pydartdiags 0.0.42__tar.gz → 0.0.43__tar.gz

pydartdiags-0.0.43/PKG-INFO

@@ -0,0 +1,45 @@
+Metadata-Version: 2.1
+Name: pydartdiags
+Version: 0.0.43
+Summary: Observation Sequence Diagnostics for DART
+Home-page: https://github.com/NCAR/pyDARTdiags.git
+Author: Helen Kershaw
+Author-email: Helen Kershaw <hkershaw@ucar.edu>
+Project-URL: Homepage, https://github.com/NCAR/pyDARTdiags.git
+Project-URL: Issues, https://github.com/NCAR/pyDARTdiags/issues
+Project-URL: Documentation, https://ncar.github.io/pyDARTdiags
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pandas>=2.2.0
+Requires-Dist: numpy>=1.26
+Requires-Dist: plotly>=5.22.0
+Requires-Dist: pyyaml>=6.0.2
+
+[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
+[![codecov](https://codecov.io/gh/NCAR/pyDARTdiags/graph/badge.svg?token=VK55SQZSVD)](https://codecov.io/gh/NCAR/pyDARTdiags)
+[![PyPI version](https://badge.fury.io/py/pydartdiags.svg)](https://pypi.org/project/pydartdiags/)
+
+
+# pyDARTdiags
+
+pyDARTdiags is a Python library for obsevation space diagnostics for the Data Assimilation Research Testbed ([DART](https://github.com/NCAR/DART)).
+
+pyDARTdiags is under initial development, so please use caution.
+The MATLAB [observation space diagnostics](https://docs.dart.ucar.edu/en/latest/guide/matlab-observation-space.html) are available through [DART](https://github.com/NCAR/DART).
+
+
+pyDARTdiags can be installed through pip: https://pypi.org/project/pydartdiags/  
+Documenation : https://ncar.github.io/pyDARTdiags/
+
+## Contributing
+Contributions are welcome! If you have a feature request, bug report, or a suggestion, please open an issue on our GitHub repository.
+Please read our [Contributors Guide](https://github.com/NCAR/pyDARTdiags/blob/main/CONTRIBUTING.md) if you would like to contribute to
+pyDARTdiags.
+
+## License
+
+pyDARTdiags is released under the Apache License 2.0. For more details, see the LICENSE file in the root directory of this source tree or visit [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0).

pydartdiags-0.0.43/README.md

@@ -0,0 +1,24 @@
+[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
+[![codecov](https://codecov.io/gh/NCAR/pyDARTdiags/graph/badge.svg?token=VK55SQZSVD)](https://codecov.io/gh/NCAR/pyDARTdiags)
+[![PyPI version](https://badge.fury.io/py/pydartdiags.svg)](https://pypi.org/project/pydartdiags/)
+
+
+# pyDARTdiags
+
+pyDARTdiags is a Python library for obsevation space diagnostics for the Data Assimilation Research Testbed ([DART](https://github.com/NCAR/DART)).
+
+pyDARTdiags is under initial development, so please use caution.
+The MATLAB [observation space diagnostics](https://docs.dart.ucar.edu/en/latest/guide/matlab-observation-space.html) are available through [DART](https://github.com/NCAR/DART).
+
+
+pyDARTdiags can be installed through pip: https://pypi.org/project/pydartdiags/  
+Documenation : https://ncar.github.io/pyDARTdiags/
+
+## Contributing
+Contributions are welcome! If you have a feature request, bug report, or a suggestion, please open an issue on our GitHub repository.
+Please read our [Contributors Guide](https://github.com/NCAR/pyDARTdiags/blob/main/CONTRIBUTING.md) if you would like to contribute to
+pyDARTdiags.
+
+## License
+
+pyDARTdiags is released under the Apache License 2.0. For more details, see the LICENSE file in the root directory of this source tree or visit [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0).

