Merge branch 'main' into metrics/from_fork

CONTRIBUTING.md

@@ -11,7 +11,20 @@ Please come join us on our [Slack](https://join.slack.com/t/neuralprophet/shared
 
 ## Process
 If this is your first time contributing to NeuralProphet, please read our [wiki summary](https://github.com/ourownstory/neural_prophet/wiki#contributing-process) of the steps involved.
-## Dev Install
+
+## Dev Install using Poetry
+Prerequisite:  [install poetry](https://python-poetry.org/docs/#installing-with-the-official-installer).
+* Start by cloning the repo with `git clone <copied link from github>`.
+* Make sure you have changed directories to your cloned neuralprophet github `cd neural_prophet`. There, run `poetry shell` to (create and) start a (new) poetry virtual environment. If you run `poetry env info --path` you should see the path to the venv.
+* To complete the venv setup, install neuralprophet (in editable mode by default) with `poetry install`.
+Note: poetry will automatically use the specific dependencies in the `poetry.lock` file for reproducibility. If you want to install the latest dependencies instead, run `poetry update`. This will update all dependencies and update the `poetry.lock` file. Be mindful to not track the `poetry.lock` file with git when commiting, unless the purpose of your pull request is to update it.
+
+Warning, you are still lacking some git hooks to auto-format your code pre-commit and to run pytests pre-push.
+Currently these need to be self-added. Simplified instructions to follow.
+
+[Tutorial Link](https://realpython.com/dependency-management-python-poetry/)
+
+## Dev Install (old)
 Before starting it's a good idea to first create and activate a new virtual environment:
 ```
 python3 -m venv <path-to-new-env>

README.md

@@ -82,7 +82,7 @@ pip install neuralprophet[live]
 ```
 This will allow you to enable `plot_live_loss` in the `fit` function to get a live plot of train (and validation) loss.
 
