np.array attr data for object_ids

`read_target_cols(usgs_id_lst=None, t_range_list=None, target_cols=None, **kwargs)` ¶

Read USGS daily average streamflow data according to id and time

Parameters¶

usgs_id_lst site information t_range_list must be time range for downloaded data target_cols

kwargs optional

Returns¶

np.array streamflow data, 1d-axis: gages, 2d-axis: day, 3d-axis: streamflow

`read_ts_xrdataset(gage_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

Read time-series xarray dataset from multiple NetCDF files and organize them by time units.

Parameters:¶

gage_id_lst: list List of gage IDs to select. t_range: list List of two elements [start_time, end_time] to select time range. var_lst: list List of variables to select. **kwargs Additional arguments.

Returns:¶

dict: A dictionary where each key is a time unit and each value is an xarray.Dataset containing the selected gage IDs, time range, and variables.

`read_usgs_gage(usgs_id, t_lst)` ¶

read data for one gage

Parameters¶

usgs_id : [type] [description] t_lst : [type] [description]

Returns¶

[type] [description]

`hydrodatasource.reader.grdc.Grdc` ¶

Bases: HydroData

Reading GRDC streamflow data.

`cache_grdc_daily(station_ids=None, time_range=None, batch_size=1000)` ¶

Save GRDC daily data to a NetCDF file.

Parameters¶

station_ids: list of str List of station IDs to read data for. time_range: list of str List of [start_time, end_time] in UTC and ISO format strings e.g. ['YYYY-MM-DDTHH:MM:SSZ', 'YYYY-MM-DDTHH:MM:SSZ']. batch_size: int Number of stations to process in each batch

`map_station_to_continent(station_id)` ¶

Maps a station ID to its corresponding continent based on rules.

`read_grdc_daily_data(station_id, time_range, parameter='Q', column='streamflow')` ¶

read daily river discharge data from Global Runoff Data Centre (GRDC).

Requires the GRDC daily data files in a local directory. The GRDC daily data files can be ordered at https://www.bafg.de/GRDC/EN/02_srvcs/21_tmsrs/riverdischarge_node.html

Parameters¶

station_id: The station id to get. The station id can be found in the
    catalogues at
    https://www.bafg.de/GRDC/EN/02_srvcs/21_tmsrs/212_prjctlgs/project_catalogue_node.html
time_range: A list of [start_time, end_time] in UTC and ISO format strings e.g.
    ['YYYY-MM-DDTHH:MM:SSZ', 'YYYY-MM-DDTHH:MM:SSZ'].
parameter: optional. The parameter code to get, e.g. ('Q') discharge,
    cubic meters per second.
column: optional. Name of column in dataframe. Default: "streamflow".

Returns:

Type	Description
	grdc data in a dataframe and metadata.

Examples:

.. code-block:: python

from ewatercycle.observation.grdc import get_grdc_data

df, meta = get_grdc_data('6335020',
                        ['2000-01-01T00:00Z', '2001-01-01T00:00Z'])
df.describe()
        streamflow
count   4382.000000
mean    2328.992469
std     1190.181058
min      881.000000
25%     1550.000000
50%     2000.000000
75%     2730.000000
max    11300.000000

meta
{'grdc_file_name': '/home/myusername/git/eWaterCycle/ewatercycle/6335020_Q_Day.Cmd.txt',
'id_from_grdc': 6335020,
'file_generation_date': '2019-03-27',
'river_name': 'RHINE RIVER',
'station_name': 'REES',
'country_code': 'DE',
'grdc_latitude_in_arc_degree': 51.756918,
'grdc_longitude_in_arc_degree': 6.395395,
'grdc_catchment_area_in_km2': 159300.0,
'altitude_masl': 8.0,
'dataSetContent': 'MEAN DAILY DISCHARGE (Q)',
'units': 'm³/s',
'time_series': '1814-11 - 2016-12',
'no_of_years': 203,
'last_update': '2018-05-24',
'nrMeasurements': 'NA',
'UserStartTime': '2000-01-01T00:00Z',
'UserEndTime': '2001-01-01T00:00Z',
'nrMissingData': 0}

`read_site_info()` ¶

Reads the shapefile and extracts the 'id' column as a list.