{pydartdiags-0.0.42 → pydartdiags-0.0.43}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pydartdiags"
-version = "0.0.42"
+version = "0.0.43"
 authors = [
   { name="Helen Kershaw", email="hkershaw@ucar.edu" },
 ]

{pydartdiags-0.0.42 → pydartdiags-0.0.43}/src/pydartdiags/obs_sequence/obs_sequence.py

@@ -5,6 +5,23 @@ import os
 import yaml
 import struct
 
+def requires_assimilation_info(func):
+    def wrapper(self, *args, **kwargs):
+        if self.has_assimilation_info:
+            return func(self, *args, **kwargs)
+        else:
+            raise ValueError("Assimilation information is required to call this function.")
+    return wrapper
+
+def requires_posterior_info(func):
+    def wrapper(self, *args, **kwargs):
+        if self.has_posterior_info:
+            return func(self, *args, **kwargs)
+        else:
+            raise ValueError("Posterior information is required to call this function.")
+    return wrapper
+
+
 class obs_sequence:
     """Create an obs_sequence object from an ascii observation sequence file.
 
@@ -59,6 +76,8 @@ class obs_sequence:
     
     def __init__(self, file, synonyms=None):
         self.loc_mod = 'None'
+        self.has_assimilation_info = False
+        self.has_posterior = False
         self.file = file
         self.synonyms_for_obs = ['NCEP BUFR observation',
                                  'AIRS observation', 
@@ -72,6 +91,17 @@ class obs_sequence:
             else:
                 self.synonyms_for_obs.append(synonyms)
 
+        if file is None:
+            # Early exit for testing purposes
+            self.df = pd.DataFrame()
+            self.types = {}
+            self.reverse_types = {}
+            self.copie_names = []
+            self.n_copies = 0
+            self.seq = []
+            self.all_obs = []
+            return
+
         module_dir = os.path.dirname(__file__)
         self.default_composite_types = os.path.join(module_dir,"composite_types.yaml")
 
@@ -103,11 +133,16 @@ class obs_sequence:
         self.synonyms_for_obs = [synonym.replace(' ', '_') for synonym in self.synonyms_for_obs]
         rename_dict = {old: 'observation' for old in self.synonyms_for_obs  if old in self.df.columns}
         self.df = self.df.rename(columns=rename_dict)
+
         # calculate bias and sq_err is the obs_seq is an obs_seq.final
         if 'prior_ensemble_mean'.casefold() in map(str.casefold, self.columns):
-            self.df['bias'] = (self.df['prior_ensemble_mean'] - self.df['observation'])
-            self.df['sq_err'] = self.df['bias']**2  # squared error
-	
+            self.has_assimilation_info = True
+            self.df['prior_bias'] = (self.df['prior_ensemble_mean'] - self.df['observation'])
+            self.df['prior_sq_err'] = self.df['prior_bias']**2  # squared error
+        if 'posterior_ensemble_mean'.casefold() in map(str.casefold, self.columns):
+            self.has_posterior_info = True
+            self.df['posterior_bias'] = (self.df['posterior_ensemble_mean'] - self.df['observation'])
+            self.df['posterior_sq_err'] = self.df['posterior_bias']**2
 
     def create_all_obs(self):
         """ steps through the generator to create a
@@ -152,14 +187,38 @@ class obs_sequence:
             data.append(self.types[type_value]) # observation type
             
         # any observation specific obs def info is between here and the end of the list
+        # can be obs_def & external forward operator
+        metadata = obs[typeI+2:-2]
+        obs_def_metadata, external_metadata = self.split_metadata(metadata)
+        data.append(obs_def_metadata)
+        data.append(external_metadata)
+
         time = obs[-2].split()
         data.append(int(time[0])) # seconds
         data.append(int(time[1])) # days
         data.append(convert_dart_time(int(time[0]), int(time[1]))) # datetime   # HK todo what is approprate for 1d models?
         data.append(float(obs[-1])) # obs error variance ?convert to sd?
-        
+
         return data
 
