Skip to content

CoMPASS Level 2 - Models

Bayesian Gaussian Mixture Models and GMM-HMM for cognitive state inference.

Overview

This module implements the Level 2 models that identify goal-directed cognitive states:

  • BGMM (Bayesian Gaussian Mixture Model): Clusters behavioral features
  • GMM-HMM: Models temporal dependencies across states
  • Cross-validation for model selection
  • State sequence decoding

The models identify:

  • Oriented states: Goal-directed behavior toward target
  • Non-Oriented states: Exploratory or non-goal-directed behavior

Model Selection

Use run_compass with cross-validation to automatically select the optimal number of states.

Main Modeling Function

run_compass

compass_labyrinth.compass.level_2.model.run_compass 

run_compass(
    config: dict,
    df: DataFrame,
    features: list,
    phase_options: list = [5],
    ncomp_options: list[int] = [2, 3],
    k_options: list[int] = [2, 3],
    reg_options: list = [0.0001, 1e-05, 1e-06],
    terminal_values: list = [47],
    bout_col: str = "Bout_ID",
    patience: None | str = None,
    patience_candidates: list = [2, 3, 5, 10],
    verbose: bool = False,
) -> tuple[pd.DataFrame, list]

Run CoMPASS.

Parameters:

  • config (dict) –

    Configuration dictionary.

  • df (DataFrame) –

    Input dataframe.

  • features (list) –

    List of feature column names to use.

  • phase_options (list, default: [5] ) –

    List of phase options to test (default is [5]).

  • ncomp_options (list, default: [2, 3] ) –

    Range of number of components to test (default is [2, 3]).

  • k_options (list, default: [2, 3] ) –

    Range of k values to test (default is [2, 3]).

  • reg_options (list, default: [0.0001, 1e-05, 1e-06] ) –

    List of regularization values to test (default is [1e-4, 1e-5, 1e-6]).

  • terminal_values (list, default: [47] ) –

    List of terminal grid values (default is [47]).

  • bout_col (str, default: 'Bout_ID' ) –

    Name of the bout column (default is "Bout_ID").

  • patience (None or str, default: None ) –

    Patience setting for early stopping (default is None). Set to 'tune' if wanted to apply patience window.

  • patience_candidates (list, default: [2, 3, 5, 10] ) –

    List of patience candidates to test if patience is 'tune' (default is [2, 3, 5, 10]). Only used if patience is set to 'tune'.

  • verbose (bool, default: False ) –

    Whether to print detailed logs during model training (default is False).

Returns:

  • tuple –

    A tuple containing: - pd.DataFrame: DataFrame with assigned Level 2 states. - list: List of all CV results for visualization.

Source code in src/compass_labyrinth/compass/level_2/model.py 
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
def run_compass(
    config: dict,
    df: pd.DataFrame,
    features: list,
    phase_options: list = [5],
    ncomp_options: list[int] = [2, 3],
    k_options: list[int] = [2, 3],
    reg_options: list = [1e-4, 1e-5, 1e-6],
    terminal_values: list = [47],
    bout_col: str = "Bout_ID",
    patience: None | str = None,
    patience_candidates: list = [2, 3, 5, 10],
    verbose: bool = False,
) -> tuple[pd.DataFrame, list]:
    """
    Run CoMPASS.

    Parameters
    -----------
    config : dict
        Configuration dictionary.
    df : pd.DataFrame
        Input dataframe.
    features : list
        List of feature column names to use.
    phase_options : list, optional
        List of phase options to test (default is [5]).
    ncomp_options : list, optional
        Range of number of components to test (default is [2, 3]).
    k_options : list, optional
        Range of k values to test (default is [2, 3]).
    reg_options : list, optional
        List of regularization values to test (default is [1e-4, 1e-5, 1e-6]).
    terminal_values : list, optional
        List of terminal grid values (default is [47]).
    bout_col : str, optional
        Name of the bout column (default is "Bout_ID").
    patience : None or str, optional
        Patience setting for early stopping (default is None).
        Set to 'tune' if wanted to apply patience window.
    patience_candidates : list, optional
        List of patience candidates to test if patience is 'tune' (default is [2, 3, 5, 10]).
        Only used if patience is set to 'tune'.
    verbose : bool, optional
        Whether to print detailed logs during model training (default is False).