`read_streamflow_xrdataset(station_id_lst=None, time_range=None, **kwargs)` ¶

Read GRDC daily data from multiple NetCDF files and organize them by station IDs.

Parameters¶

station_id_lst : list List of station IDs to select. time_range : list List of two elements [start_time, end_time] to select time range. **kwargs: Additional arguments.

Returns¶

dict: A dictionary where each key is a station ID and each value is an xarray.Dataset containing the selected station IDs, time range, and variable.

`hydrodatasource.reader.rainfall_reader.RainfallReader` ¶

Bases: HydroData

`init(uri)` ¶

A reader for basin rainfall data.

Parameters¶

uri : str Absolute path pointing to the rainfall data directory.

`read_1basin_rainfall(basin_id)` ¶

Read basin-mean rainfall data for a single basin.

Parameters¶

basin_id : str Basin ID

Returns¶

pd.DataFrame A DataFrame containing rainfall data for the specified basin.

`read_basin_1station_rainfall(basin_id, station_id)` ¶

Read rainfall data for a single station.

Parameters¶

station_id : str Station ID

Returns¶

pd.DataFrame A DataFrame containing rainfall data for the specified station.

`read_basin_rainfall(basin_ids)` ¶

Read rainfall data for all basins.

Parameters¶

basin_ids : list List of basin IDs

Returns¶

pd.DataFrame A DataFrame containing rainfall data for all basins.

`hydrodatasource.reader.reservoir_datasets.Crd` ¶

Bases: HydroData

Reading CRD reservoir dataset

`hydrodatasource.reader.rsvr_inflow_reader.RsvrInflowReader` ¶

Bases: HydroData

`read_1rsvr_inflow(rsvr_id)` ¶

Read inflow data for a single reservoir.

Parameters¶

rsvr_id : str Reservoir ID

Returns¶

pd.DataFrame A DataFrame containing inflow data for the specified reservoir.

Processor¶

Basin Mean Rainfall¶

`hydrodatasource.processor.basin_mean_rainfall.basin_mean_func(df, weights_dict=None)` ¶

Generic basin averaging method that supports both arithmetic mean and weighted mean (e.g. Thiessen polygon weights)

When some columns have missing values in a row, the function automatically switches to arithmetic mean for that row instead of using weights. This ensures robustness when dealing with incomplete data.

Parameters¶

df : DataFrame Time series DataFrame for multiple stations, with station names as column names; each column should be a time series of rainfall data for a specific station weights_dict : dict, optional Dictionary with tuple of station names as keys and list of weights as values. If None, arithmetic mean is used.

the keys of list must be in the same order as the columns of df.

hence, an easy way is you give your df with a sorted column names and then use the same order to create the keys of weights_dict. for example: weights_dict = { ("st1", "st2", "st3", "st4"): [0.25, 0.5, 0.1, 0.15], } df = df[["st1", "st2", "st3", "st4"]] then the keys of weights_dict must be in the same order as the columns of df.

NOTE

we set the format of weights_dict like this because we want to extend it to match the weights_dict key based on the missing data situation and the key in weights_dict. This is a TODO item. if the key in weights_dict matches the columns of df, we use the weights in weights_dict; if the key in weights_dict does not match the columns of df, we use the arithmetic mean. For example, if the columns of df are ["st1", "st2", "st3", "st4"], and the weights_dict is: weights_dict = { ("st1", "st2", "st3", "st4"): [0.25, 0.5, 0.1, 0.15], ("st1", "st2", "st3"): [0.25, 0.5, 0.1], ("st3", "st4"): [0.1, 0.15], } then when st4 has missing data, we use the weights in ("st1", "st2", "st3") to calculate the weighted mean; and when st1 and st2 have missing data, we use the weights in ("st3", "st4") to calculate the weighted mean. Otherwise, we use the arithmetic mean.

But this function is not finished yet, and the weights_dict now only supports the case that the keys of weights_dict has all the columns of df; if any column in df is missing, the function will use the arithmetic mean.