+    @staticmethod
+    def split_metadata(metadata):
+        """
+        Split the metadata list at the first occurrence of an element starting with 'externalF0'.
+
+        Args:
+            metadata (list of str): The metadata list to be split.
+
+        Returns:
+            tuple: Two sublists, the first containing elements before 'externalF0', and the second
+                containing 'externalF0' and all elements after it. If 'externalF0' is not found,
+                the first sublist contains the entire metadata list, and the second is empty.
+        """
+        for i, item in enumerate(metadata):
+            if item.startswith('external_FO'):
+                return metadata[:i], metadata[i:]
+        return metadata, []
+
     def list_to_obs(self, data):
         obs = []
         obs.append('OBS        ' + str(data[0]))  # obs_num lots of space
@@ -171,10 +230,16 @@ class obs_sequence:
             obs.append('   '.join(map(str, data[self.n_copies+2:self.n_copies+5])) + '   ' + str(self.reversed_vert[data[self.n_copies+5]]) )  # location x, y, z, vert
             obs.append('kind') # this is type of observation
             obs.append(self.reverse_types[data[self.n_copies + 6]])  # observation type
+            # Convert metadata to a string and append
+            obs.extend(data[self.n_copies + 7])  # metadata
         elif self.loc_mod == 'loc1d':
             obs.append(data[self.n_copies+2])  # 1d location
             obs.append('kind') # this is type of observation
             obs.append(self.reverse_types[data[self.n_copies + 3]])  # observation type
+            # Convert metadata to a string and append
+            metadata = ' '.join(map(str, data[self.n_copies + 4:-4]))
+            if metadata:
+                obs.append(metadata)  # metadata
         obs.append(' '.join(map(str, data[-4:-2])))  # seconds, days
         obs.append(data[-1])  # obs error variance
 
@@ -273,12 +338,70 @@ class obs_sequence:
         elif self.loc_mod == 'loc1d':
             heading.append('location')
         heading.append('type')
+        heading.append('metadata')
+        heading.append('external_FO')
         heading.append('seconds')
         heading.append('days')
         heading.append('time')
         heading.append('obs_err_var')
         return heading
 
+    @requires_assimilation_info    
+    def select_by_dart_qc(self, dart_qc):
+        """
+        Selects rows from a DataFrame based on the DART quality control flag.
+
+        Parameters:
+            df (DataFrame): A pandas DataFrame.
+            dart_qc (int): The DART quality control flag to select.
+
+        Returns:
+            DataFrame: A DataFrame containing only the rows with the specified DART quality control flag.
+
+        Raises:
+            ValueError: If the DART quality control flag is not present in the DataFrame.
+        """
+        if dart_qc not in self.df['DART_quality_control'].unique():
+            raise ValueError(f"DART quality control flag '{dart_qc}' not found in DataFrame.")
+        else:
+            return self.df[self.df['DART_quality_control'] == dart_qc]
+
+    @requires_assimilation_info
+    def select_failed_qcs(self):
+        """
+        Select rows from the DataFrame where the DART quality control flag is greater than 0.
+
+        Returns:
+            pandas.DataFrame: A DataFrame containing only the rows with a DART quality control flag greater than 0.
+        """
+        return self.df[self.df['DART_quality_control'] > 0]
+
+    @requires_assimilation_info
+    def possible_vs_used(self):
+        """
+        Calculates the count of possible vs. used observations by type.
+
+        This function takes a DataFrame containing observation data, including a 'type' column for the observation
+        type and an 'observation' column. The number of used observations ('used'), is the total number