-If you would like the most up to date version, you can instead install direclty from github:
+If you would like the most up to date version, you can instead install directly from github:
 ```shell
 git clone <copied link from github>
 cd neural_prophet

neuralprophet/df_utils.py

@@ -1059,6 +1059,46 @@ def add_missing_dates_nan(df, freq):
     return df_resampled, num_added
 
 
+def create_dummy_datestamps(
+    df, freq="S", startyear=1970, startmonth=1, startday=1, starthour=0, startminute=0, startsecond=0
+):
+    """
+    Helper function to create a dummy series of datestamps for equidistant data without ds.
+    Parameters
+    ----------
+        df : pd.DataFrame
+            dataframe with column 'y' and without column 'ds'
+        freq : str
+            Frequency of data recording, any valid frequency for pd.date_range, such as ``D`` or ``M``
+        startyear, startmonth, startday, starthour, startminute, startsecond : int
+            Defines the first datestamp
+    Returns
+    -------
+        pd.DataFrame
+            dataframe with dummy equidistant datestamps
+
+    Examples
+    --------
+    Adding dummy datestamps to a dataframe without datestamps.
+    To prepare the dataframe for training, import df_utils and insert your prefered dates.
+        >>> from neuralprophet import df_utils
+        >>> df_drop = df.drop("ds", axis=1)
+        >>> df_dummy = df_utils.create_dummy_datestamps(
+        >>> df_drop, freq="S", startyear=1970, startmonth=1, startday=1, starthour=0, startminute=0, startsecond=0
+        >>> )
+    """
+    if "ds" in df:
+        raise ValueError("Column 'ds' in df detected.")
+    log.info(f"Dummy equidistant datestamps added. Frequency={freq}.")
+    df_length = len(df)
+    startdate = pd.Timestamp(
+        year=startyear, month=startmonth, day=startday, hour=starthour, minute=startminute, second=startsecond
+    )
+    datestamps = pd.date_range(startdate, periods=df_length, freq=freq)
+    df_dummy = pd.DataFrame({"ds": datestamps, "y": df["y"]})
+    return df_dummy
+
+
 def fill_linear_then_rolling_avg(series, limit_linear, rolling):
     """Adds missing dates, fills missing values with linear imputation or trend.
 

neuralprophet/forecaster.py

@@ -693,7 +693,7 @@ def add_country_holidays(
         and create the corresponding configs such as lower, upper windows and the regularization
         parameters
 
-        Holidays can only be added for a single country. Calling the function
+        Holidays can only be added for a single country or country list. Calling the function
         multiple times will override already added country holidays.
 
         Parameters

neuralprophet/plot_utils.py

@@ -193,7 +193,7 @@ def check_if_configured(m, components, error_flag=False):  # move to utils
     if "lagged_regressors" in components and m.config_lagged_regressors is None:
         components.remove("lagged_regressors")
         invalid_components.append("lagged_regressors")
-    if "events" in components and (m.config_events and m.config_country_holidays) is None:
+    if "events" in components and (m.config_events is None and m.config_country_holidays is None):
         components.remove("events")
         invalid_components.append("events")
     if "future_regressors" in components and m.config_regressors is None:
@@ -419,31 +419,37 @@ def get_valid_configuration(  # move to utils
     if "events" in components:
         additive_events_flag = False
         muliplicative_events_flag = False
-        for event, configs in m.config_events.items():
-            if validator == "plot_components" and configs.mode == "additive":
-                additive_events_flag = True
-            elif validator == "plot_components" and configs.mode == "multiplicative":
-                muliplicative_events_flag = True
-            elif validator == "plot_parameters":
-                event_params = m.model.get_event_weights(event)
-                weight_list = [(key, param.detach().numpy()[quantile_index, :]) for key, param in event_params.items()]
-                if configs.mode == "additive":
-                    additive_events = additive_events + weight_list
-                elif configs.mode == "multiplicative":
-                    multiplicative_events = multiplicative_events + weight_list
-
-        for country_holiday in m.config_country_holidays.holiday_names:
-            if validator == "plot_components" and m.config_country_holidays.mode == "additive":
-                additive_events_flag = True
-            elif validator == "plot_components" and m.config_country_holidays.mode == "multiplicative":
-                muliplicative_events_flag = True
-            elif validator == "plot_parameters":
-                event_params = m.model.get_event_weights(country_holiday)
-                weight_list = [(key, param.detach().numpy()[quantile_index, :]) for key, param in event_params.items()]
-                if m.config_country_holidays.mode == "additive":
-                    additive_events = additive_events + weight_list
-                elif m.config_country_holidays.mode == "multiplicative":
-                    multiplicative_events = multiplicative_events + weight_list
+        if m.config_events is not None:
+            for event, configs in m.config_events.items():
+                if validator == "plot_components" and configs.mode == "additive":
+                    additive_events_flag = True
+                elif validator == "plot_components" and configs.mode == "multiplicative":
+                    muliplicative_events_flag = True
+                elif validator == "plot_parameters":
+                    event_params = m.model.get_event_weights(event)
+                    weight_list = [
+                        (key, param.detach().numpy()[quantile_index, :]) for key, param in event_params.items()
+                    ]
+                    if configs.mode == "additive":
+                        additive_events = additive_events + weight_list
+                    elif configs.mode == "multiplicative":
+                        multiplicative_events = multiplicative_events + weight_list
+
+        if m.config_country_holidays is not None:
+            for country_holiday in m.config_country_holidays.holiday_names:
+                if validator == "plot_components" and m.config_country_holidays.mode == "additive":
+                    additive_events_flag = True
+                elif validator == "plot_components" and m.config_country_holidays.mode == "multiplicative":
+                    muliplicative_events_flag = True
+                elif validator == "plot_parameters":
+                    event_params = m.model.get_event_weights(country_holiday)
+                    weight_list = [
+                        (key, param.detach().numpy()[quantile_index, :]) for key, param in event_params.items()
+                    ]
+                    if m.config_country_holidays.mode == "additive":
+                        additive_events = additive_events + weight_list
+                    elif m.config_country_holidays.mode == "multiplicative":
+                        multiplicative_events = multiplicative_events + weight_list
 
         if additive_events_flag:
             plot_components.append(

neuralprophet/utils.py

@@ -338,12 +338,13 @@ def get_holidays_from_country(country: Union[str, Iterable[str]], df=None):
     if isinstance(country, str):
         country = [country]
 
-    holidays = {}
+    unique_holidays = {}
     for single_country in country:
         holidays_country = get_country_holidays(single_country, years)
-        # only add holiday if it is not already in the dict
-        holidays.update(holidays_country)
-    holiday_names = holidays.values()
+        for date, name in holidays_country.items():
+            if date not in unique_holidays:
+                unique_holidays[date] = name
+    holiday_names = unique_holidays.values()
     return set(holiday_names)