Returns¶

Series Basin-averaged time series

Rainfall-Runoff Event Identification¶

`hydrodatasource.processor.dmca_esr.get_rr_events(rain, flow, basin_area, max_window=100, max_flow_min=None)` ¶

use DMCA-ESR method to identify rainfall-runoff events

Parameters¶

rain : xr.DataArray the rainfall data flow : xr.DataArray the streamflow data basin_area : xr.Dataset a dataset with a variable named area and for each basin max_window: int number of time intervals for find events; default 100 (for hourly) max_flow_min: list the minimum of max flow for each basin value below this will not be considered to look for an event; default 100 m^3/s Returns

dict the rainfall-runoff events for each basin

Raises¶

ValueError Invalid unit format ValueError Unsupported unit

Cleaner¶

Cleaner (base class)¶

`hydrodatasource.cleaner.cleaner.Cleaner` ¶

RainfallCleaner¶

`hydrodatasource.cleaner.rainfall_cleaner.RainfallCleaner` ¶

Bases: Cleaner

`init(data_folder, output_folder)` ¶

All files to be cleaned are in the data_dir

Parameters¶

data_dir : type description

`data_check_hourly_extreme(basin_id, climate_extreme_value=None, modify=False)` ¶

Check if the daily precipitation values at chosen stations are within a reasonable range. Values larger than the climate extreme value are treated as anomalies. If no climate_extreme_value is provided, the maximum value in the data is used.

Parameters¶

climate_extreme_value : float, optional Climate extreme threshold for the region, calculated as 95% of the maximum observed DRP. If not provided, will be calculated as 95% of the maximum DRP value in the data.

Returns¶

df_anomaly_stations_periods : pd.DataFrame DataFrame of anomalies with columns: 'STCD', 'TM', 'DRP'.

`data_check_time_series(basin_id, check_type=None, gradient_limit=None, window_size=None, consistent_value=None, modify=False)` ¶

Check daily precipitation values at chosen stations for gradient or time consistency anomalies.

Parameters¶

basin_id: str Basin ID. check_type : str Type of check to perform: "gradient" for gradient check, "consistency" for time consistency check. gradient_limit : float, optional Maximum allowable gradient change in precipitation between consecutive days. Used in "gradient" check. Default is 10 mm. window_size : int, optional Size of the window (in hours) to check for time consistency (used in "consistency" check). Default is 24 hours. consistent_value : float, optional The specific precipitation value to check for consistency (used in "consistency" check). Default is 0.1 mm.

Returns¶

pd.DataFrame DataFrame of detected anomalies with columns: 'STCD', 'TM', 'DRP', 'Issue' (where applicable).

`data_check_yearly(basin_id, year_range=None, diff_range=None, min_true_percentage=0.75, min_consecutive_years=3, modify=False)` ¶

计算遥感数据与站点数据之间的降水差异，评估站点可靠性，并返回可信任的站点列表。

参数:¶

basin_id : str Basin ID year_range : list, 可选要筛选的年份范围，默认是 [2010, 2024]。 diff_range : list, 可选站点数据和遥感数据之间的ratio差异范围 0.5 means station data is 0.5 times of reanalysis data 2.0 means station data is 2 times of reanalysis data min_true_percentage : float, 可选要求可信年份的最小比例，默认 0.75。 min_consecutive_years : int, 可选最小连续可信年份数，默认 3。

返回:¶

result_df : pd.DataFrame 可信站点的 DataFrame，包含 'STCD'、'Latitude'、'Longitude' 和 'Reason' 列。

`rainfall_clean(basin_id, **kwargs)` ¶

the station gauged rainfall data cleaning pipeline

`read_and_concat_csv(basin_id)` ¶

读取并合并文件夹下的所有 CSV 文件

ReservoirInflowBacktrack¶

`hydrodatasource.cleaner.rsvr_inflow_cleaner.ReservoirInflowBacktrack` ¶

Bases: Cleaner

`init(data_folder, output_folder)` ¶

Back-calculating inflow of reservior

Parameters¶

data_folder : str the folder of reservoir data output_folder : type where we put inflow data

`back_calculation(rsvr_id, clean_w_path, original_file, output_folder)` ¶

Back-calculate inflow from reservoir storage data NOTE: each time has three columns: I Q W -- I is the inflow, Q is the outflow, W is the reservoir storage Generally, in sql database, a time means the end of previous time period For example, a hourly database, 13:00 means 12:00-13:00 period because the data is GOT at 13:00 (we cannot observe future) Hence, for this function, W means the storage at the end of the time period, I and Q means the inflow and outflow of the time period So we need to use W of the previous time as the initial water storage of the time period. Hence, I1 = Q1 + (W1 - W0)