+        minus the observations that failed quality control checks (as determined by the `select_failed_qcs` function).
+        The result is a DataFrame with each observation type, the count of possible observations, and the count of
+        used observations.
+
+        Returns:
+            pd.DataFrame: A DataFrame with three columns: 'type', 'possible', and 'used'. 'type' is the observation type,
+            'possible' is the count of all observations of that type, and 'used' is the count of observations of that type
+            that passed quality control checks.
+        """
+        possible = self.df.groupby('type')['observation'].count()
+        possible.rename('possible', inplace=True)
+        
+        failed_qcs = self.select_failed_qcs().groupby('type')['observation'].count()
+        used = possible - failed_qcs.reindex(possible.index, fill_value=0)
+        used.rename('used', inplace=True)
+        
+        return pd.concat([possible, used], axis=1).reset_index()
+
+
     @staticmethod
     def is_binary(file):
         """Check if a file is binary file."""
@@ -659,65 +782,6 @@ def convert_dart_time(seconds, days):
     """
     time = dt.datetime(1601,1,1) + dt.timedelta(days=days, seconds=seconds)
     return time
-    
-def select_by_dart_qc(df, dart_qc):
-    """
-    Selects rows from a DataFrame based on the DART quality control flag.
-
-    Parameters:
-        df (DataFrame): A pandas DataFrame.
-        dart_qc (int): The DART quality control flag to select.
-
-    Returns:
-        DataFrame: A DataFrame containing only the rows with the specified DART quality control flag.
-
-    Raises:
-        ValueError: If the DART quality control flag is not present in the DataFrame.
-    """
-    if dart_qc not in df['DART_quality_control'].unique():
-        raise ValueError(f"DART quality control flag '{dart_qc}' not found in DataFrame.")
-    else:
-        return df[df['DART_quality_control'] == dart_qc]
-
-def select_failed_qcs(df):
-    """
-    Selects rows from a DataFrame where the DART quality control flag is greater than 0.
-
-    Parameters:
-        df (DataFrame): A pandas DataFrame.
-
-    Returns:
-        DataFrame: A DataFrame containing only the rows with a DART quality control flag greater than 0.
-    """
-    return df[df['DART_quality_control'] > 0]
-
-def possible_vs_used(df):
-    """
-    Calculates the count of possible vs. used observations by type.
-
-    This function takes a DataFrame containing observation data, including a 'type' column for the observation
-    type and an 'observation' column. The number of used observations ('used'), is the total number
-    minus the observations that failed quality control checks (as determined by the `select_failed_qcs` function).
-    The result is a DataFrame with each observation type, the count of possible observations, and the count of
-    used observations.
-
-    Parameters:
-        df (pd.DataFrame): A DataFrame with at least two columns: 'type' for the observation type and 'observation'
-        for the observation data. It may also contain other columns required by the `select_failed_qcs` function
-        to determine failed quality control checks.
-
-    Returns:
-        pd.DataFrame: A DataFrame with three columns: 'type', 'possible', and 'used'. 'type' is the observation type,
-        'possible' is the count of all observations of that type, and 'used' is the count of observations of that type
-        that passed quality control checks.
-
-    """
-    possible = df.groupby('type')['observation'].count()
-    possible.rename('possible', inplace=True)
-    used = df.groupby('type')['observation'].count() - select_failed_qcs(df).groupby('type')['observation'].count()
-    used.rename('used', inplace=True)
-    return pd.concat([possible, used], axis=1).reset_index()
-
 