    Returns
    --------
    tuple
        A tuple containing:
        - pd.DataFrame: DataFrame with assigned Level 2 states.
        - list: List of all CV results for visualization.
    """
    all_results = []
    final_sess_data = []

    df = assign_bouts_per_session(df, terminal_values=terminal_values, bout_col=bout_col)

    for n_phases in phase_options:
        sessions = df.Session.unique()
        phase_labels = range(n_phases)
        phase_map = build_phase_map(df, n_phases)

        for phase_index in phase_labels:
            for test_sess in sessions:
                print(f"\n=== CV | Test: Session {test_sess} - Phase {phase_index+1}/{n_phases} ===")

                test_bouts = phase_map[(test_sess, phase_index)]
                df_test = df[(df.Session == test_sess) & (df[bout_col].isin(test_bouts))]

                train_sessions = [s for s in sessions if s != test_sess]
                df_train_pool = pd.concat(
                    [df[(df.Session == s) & (df[bout_col].isin(phase_map[(s, phase_index)]))] for s in train_sessions]
                )

                inner_sessions = df_train_pool.Session.unique()
                best_log_lik = -np.inf
                best_aic = np.inf
                best_model = None

                best_patience = patience_candidates[0] if patience == "tune" else patience
                patience_results = {}

                for test_patience in (patience_candidates if patience == "tune" else [best_patience]):
                    log_liks, aics, param_labels = [], [], []
                    no_improve = 0
                    best_inner_loglik = -np.inf

                    early_stopped = False

                    for val_sess in inner_sessions:
                        df_val = df_train_pool[df_train_pool.Session == val_sess]
                        df_train = df_train_pool[df_train_pool.Session != val_sess]

                        for ncomp in ncomp_options:
                            for k in k_options:
                                for reg_val in reg_options:
                                    try:
                                        X_train = df_train[features].values
                                        X_val = df_val[features].values

                                        gmm = initialize_bgmm(ncomp, reg_val)
                                        gmm.fit(X_train)
                                        covariances = regularize_covariances(gmm.covariances_, reg_val)

                                        means = np.tile(gmm.means_[None, :, :], (ncomp, 1, 1))
                                        covars = np.tile(covariances[None, :, :, :], (ncomp, 1, 1, 1))
                                        weights = np.tile(gmm.weights_[None, :], (ncomp, 1))

                                        model = initialize_gmmhmm(ncomp, k, means, covars, weights)
                                        model.fit(X_train)

                                        log_lik = model.score(X_val)
                                        aic = compute_aic(model, X_train)
                                        label = f"n={ncomp},k={k},r={reg_val:.0e}"

                                        log_liks.append(log_lik)
                                        aics.append(aic)
                                        param_labels.append(label)

                                        if log_lik > best_inner_loglik:
                                            best_inner_loglik = log_lik
                                            best_model = model
                                            no_improve = 0
                                        else:
                                            no_improve += 1

                                        if test_patience is not None and no_improve >= test_patience:
                                            early_stopped = True
                                            break
                                    except Exception as e:
                                        if verbose:
                                            print(f"Model failed for n={ncomp}, k={k}, reg={reg_val:.0e}: {e}")
                                        continue
                                if early_stopped:
                                    break
                            if early_stopped:
                                break

                    avg_loglik = np.mean(log_liks)
                    patience_results[test_patience] = (avg_loglik, best_model, log_liks, aics, param_labels)

                # Pick best patience if tuning
                if patience == "tune":
                    best_patience, (best_avg_loglik, best_model, log_liks, aics, param_labels) = max(
                        patience_results.items(), key=lambda x: x[1][0]
                    )
                    print(f"Optimal patience for Session {test_sess}, Phase {phase_index+1}: {best_patience}")

                if best_model is not None:
                    X_test = df_test[features].values
                    df_test = df_test.copy()
                    df_test["Level_2_States"] = best_model.predict(X_test)
                    final_sess_data.append(df_test)

                tag = f"Session:{test_sess}_PhaseIndex:{phase_index+1}_NumPhases:{n_phases}_Patience:{best_patience}"
                all_results.append((tag, log_liks, aics, param_labels))

    df_hier = pd.concat(final_sess_data)