Parameters¶

rsvr_id : str The ID of the reservoir data_path : str the path to the cleaned_w_data file original_file: str the path to the original file output_folder : str where to save the back calculated data

Returns¶

str the path to the result file

`clean_w(rsvr_id, file_path, output_folder, fit_method='quadratic', zw_curve_std_times=3.0, remove_zw_outliers=False)` ¶

Remove abnormal reservoir capacity data

Parameters¶

rsvr_id : str The ID of the reservoir file_path : str Path to the input file output_folder : str Path to the output folder fit_method : str, optional z-w curve fitting method, by default "quadratic" TODO: MORE METHODS need to be supported; power is also need to be debugged zw_curve_std_times: float, optional the times of standard deviation to remove outliers, by default 3 remove_zw_outliers: bool, optional whether to remove outliers for z-w curve fitting, by default False

Returns¶

str Path to the cleaned data file

`delete_negative_inq(rsvr_id, inflow_data_path, original_file, output_folder, negative_deal_window=7, negative_deal_stride=4)` ¶

remove negative inflow values with a rolling window the negative value will be adjusted to positvie ones to make the total inflow consistent for example, 1, -1, 1, -1 will be adjusted to 0, 0, 0, 0 so that wate balance is kept but note that as the window has stride, maybe the final few values will not be adjusted

Parameters¶

rsvr_id : str the id of the reservoir inflow_data_path : str the data file after back_calculation original_file : str the original file output_folder : str where to save the data negative_deal_window : int, optional the window to deal with negative values, by default 7 negative_deal_stride : int, optional the stride of window, by default 4

Returns¶

str the path to the result file

`insert_inq(rsvr_id, inflow_data_path, original_file, output_folder)` ¶

make inflow data as hourly data as original data is not strictly hourly data and insert inq with linear interpolation

Parameters¶

rsvr_id : str the id of the reservoir inflow_data_path : str the data file after delete negative inflow values original_file : str the original file output_folder : str where to save the data

Returns¶

str the path to the result file

`rsvr_inflow_clean(**kwargs)` ¶

The reservoir inflow data cleaning pipeline

Parameters¶

zw_curve_std_times : float the times of standard deviation to remove outliers, by default 3.0 remove_zw_outliers : bool whether to remove outliers for z-w curve fitting, by default False

StreamflowCleaner¶

`hydrodatasource.cleaner.streamflow_cleaner.StreamflowCleaner` ¶

Bases: Cleaner

`FFT(streamflow_data)` ¶

对流量数据进行迭代的傅里叶滤波处理，包括非负值调整和流量总量调整。 :cutoff_frequency: 傅里叶滤波的截止频率。 :time_step: 数据采样间隔。 :iterations: 迭代次数。

`data_balanced(origin_data, transform_data)` ¶

对一维流量数据进行总量平衡变换。 :origin_data: 原始一维流量数据。 :transform_data: 平滑转换后的一维流量数据。

`kalman_filter(streamflow_data)` ¶

对流量数据应用卡尔曼滤波进行平滑处理，并保持流量总量平衡。 :param streamflow_data: 原始流量数据

`lowpass_filter(streamflow_data)` ¶

对一维流量数据应用调整后的低通滤波器。 :cutoff_frequency: 低通滤波器的截止频率。 :sampling_rate: 数据的采样率。 :order: 滤波器的阶数，默认为5。

`moving_average(streamflow_data)` ¶

对流量数据应用滑动平均进行平滑处理，并保持流量总量平衡。 :param streamflow_data: 输入的流量数据数组 :return: 平滑处理后的流量数据

`moving_average_difference(streamflow_data)` ¶

对流量数据应用滑动平均差算法进行平滑处理，并保持流量总量平衡。 :window_size: 滑动窗口的大小

`robust_fitting(streamflow_data, k=1.5)` ¶

对流量数据应用抗差修正算法进行平滑处理，并保持流量总量平衡。默认采用二次曲线进行拟合优化，该算法处理性能较差

`wavelet(streamflow_data)` ¶

对一维流量数据进行小波变换分析前后拓展数据以减少边缘失真，然后调整总流量。 :cwt_row: 小波变换中使用的特定宽度。

WaterlevelCleaner¶

`hydrodatasource.cleaner.waterlevel_cleaner.WaterlevelCleaner` ¶

Bases: Cleaner

API Reference¶

Data Resolver (Unified Data Interface)¶

hydrodatasource.configs.data_resolver.open_dataset(dataset_id, *, source=None, ctx=None, **reader_kwargs) ¶