 def construct_composit(df_comp, composite, components):
     """

pydartdiags-0.0.43/src/pydartdiags/plots/plots.py

@@ -0,0 +1,339 @@
+
+import numpy as np
+import plotly.express as px
+import plotly.graph_objects as go
+import pandas as pd
+
+def plot_rank_histogram(df):  
+    """
+    Plots a rank histogram colored by observation type.
+
+    All histogram bars are initalized to be hidden and can be toggled visible in the plot's legend
+    """
+    _, _, df_hist = calculate_rank(df)
+    fig = px.histogram(df_hist, x='rank', color='obstype', title='Histogram Colored by obstype')
+    for trace in fig.data:
+        trace.visible = 'legendonly'
+    fig.show()
+
+
+def calculate_rank(df):
+    """
+    Calculate the rank of observations within an ensemble.
+
+    This function takes a DataFrame containing ensemble predictions and observed values,
+    adds sampling noise to the ensemble predictions, and calculates the rank of the observed
+    value within the perturbed ensemble for each observation. The rank indicates the position
+    of the observed value within the sorted ensemble values, with 1 being the lowest. If the
+    observed value is larger than the largest ensemble member, its rank is set to the ensemble
+    size plus one.
+
+    Parameters:
+        df (pd.DataFrame): A DataFrame with columns for mean, standard deviation, observed values,
+                           ensemble size, and observation type. The DataFrame should have one row per observation.
+
+    Returns:
+        tuple: A tuple containing the rank array, ensemble size, and a result DataFrame. The result
+        DataFrame contains columns for 'rank' and 'obstype'.
+    """
+    ensemble_values = df.filter(regex='prior_ensemble_member').to_numpy().copy()
+    std_dev = np.sqrt(df['obs_err_var']).to_numpy()
+    obsvalue = df['observation'].to_numpy()
+    obstype = df['type'].to_numpy()
+    ens_size = ensemble_values.shape[1]
+    mean = 0.0 # mean of the sampling noise
+    rank = np.zeros(obsvalue.shape[0], dtype=int)
+    
+    for obs in range(ensemble_values.shape[0]):
+        sampling_noise = np.random.normal(mean, std_dev[obs], ens_size)
+        ensemble_values[obs] += sampling_noise
+        ensemble_values[obs].sort()
+        for i, ens in enumerate(ensemble_values[obs]):
+            if obsvalue[obs] <= ens:
+                rank[obs] = i + 1
+                break
+
+        if rank[obs] == 0: # observation is larger than largest ensemble member
+            rank[obs] = ens_size + 1
+
+    result_df = pd.DataFrame({
+        'rank': rank,
+        'obstype': obstype
+    })
+
+    return (rank, ens_size, result_df)
+
+def plot_profile(df, levels, verticalUnit = "pressure (Pa)"):
+    """
+    Plots RMSE, bias, and total spread profiles for different observation types across specified vertical levels.
+
+    This function takes a DataFrame containing observational data and model predictions, categorizes
+    the data into specified vertical levels, and calculates the RMSE, bias and total spread for each level and
+    observation type. It then plots three line charts: one for RMSE, one for bias, one for total spread, as functions
+    of vertical level. The vertical levels are plotted on the y-axis in reversed order to represent
+    the vertical profile in the atmosphere correctly if the vertical units are pressure.
+
+    Parameters:
+        df (pd.DataFrame): The input DataFrame containing at least the 'vertical' column for vertical levels,
+        the vert_unit column, and other columns required by the `rmse_bias` function for calculating RMSE and 
+        Bias.
+        levels (array-like): The bin edges for categorizing the 'vertical' column values into the desired 
+        vertical levels.
+        verticalUnit (string) (optional): The vertical unit to be used. Only observations in df which have this
+        string in the vert_unit column will be plotted. Defaults to 'pressure (Pa)'.
+
+    Returns:
+        tuple: A tuple containing the DataFrame with RMSE, bias and total spread calculations,
+        The DataFrame includes a 'vlevels' column representing the categorized vertical levels
+        and 'midpoint' column representing the midpoint of each vertical level bin. And the three figures.
+
+    Raises:
+        ValueError: If there are missing values in the 'vertical' column of the input DataFrame.
+        ValueError: If none of the input obs have 'verticalUnit' in the 'vert_unit' column of the input DataFrame.
+
+    Note:
+        - The function modifies the input DataFrame by adding 'vlevels' and 'midpoint' columns.
+        - The 'midpoint' values are calculated as half the midpoint of each vertical level bin, which may need
+          adjustment based on the specific requirements for vertical level representation.