    # Save results
    save_path = Path(config["project_path_full"]) / "csvs" / "combined" / "hhmm_state_file.csv"
    df_hier.to_csv(save_path, index=False)
    print(f"HHMM state file saved at: {save_path}")

    return df_hier, all_results

Model Initialization

initialize_bgmm

compass_labyrinth.compass.level_2.model.initialize_bgmm 

initialize_bgmm(n_components, reg_val, random_state=3)
Source code in src/compass_labyrinth/compass/level_2/model.py 
44
45
46
47
def initialize_bgmm(n_components, reg_val, random_state=3):
    return BayesianGaussianMixture(
        n_components=n_components, covariance_type="full", reg_covar=reg_val, random_state=random_state, max_iter=500
    )

initialize_gmmhmm

compass_labyrinth.compass.level_2.model.initialize_gmmhmm 

initialize_gmmhmm(
    n_states,
    n_mix,
    means,
    covariances,
    weights,
    random_state=3,
)
Source code in src/compass_labyrinth/compass/level_2/model.py 
53
54
55
56
57
58
def initialize_gmmhmm(n_states, n_mix, means, covariances, weights, random_state=3):
    model = GMMHMM(n_components=n_states, n_mix=n_mix, covariance_type="full", random_state=random_state, n_iter=500)
    model.means_ = means[:, :n_mix, :]
    model.covars_ = covariances[:, :n_mix, :, :]
    model.weights_ = weights[:, :n_mix]
    return model

Model Evaluation

compute_aic

compass_labyrinth.compass.level_2.model.compute_aic 

compute_aic(model, X)
Source code in src/compass_labyrinth/compass/level_2/model.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
def compute_aic(model, X):
    log_likelihood = model.score(X)
    n_states = model.n_components
    n_features = model.means_.shape[-1]
    n_mixtures = model.n_mix
    num_params = (
        (n_states - 1)  # start prob
        + n_states * (n_states - 1)  # transition matrix
        + n_states * n_mixtures * n_features  # means
        + n_states * n_mixtures * n_features  # covariances
        + n_states * n_mixtures  # weights
    )
    return 2 * num_params - 2 * log_likelihood

visualize_cv_results

compass_labyrinth.compass.level_2.model.visualize_cv_results 

visualize_cv_results(
    config: dict,
    all_results: list,
    save_fig: bool = True,
    show_fig: bool = True,
    return_fig: bool = False,
) -> None | list[plt.Figure]

Visualize cross-validation results.

Parameters:

  • config (dict) –

    Configuration dictionary.

  • all_results (list) –

    List of tuples containing CV results.

  • save_fig (bool, default: True ) –

    Whether to save the figures (default is True).

  • show_fig (bool, default: True ) –

    Whether to show the figures (default is True).

  • return_fig (bool, default: False ) –

    Whether to return the figures (default is False).

Returns:

  • None or list of plt.Figure –

    List of figures if return_fig is True, otherwise None.

Source code in src/compass_labyrinth/compass/level_2/model.py 
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
def visualize_cv_results(
    config: dict,
    all_results: list,
    save_fig: bool = True,
    show_fig: bool = True,
    return_fig: bool = False,
) -> None | list[plt.Figure]:
    """
    Visualize cross-validation results.

    Parameters
    -----------
    config : dict
        Configuration dictionary.
    all_results : list
        List of tuples containing CV results.
    save_fig : bool, optional
        Whether to save the figures (default is True).
    show_fig : bool, optional
        Whether to show the figures (default is True).
    return_fig : bool, optional
        Whether to return the figures (default is False).