Parameters¶

Returns¶

Raises¶

Examples¶

hydrodatasource.configs.data_resolver.resolve_data_path(dataset_id, *, source=None, ctx=None) ¶

Parameters¶

Returns¶

Raises¶

Examples¶

Default: zero-boilerplate, reads ~/hydro_setting.yml¶

Custom storage root via ResolverContext¶

Reader¶

HydroData (base class)¶

hydrodatasource.reader.data_source.HydroData ¶

Parameters¶

SelfMadeHydroDataset¶

hydrodatasource.reader.data_source.SelfMadeHydroDataset ¶

__init__(uri=None, time_unit=None, **kwargs) ¶

Parameters¶

cache_attributes_xrdataset() ¶

Returns¶

cache_timeseries_xrdataset(**kwargs) ¶

Parameters¶

cache_xrdataset(t_range=None, time_units=None) ¶

get_attributes_cols() ¶

get_timeseries_cols() ¶

read_area(gage_id_lst=None) ¶

read_attributes(object_ids=None, constant_cols=None, **kwargs) ¶

read_mean_prcp(gage_id_lst=None, unit='mm/d') ¶

Parameters¶

Returns¶

read_timeseries(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs) ¶

Parameters¶

Returns¶

read_ts_xrdataset(gage_id_lst=None, t_range=None, var_lst=None, **kwargs) ¶

Parameters:¶

Returns:¶

Other Readers¶

hydrodatasource.reader.data_source.LongTermDataset ¶

cache_global_dataset(object_ids=None, t_range_list=None) ¶

hydrodatasource.reader.data_source.SelfMadeForecastDataset ¶

__init__(uri=None, time_unit=None, **kwargs) ¶

Parameters¶

cache_forecast_xrdataset(t_range=None, **kwargs) ¶

Parameters¶

read_forecast(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs) ¶

Parameters¶

read_forecast_xrdataset(gage_id_lst, t_range, var_lst, **kwargs) ¶

Parameters¶

Returns¶

read_ts_xrdataset(gage_id_lst, t_range, var_lst, **kwargs) ¶

Parameters¶

Returns¶

set_data_source_describe() ¶

Returns¶

hydrodatasource.reader.data_source.StationHydroDataset ¶

__init__(uri=None, time_unit=None, **kwargs) ¶

Parameters¶

cache_adjacency_xrdataset() ¶

cache_all_station_data(**kwargs) ¶

cache_station_info_xrdataset() ¶

cache_station_timeseries_xrdataset(**kwargs) ¶

Parameters¶

get_stations_by_basin(basin_id) ¶

Parameters¶

Returns¶

read_adjacency_xrdataset(basin_id) ¶

Parameters¶

Returns¶

read_basin_adjacency(basin_id) ¶

Parameters¶

Returns¶

read_basin_stations(basin_id) ¶

Parameters¶

Returns¶

read_station_info() ¶

read_station_info_xrdataset(**kwargs) ¶

`hydrodatasource.configs.data_resolver.open_dataset(dataset_id, *, source=None, ctx=None, **reader_kwargs)` ¶

`hydrodatasource.configs.data_resolver.resolve_data_path(dataset_id, *, source=None, ctx=None)` ¶