+        - The plots are generated using Plotly Express and are displayed inline. The y-axis of the plots is
+          reversed to align with standard atmospheric pressure level representation if the vertical units
+          are atmospheric pressure.
+    """
+
+    pd.options.mode.copy_on_write = True
+    if df['vertical'].isnull().values.any(): # what about horizontal observations?
+        raise ValueError("Missing values in 'vertical' column.")
+    elif verticalUnit not in df['vert_unit'].values:
+        raise ValueError("No obs with expected vertical unit '"+verticalUnit+"'.")
+    else:
+        df = df[df["vert_unit"].isin({verticalUnit})] # Subset to only rows with the correct vertical unit
+        df.loc[:,'vlevels'] = pd.cut(df['vertical'], levels)
+        if verticalUnit == "pressure (Pa)":
+            df.loc[:,'midpoint'] = df['vlevels'].apply(lambda x: x.mid / 100.) # HK todo units
+        else:
+            df.loc[:,'midpoint'] = df['vlevels'].apply(lambda x: x.mid)
+
+    # Calculations
+    df_profile_prior = rmse_bias_totalspread(df, phase='prior')
+    df_profile_posterior = None
+    if 'posterior_ensemble_mean' in df.columns:
+        df_profile_posterior = rmse_bias_totalspread(df, phase='posterior')
+
+    # Merge prior and posterior dataframes
+    if df_profile_posterior is not None:
+        df_profile = pd.merge(df_profile_prior, df_profile_posterior, on=['midpoint', 'type'], suffixes=('_prior', '_posterior'))
+        fig_rmse = plot_profile_prior_post(df_profile, 'rmse', verticalUnit)
+        fig_rmse.show()
+        fig_bias = plot_profile_prior_post(df_profile, 'bias', verticalUnit)
+        fig_bias.show()
+        fig_ts = plot_profile_prior_post(df_profile, 'totalspread', verticalUnit)
+        fig_ts.show()
+    else:
+        df_profile = df_profile_prior
+        fig_rmse = plot_profile_prior(df_profile, 'rmse', verticalUnit)
+        fig_rmse.show()
+        fig_bias = plot_profile_prior(df_profile, 'bias', verticalUnit)
+        fig_bias.show()
+        fig_ts = plot_profile_prior(df_profile, 'totalspread', verticalUnit)
+        fig_ts.show()   
+
+    return df_profile, fig_rmse, fig_ts, fig_bias
+    
+def plot_profile_prior_post(df_profile, stat, verticalUnit):
+    """
+    Plots prior and posterior statistics by vertical level for different observation types.
+
+    Parameters:
+        df_profile (pd.DataFrame): DataFrame containing the prior and posterior statistics.
+        stat (str): The statistic to plot (e.g., 'rmse', 'bias', 'totalspread').
+        verticalUnit (str): The unit of the vertical axis (e.g., 'pressure (Pa)').
+
+    Returns:
+        plotly.graph_objects.Figure: The generated Plotly figure.
+    """
+    # Reshape DataFrame to long format for easier plotting
+    df_long = pd.melt(
+        df_profile,
+        id_vars=["midpoint", "type"],
+        value_vars=["prior_"+stat, "posterior_"+stat],
+        var_name=stat+"_type",
+        value_name=stat+"_value"
+    )
+
+    # Define a color mapping for observation each type
+    unique_types = df_long["type"].unique()
+    colors = px.colors.qualitative.Plotly
+    color_mapping = {type_: colors[i % len(colors)] for i, type_ in enumerate(unique_types)}
+
+    # Create a mapping for line styles based on stat
+    line_styles = {"prior_"+stat: "solid", "posterior_"+stat: "dash"}
+
+    # Create the figure
+    fig_stat = go.Figure()
+
+    # Loop through each type and type to add traces
+    for t in df_long["type"].unique():
+        for stat_type, dash_style in line_styles.items():
+            # Filter the DataFrame for this type and stat
+            df_filtered = df_long[(df_long[stat+"_type"] == stat_type) & (df_long["type"] == t)]
+            
+            # Add a trace
+            fig_stat.add_trace(go.Scatter(
+                x=df_filtered[stat+"_value"],
+                y=df_filtered["midpoint"],
+                mode='lines+markers',
+                name='prior '+t if stat_type == "prior_"+stat else 'post ',  # Show legend for "prior_stat OBS TYPE" only
+                line=dict(dash=dash_style, color=color_mapping[t]),  # Same color for all traces in group
+                marker=dict(size=8, color=color_mapping[t]),
+                legendgroup=t  # Group traces by type
+            ))
+
+        # Update layout
+        fig_stat.update_layout(
+            title= stat+' by Level',
+            xaxis_title=stat,
+            yaxis_title=verticalUnit,
+            width=800,
+            height=800,
+            template="plotly_white"
+        )
+
+    if verticalUnit == "pressure (Pa)":
+        fig_stat.update_yaxes(autorange="reversed")
+
+    return fig_stat
+
+
+def plot_profile_prior(df_profile, stat, verticalUnit):
+    """