    Returns
    --------
    None or list of plt.Figure
        List of figures if return_fig is True, otherwise None.
    """
    all_figs = list()
    for tag, log_liks, aics, param_labels in all_results:
        fig = plt.figure(figsize=(14, 5))

        plt.subplot(1, 2, 1)
        sns.lineplot(x=np.arange(len(log_liks)), y=log_liks, marker="o")
        plt.xticks(ticks=np.arange(len(param_labels)), labels=param_labels, rotation=90)
        plt.title(f"{tag} - Log-Likelihoods")
        plt.xlabel("Param Config (n,k,reg)")
        plt.ylabel("Log-Likelihood")

        plt.subplot(1, 2, 2)
        sns.lineplot(x=np.arange(len(aics)), y=aics, marker="o")
        plt.xticks(ticks=np.arange(len(param_labels)), labels=param_labels, rotation=90)
        plt.title(f"{tag} - AIC")
        plt.xlabel("Param Config (n,k,reg)")
        plt.ylabel("AIC")

        plt.tight_layout()

        # Save figure
        if save_fig:
            prefix = tag.replace(":", "-").replace(",", "_").replace(" ", "_")
            save_path = Path(config["project_path_full"]) / "figures" / f"level_2_cv_performance_{prefix}.pdf"
            plt.savefig(save_path, bbox_inches="tight", dpi=300)
            print(f"Figure saved at: {save_path}")

        # Show figure
        if show_fig:
            plt.show()

        # Return figure
        if return_fig:
            all_figs.append(fig)

    if return_fig:
        return all_figs

regularize_covariances

compass_labyrinth.compass.level_2.model.regularize_covariances 

regularize_covariances(covariances, reg_val=1e-06)
Source code in src/compass_labyrinth/compass/level_2/model.py 
37
38
def regularize_covariances(covariances, reg_val=1e-6):
    return np.array([cov + np.eye(cov.shape[0]) * reg_val for cov in covariances])

State Analysis

get_unique_states

compass_labyrinth.compass.level_2.model.get_unique_states 

get_unique_states(df, state_col='Level_2_States')

Return sorted unique states from the specified column.

Source code in src/compass_labyrinth/compass/level_2/model.py 
305
306
307
def get_unique_states(df, state_col="Level_2_States"):
    """Return sorted unique states from the specified column."""
    return sorted(df[state_col].dropna().unique())

generate_state_color_map

compass_labyrinth.compass.level_2.model.generate_state_color_map 

generate_state_color_map(unique_states, palette='tab10')

Assign each state a color from the selected Seaborn palette.

Source code in src/compass_labyrinth/compass/level_2/model.py 
310
311
312
313
def generate_state_color_map(unique_states, palette="tab10"):
    """Assign each state a color from the selected Seaborn palette."""
    colors = sns.color_palette(palette, len(unique_states))
    return {state: colors[i] for i, state in enumerate(unique_states)}

assign_reward_orientation

compass_labyrinth.compass.level_2.model.assign_reward_orientation 

assign_reward_orientation(
    df: DataFrame,
    angle_col: str = "Targeted_Angle_smooth_abs",
    level_2_state_col: str = "Level_2_States",
    session_col: str = "Session",
) -> pd.DataFrame

Assigns reward orientation labels ('Reward Oriented' or 'Non-Reward Oriented') to Level 2 states per session, based on the median Targeted_Angle_smooth within each state and relative to session median.

Parameters:

  • df (DataFrame) –

    Input dataframe with columns for session, level 2 state, and angle.

  • angle_col (str, default: 'Targeted_Angle_smooth_abs' ) –

    Column name representing the smoothed targeted angle.

  • level_2_state_col (str, default: 'Level_2_States' ) –

    Column name for level 2 HMM states.

  • session_col (str, default: 'Session' ) –

    Column name for session identifier.

Returns:

  • DataFrame –

    Updated dataframe with 'Reward_Oriented' column.

Source code in src/compass_labyrinth/compass/level_2/model.py 
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
def assign_reward_orientation(
    df: pd.DataFrame,
    angle_col: str = "Targeted_Angle_smooth_abs",
    level_2_state_col: str = "Level_2_States",
    session_col: str = "Session",
) -> pd.DataFrame:
    """
    Assigns reward orientation labels ('Reward Oriented' or 'Non-Reward Oriented') to Level 2 states per session,
    based on the median Targeted_Angle_smooth within each state and relative to session median.

    Parameters
    -----------
    df : pd.DataFrame
        Input dataframe with columns for session, level 2 state, and angle.
    angle_col : str
        Column name representing the smoothed targeted angle.
    level_2_state_col : str
        Column name for level 2 HMM states.
    session_col : str
        Column name for session identifier.