`hydrodatasource.reader.data_source.HydroData` ¶

`hydrodatasource.reader.data_source.SelfMadeHydroDataset` ¶

`init(uri=None, time_unit=None, **kwargs)` ¶

`cache_attributes_xrdataset()` ¶

`cache_timeseries_xrdataset(**kwargs)` ¶

`cache_xrdataset(t_range=None, time_units=None)` ¶

`get_attributes_cols()` ¶

`get_timeseries_cols()` ¶

`read_area(gage_id_lst=None)` ¶

`read_attributes(object_ids=None, constant_cols=None, **kwargs)` ¶

`read_mean_prcp(gage_id_lst=None, unit='mm/d')` ¶

`read_timeseries(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

`read_ts_xrdataset(gage_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

`hydrodatasource.reader.data_source.LongTermDataset` ¶

`cache_global_dataset(object_ids=None, t_range_list=None)` ¶

`hydrodatasource.reader.data_source.SelfMadeForecastDataset` ¶

`init(uri=None, time_unit=None, **kwargs)` ¶

`cache_forecast_xrdataset(t_range=None, **kwargs)` ¶

`read_forecast(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

`read_forecast_xrdataset(gage_id_lst, t_range, var_lst, **kwargs)` ¶

`read_ts_xrdataset(gage_id_lst, t_range, var_lst, **kwargs)` ¶

`set_data_source_describe()` ¶

`hydrodatasource.reader.data_source.StationHydroDataset` ¶

`init(uri=None, time_unit=None, **kwargs)` ¶

`cache_adjacency_xrdataset()` ¶

`cache_all_station_data(**kwargs)` ¶

`cache_station_info_xrdataset()` ¶

`cache_station_timeseries_xrdataset(**kwargs)` ¶

`get_stations_by_basin(basin_id)` ¶

`read_adjacency_xrdataset(basin_id)` ¶

`read_basin_adjacency(basin_id)` ¶

`read_basin_stations(basin_id)` ¶

`read_station_info()` ¶

`read_station_info_xrdataset(**kwargs)` ¶

`read_station_object_ids()` ¶

`read_station_timeseries(station_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

`read_station_ts_xrdataset(station_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

`set_data_source_describe()` ¶

`hydrodatasource.reader.data_source.TgHydroDatasource` ¶

`init(uri=None, time_unit=None, **kwargs)` ¶

`cache_intermediate_attributes_xrdataset()` ¶

`cache_intermediate_timeseries_xrdataset(**kwargs)` ¶

`classify_nodes(basin_names=None)` ¶

`get_name()` ¶

`read_attr_xrdataset(gage_id_lst=None, var_lst=None, **kwargs)` ¶

`read_graph_data(object_ids=None, as_numpy=False, include_optional=True)` ¶

`read_intermediate_timeseries(object_ids=None, t_range_list=None, relevant_cols=None, **kwargs)` ¶

`read_ts_xrdataset(gage_id_lst=None, t_range=None, var_lst=None, **kwargs)` ¶

`set_data_source_describe()` ¶

`hydrodatasource.reader.floodevent.FloodEventDatasource` ¶

`init(uri=None, time_unit=None, *, rain_key='rain', pet_key='ES', net_rain_key='net_rain', obs_flow_key='inflow', warmup_length=0, **kwargs)` ¶

`adjust_warmup_time_to_augmented_year(warmup_df, augmented_start_time)` ¶

`check_event_data_nan(all_event_data, exclude_warmup=True)` ¶

`concatenate_warmup_and_augmented_data(warmup_df, augmented_file_path, use_score_as_weight=False, score_threshold=60)` ¶

`create_xarray_dataset_from_augdf(df, station_id, time_unit='3h')` ¶

`discover_augmented_files(augmented_files_dir, source_event=None, modified_by=None, time_range=None, latest_only=False)` ¶

`extract_flood_events(df, include_peak_obs=True)` ¶

`get_constants()` ¶

`get_user_contributions_summary(augmented_files_dir)` ¶

`get_warmup_period_data(original_start_time, original_end_time, station_id, warmup_hours=240)` ¶

`load_1basin_flood_events(station_id=None, flow_unit='mm/3h', include_peak_obs=True, verbose=True, **kwargs)` ¶

`parse_augmented_file_metadata(augmented_file_path)` ¶

`process_augmented_files_by_discovery(station_ids, augmented_files_dir, source_event=None, modified_by=None, time_range=None, latest_only=True, warmup_hours=240, time_unit='3h', score_threshold=60, use_score_as_weight=False)` ¶

`read_ts_xrdataset_augmented(gage_id_lst=None, t_range=None, var_lst=None, time_unit='3h', **kwargs)` ¶

`rename_dataframe_columns(df, custom_mapping=None)` ¶

`save_augmented_timeseries_to_cache(ds, station_ids, time_unit='3h')` ¶