+    Plots prior statistics by vertical level for different observation types.
+
+    Parameters:
+        df_profile (pd.DataFrame): DataFrame containing the prior statistics.
+        stat (str): The statistic to plot (e.g., 'rmse', 'bias', 'totalspread').
+        verticalUnit (str): The unit of the vertical axis (e.g., 'pressure (Pa)').
+
+    Returns:
+        plotly.graph_objects.Figure: The generated Plotly figure.
+    """
+    # Reshape DataFrame to long format for easier plotting - not needed for prior only, but
+    #   leaving it in for consistency with the plot_profile_prior_post function for now
+    df_long = pd.melt(
+        df_profile,
+        id_vars=["midpoint", "type"],
+        value_vars=["prior_"+stat],
+        var_name=stat+"_type",
+        value_name=stat+"_value"
+    )
+
+    # Define a color mapping for observation each type
+    unique_types = df_long["type"].unique()
+    colors = px.colors.qualitative.Plotly
+    color_mapping = {type_: colors[i % len(colors)] for i, type_ in enumerate(unique_types)}
+
+    # Create the figure
+    fig_stat = go.Figure()
+
+    # Loop through each type to add traces
+    for t in df_long["type"].unique():
+        # Filter the DataFrame for this type and stat
+        df_filtered = df_long[(df_long["type"] == t)]
+        
+        # Add a trace
+        fig_stat.add_trace(go.Scatter(
+            x=df_filtered[stat+"_value"],
+            y=df_filtered["midpoint"],
+            mode='lines+markers',
+            name='prior ' + t,
+            line=dict(color=color_mapping[t]),  # Same color for all traces in group
+            marker=dict(size=8, color=color_mapping[t]),
+            legendgroup=t  # Group traces by type
+        ))
+
+    # Update layout
+    fig_stat.update_layout(
+        title=stat + ' by Level',
+        xaxis_title=stat,
+        yaxis_title=verticalUnit,
+        width=800,
+        height=800,
+        template="plotly_white"
+    )
+
+    if verticalUnit == "pressure (Pa)":
+        fig_stat.update_yaxes(autorange="reversed")
+        
+    return fig_stat
+
+    
+def mean_then_sqrt(x):
+    """
+    Calculates the mean of an array-like object and then takes the square root of the result.
+
+    Parameters:
+        arr (array-like): An array-like object (such as a list or a pandas Series). 
+                          The elements should be numeric.
+
+    Returns:
+        float: The square root of the mean of the input array.
+
+    Raises:
+        TypeError: If the input is not an array-like object containing numeric values.
+         ValueError: If the input array is empty.
+    """
+        
+    return np.sqrt(np.mean(x))
+
+def rmse_bias_totalspread(df, phase='prior'): 
+    if phase == 'prior':
+        sq_err_column = 'prior_sq_err'
+        bias_column = 'prior_bias'
+        rmse_column = 'prior_rmse'
+        spread_column = 'prior_ensemble_spread'
+        totalspread_column = 'prior_totalspread'
+    elif phase == 'posterior':
+        sq_err_column = 'posterior_sq_err'
+        bias_column = 'posterior_bias'
+        rmse_column = 'posterior_rmse'
+        spread_column = 'posterior_ensemble_spread'
+        totalspread_column = 'posterior_totalspread'
+    else:
+        raise ValueError("Invalid phase. Must be 'prior' or 'posterior'.")
+
+    rmse_bias_ts_df = df.groupby(['midpoint', 'type'], observed=False).agg({
+        sq_err_column: mean_then_sqrt,
+        bias_column: 'mean',
+        spread_column: mean_then_sqrt,
+        'obs_err_var': mean_then_sqrt     
+    }).reset_index()
+
+    # Add column for totalspread
+    rmse_bias_ts_df[totalspread_column] = np.sqrt(rmse_bias_ts_df[spread_column] + rmse_bias_ts_df['obs_err_var'])
+
+    # Rename square error to root mean square error
+    rmse_bias_ts_df.rename(columns={sq_err_column: rmse_column}, inplace=True)
+    
+    return rmse_bias_ts_df
+
+def rmse_bias_by_obs_type(df, obs_type):
+    """
+    Calculate the RMSE and bias for a given observation type.
+
+    Parameters:
+        df (DataFrame): A pandas DataFrame.
+        obs_type (str): The observation type for which to calculate the RMSE and bias.
+
+    Returns:
+        DataFrame: A DataFrame containing the RMSE and bias for the given observation type.
+
+    Raises:
+        ValueError: If the observation type is not present in the DataFrame.
+    """
+    if obs_type not in df['type'].unique():
+        raise ValueError(f"Observation type '{obs_type}' not found in DataFrame.")
+    else:
+        obs_type_df = df[df['type'] == obs_type]
+        obs_type_agg = obs_type_df.groupby('vlevels', observed=False).agg({'sq_err':mean_then_sqrt, 'bias':'mean'}).reset_index()
+        obs_type_agg.rename(columns={'sq_err':'rmse'}, inplace=True)
+        return obs_type_agg
+