    Returns
    --------
    pd.DataFrame
        Updated dataframe with 'Reward_Oriented' column.
    """
    df = df.copy()
    df["Reward_Oriented"] = np.nan

    for sess in df[session_col].unique():
        df_sess = df[df[session_col] == sess]
        state_medians = df_sess.groupby(level_2_state_col)[angle_col].median()
        session_median = df_sess[angle_col].median()

        if len(state_medians) == 2:
            reward_state = state_medians.idxmin()
            non_reward_state = state_medians.idxmax()
            df.loc[(df[session_col] == sess) & (df[level_2_state_col] == reward_state), "Reward_Oriented"] = (
                "Reward Oriented"
            )
            df.loc[(df[session_col] == sess) & (df[level_2_state_col] == non_reward_state), "Reward_Oriented"] = (
                "Non-Reward Oriented"
            )

        elif len(state_medians) == 3:
            sorted_states = state_medians.sort_values()
            reward_state = sorted_states.index[0]
            non_reward_state = sorted_states.index[2]
            middle_state = sorted_states.index[1]

            if state_medians[middle_state] <= session_median:
                middle_label = "Reward Oriented"
            else:
                middle_label = "Non-Reward Oriented"

            df.loc[(df[session_col] == sess) & (df[level_2_state_col] == reward_state), "Reward_Oriented"] = (
                "Reward Oriented"
            )
            df.loc[(df[session_col] == sess) & (df[level_2_state_col] == non_reward_state), "Reward_Oriented"] = (
                "Non-Reward Oriented"
            )
            df.loc[(df[session_col] == sess) & (df[level_2_state_col] == middle_state), "Reward_Oriented"] = (
                middle_label
            )

    return df

assign_hhmm_state

compass_labyrinth.compass.level_2.model.assign_hhmm_state 

assign_hhmm_state(
    df: DataFrame,
    level_1_state_col: str,
    level_2_state_col: str,
) -> pd.DataFrame

Assigns a final HHMM (Hierarchical Hidden Markov Model) state to the dataframe. The final HHMM state is based on the combination of level 1 and level 2 states.

Parameters:

  • df (DataFrame) –

    Dataframe containing the level_1_state_col and level_2_state_col columns.

  • level_1_state_col (str) –

    The name of the column representing the first-level HMM state.

  • level_2_state_col (str) –

    The name of the column representing the second-level state (reward-oriented or not).

Returns:

  • df ( DataFrame ) –

    DataFrame with an additional 'HHMM State' column indicating the final HHMM state.

Source code in src/compass_labyrinth/compass/level_2/model.py 
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
def assign_hhmm_state(
    df: pd.DataFrame,
    level_1_state_col: str,
    level_2_state_col: str,
) -> pd.DataFrame:
    """
    Assigns a final HHMM (Hierarchical Hidden Markov Model) state to the dataframe.
    The final HHMM state is based on the combination of level 1 and level 2 states.

    Parameters
    -----------
    df : pd.DataFrame
        Dataframe containing the level_1_state_col and level_2_state_col columns.
    level_1_state_col : str
        The name of the column representing the first-level HMM state.
    level_2_state_col : str
        The name of the column representing the second-level state (reward-oriented or not).

    Returns
    --------
    df : pd.DataFrame
        DataFrame with an additional 'HHMM State' column indicating the final HHMM state.
    """
    # Define the conditions for assigning the HHMM states
    conds = [
        (df[level_1_state_col] == 1) & (df[level_2_state_col] == "Non-Reward Oriented"),
        (df[level_1_state_col] == 1) & (df[level_2_state_col] == "Reward Oriented"),
        (df[level_1_state_col] == 2) & (df[level_2_state_col] == "Non-Reward Oriented"),
        (df[level_1_state_col] == 2) & (df[level_2_state_col] == "Reward Oriented"),
    ]

    # Define the labels for the corresponding HHMM states
    labels = [
        "Surveillance, Non-Reward Oriented",
        "Surveillance, Reward Oriented",
        "Ambulatory, Non-Reward Oriented",
        "Ambulatory, Reward Oriented",
    ]

    # Set the default value as 'NaN' (string) to match the data type of the labels
    df["HHMM State"] = np.select(conds, labels, default="NaN")

    return df

State Sequence Visualization

plot_state_sequence_for_session

compass_labyrinth.compass.level_2.model.plot_state_sequence_for_session 

plot_state_sequence_for_session(
    df_session: DataFrame,
    state_col: str = "Level_2_States",
    color_map: None | dict = None,
    title_prefix: str = "State Sequence",
) -> plt.Figure

Plot the state sequence using color bars for one session.

Source code in src/compass_labyrinth/compass/level_2/model.py 
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
def plot_state_sequence_for_session(
    df_session: pd.DataFrame,
    state_col: str = "Level_2_States",
    color_map: None | dict = None,
    title_prefix: str = "State Sequence",
) -> plt.Figure:
    """Plot the state sequence using color bars for one session."""
    df_session = df_session.reset_index(drop=True).copy()
    df_session["color"] = df_session[state_col].map(color_map)

    fig, ax = plt.subplots(figsize=(10, 3))
    for idx, row in df_session.iterrows():
        rect = patches.Rectangle((idx, 0), 1, 1, color=row["color"])
        ax.add_patch(rect)

    ax.set_xlim(df_session.index.min(), df_session.index.max() + 1)
    ax.set_yticks([])
    ax.set_title(f"{title_prefix} - Session {df_session['Session'].iloc[0]}")

    legend_handles = [patches.Patch(color=color_map[state], label=f"State {state}") for state in color_map]
    ax.legend(
        handles=legend_handles,
        title="States",
        bbox_to_anchor=(0.5, -0.15),
        loc="upper center",
        borderaxespad=0.0,
    )

    plt.tight_layout()
    return fig

plot_state_sequences

compass_labyrinth.compass.level_2.model.plot_state_sequences 

plot_state_sequences(
    config: dict,
    df: DataFrame,
    genotype: str = "WT-WT",
    state_col: str = "Level_2_States",
    sessions_to_plot: str | list | int = "all",
    title_prefix: str = "State Sequence",
    save_fig: bool = True,
    show_fig: bool = True,
    return_fig: bool = False,
) -> None | list[plt.Figure]

Plot state sequences for specified sessions and genotype.

Source code in src/compass_labyrinth/compass/level_2/model.py 
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
def plot_state_sequences(
    config: dict,
    df: pd.DataFrame,
    genotype: str = "WT-WT",
    state_col: str = "Level_2_States",
    sessions_to_plot: str | list | int = "all",  # Can be 'all', a list of session IDs, or an int (top n)
    title_prefix: str = "State Sequence",
    save_fig: bool = True,
    show_fig: bool = True,
    return_fig: bool = False,
) -> None | list[plt.Figure]:
    """Plot state sequences for specified sessions and genotype."""
    df_geno = df[df["Genotype"] == genotype]
    unique_states = get_unique_states(df_geno, state_col=state_col)
    color_map = generate_state_color_map(unique_states)

    # Determine which sessions to plot
    all_sessions = df_geno["Session"].unique()
    if isinstance(sessions_to_plot, int):
        selected_sessions = all_sessions[:sessions_to_plot]
    elif isinstance(sessions_to_plot, list):
        selected_sessions = [s for s in sessions_to_plot if s in all_sessions]
    else:
        selected_sessions = all_sessions

    # Plot each selected session
    all_figs = []
    for sess_id in selected_sessions:
        df_sess = df_geno[df_geno["Session"] == sess_id][[state_col, "Session"]]
        fig = plot_state_sequence_for_session(
            df_sess,
            state_col=state_col,
            color_map=color_map,
            title_prefix=title_prefix,
        )

        # Save figure
        if save_fig:
            save_path = Path(config["project_path_full"]) / "figures" / f"state_sequence_session_{sess_id}.pdf"
            fig.savefig(save_path, bbox_inches="tight", dpi=300)
            print(f"Figure saved at: {save_path}")

        # Show figure
        if show_fig:
            plt.show()

        # Return figure
        if return_fig:
            all_figs.append(fig)

    if return_fig:
        return all_figs

plot_hhmm_state_sequence

compass_labyrinth.compass.level_2.model.plot_hhmm_state_sequence 

plot_hhmm_state_sequence(
    config: dict,
    df: DataFrame,
    session_col: str = "Session",
    state_col: str = "HHMM State",
    session_id: None | int = None,
    title_prefix: str = "State Sequence",
    colors: None | dict = None,
    save_fig: bool = True,
    show_fig: bool = True,
    return_fig: bool = False,
) -> None | plt.Figure

Plots a rectangular sequence of HHMM states for a given session.

Parameters:

  • config (dict) –

    Configuration dictionary for the project.

  • df (DataFrame) –

    DataFrame containing session and HHMM state columns.

  • session_col (str, default: 'Session' ) –

    Name of the column indicating session.

  • state_col (str, default: 'HHMM State' ) –

    Name of the column containing HHMM state labels.

  • session_id (None | int, default: None ) –

    Specific session to plot. If None, plots all sessions.

  • title_prefix (str, default: 'State Sequence' ) –

    Custom title prefix for plots.

  • colors (None | dict, default: None ) –

    Dictionary mapping HHMM states to colors. If None, default colors are used.

  • save_fig (bool, default: True ) –

    Whether to save the figure.

  • show_fig (bool, default: True ) –

    Whether to display the figure.

  • return_fig (bool, default: False ) –

    Whether to return the figure object.

Returns:

  • None or Figure –

    The figure object if return_fig is True, otherwise None.

Source code in src/compass_labyrinth/compass/level_2/model.py 
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
def plot_hhmm_state_sequence(
    config: dict,
    df: pd.DataFrame,
    session_col: str = "Session",
    state_col: str = "HHMM State",
    session_id: None | int = None,
    title_prefix: str = "State Sequence",
    colors: None | dict = None,
    save_fig: bool = True,
    show_fig: bool = True,
    return_fig: bool = False,
) -> None | plt.Figure:
    """
    Plots a rectangular sequence of HHMM states for a given session.

    Parameters
    -----------
    config : dict
        Configuration dictionary for the project.
    df : pd.DataFrame
        DataFrame containing session and HHMM state columns.
    session_col : str
        Name of the column indicating session.
    state_col : str
        Name of the column containing HHMM state labels.
    session_id : None | int
        Specific session to plot. If None, plots all sessions.
    title_prefix : str
        Custom title prefix for plots.
    colors : None | dict
        Dictionary mapping HHMM states to colors. If None, default colors are used.
    save_fig : bool
        Whether to save the figure.
    show_fig : bool
        Whether to display the figure.
    return_fig : bool
        Whether to return the figure object.

    Returns
    --------
    None or plt.Figure
        The figure object if return_fig is True, otherwise None.
    """
    sessions_to_plot = [session_id] if session_id is not None else df[session_col].unique()

    if colors is None:
        # Get unique states from all sessions to be plotted and generate color map
        unique_states = sorted(df[df[session_col].isin(sessions_to_plot)][state_col].dropna().unique())
        colors = generate_state_color_map(unique_states)
    all_figs = []
    for sess in sessions_to_plot:
        test = df.loc[df[session_col] == sess, [state_col]].reset_index(drop=True)
        test["color"] = test[state_col].map(colors)

        fig, ax = plt.subplots(figsize=(10, 3))
        for idx, row in test.iterrows():
            rect = patches.Rectangle((idx, 0), 1, 1, color=row["color"])
            ax.add_patch(rect)

        plt.yticks([])
        plt.xlim(test.index.min(), test.index.max() + 1)
        plt.title(f"{title_prefix} - Session {sess}")
        handles = [patches.Patch(color=color, label=label) for label, color in colors.items()]
        ax.legend(handles=handles, title="States", bbox_to_anchor=(0.5, -0.15), loc="upper center", borderaxespad=0.0)
        plt.tight_layout()

        # Save figure
        if save_fig:
            save_path = Path(config["project_path_full"]) / "figures" / f"hhmm_state_sequence_session_{sess}.pdf"
            plt.savefig(save_path, bbox_inches="tight", dpi=300)
            print(f"Figure saved at: {save_path}")

        # Show figure
        if show_fig:
            plt.show()

        # Return figure
        if return_fig:
            all_figs.append(fig)

    if return_fig:
        return all_figs