Add files via upload

added mermaid diagrams for key functionality

Author: SamoraHunter

assets/data_feature_importance_methods.mmd

@@ -0,0 +1,70 @@
+flowchart TD
+    A[ml_grid_object] --> B[feature_importance_methods]
+    
+    B --> C{Check feature_selection_method}
+    
+    C -->|"anova or None"| D[ANOVA Method]
+    C -->|"markov_blanket"| E[Markov Blanket Method]
+    
+    D --> F[feature_methods.getNfeaturesANOVAF]
+    E --> G[feature_methods.getNFeaturesMarkovBlanket]
+    
+    F --> H[Calculate F-values using f_classif]
+    F --> I[Sort features by F-value]
+    F --> J[Return top n feature names]
+    
+    G --> K[Initialize PPIMBC with SVC]
+    G --> L[Fit and transform training data]
+    G --> M[Extract top n features from MB]
+    
+    J --> N[Apply feature selection]
+    M --> N
+    
+    N --> O[Filter X_train with selected features]
+    N --> P[Filter X_test with selected features]
+    N --> Q[Filter X_test_orig with selected features]
+    
+    O --> R[Return filtered datasets]
+    P --> R
+    Q --> R
+    
+    subgraph "Input Data"
+        S[target_n_features]
+        T[X_train]
+        U[X_test]
+        V[y_train]
+        W[X_test_orig]
+    end
+    
+    subgraph "Feature Methods Class"
+        F
+        G
+        H
+        I
+        J
+        K
+        L
+        M
+    end
+    
+    subgraph "Output"
+        X[Filtered X_train]
+        Y[Filtered X_test]
+        Z[Filtered X_test_orig]
+    end
+    
+    S --> B
+    T --> B
+    U --> B
+    V --> B
+    W --> B
+    
+    R --> X
+    R --> Y
+    R --> Z
+    
+    style B fill:#e1f5fe
+    style C fill:#fff3e0
+    style D fill:#f3e5f5
+    style E fill:#e8f5e8
+    style N fill:#fff8e1

assets/data_feature_importance_methods.svg

@@ -0,0 +1,70 @@
+flowchart TD
+    A[Start Pipeline] --> B[Initialize Parameters]
+    B --> C{Read Data}
+    C -->|Sample Mode| D[Read Sample Data]
+    C -->|Full Mode| E[Read Full Data]
+    
+    D --> F[Apply Test/Column Sampling]
+    E --> F
+    
+    F --> G[Get Perturbation Columns]
+    G --> H[Set Outcome Variable]
+    H --> I[Handle Correlation Matrix]
+    
+    I --> J[Handle Percent Missing]
+    J --> K[Handle Outcome List]
+    K --> L[Remove Constant Columns]
+    
+    L --> M{Final Column List Empty?}
+    M -->|Yes| N[Safety Retention Mechanism]
+    M -->|No| O[Proceed with Final Columns]
+    
+    N --> P[Retain Protected Columns]
+    P --> Q{Still Empty?}
+    Q -->|Yes| R[Select Random Features]
+    Q -->|No| O
+    R --> O
+    
+    O --> S[Clean Up Data]
+    S --> T[Screen Non-Float Types]
+    T --> U[Handle Column Names]
+    
+    U --> V{Scale Data?}
+    V -->|Yes| W[Apply Standard Scaling]
+    V -->|No| X[Skip Scaling]
+    W --> X
+    
+    X --> Y{Time Series Mode?}
+    Y -->|Yes| Z[Convert to Time Series]
+    Y -->|No| AA[Regular Processing]
+    
+    Z --> BB[Get Max Sequence Length]
+    BB --> CC[Convert X,y to Time Series]
+    CC --> DD[Train/Test Split]
+    
+    AA --> DD
+    DD --> EE[Remove Constant Columns After Split]
+    
+    EE --> FF{Feature Reduction Needed?}
+    FF -->|Yes| GG[Apply Feature Importance Methods]
+    FF -->|No| HH[Skip Feature Reduction]
+    
+    GG --> II{Features Remain?}
+    II -->|No| JJ[Error: All Features Removed]
+    II -->|Yes| HH
+    
+    HH --> KK{Time Series Mode?}
+    KK -->|Yes| LL[Get Time Series Model List]
+    KK -->|No| MM[Get Regular Model List]
+    
+    LL --> NN[Pipeline Complete]
+    MM --> NN
+    
+    JJ --> OO[Pipeline Failed]
+    
+    style A fill:#e1f5fe
+    style NN fill:#c8e6c9
+    style OO fill:#ffcdd2
+    style N fill:#fff3e0
+    style P fill:#fff3e0
+    style R fill:#fff3e0

assets/data_pipeline.mmd

@@ -0,0 +1,68 @@
+graph TD
+    A[Grid Class Initialization] --> B[Set Global Parameters]
+    B --> C[Set sample_n default=1000]
+    C --> D[Initialize Grid Dictionary]
+    
+    D --> E[Grid Parameters]
+    E --> E1[resample: undersample/oversample/None]
+    E --> E2[scale: True/False]
+    E --> E3[feature_n: 100,95,75,50,25,5]
+    E --> E4[param_space_size: medium/xsmall]
+    E --> E5[n_unique_out: 10]
+    E --> E6[outcome_var_n: 1]
+    E --> E7[percent_missing: 99,95,80]
+    E --> E8[corr: 0.98,0.85,0.5,0.25]
+    E --> E9[feature_selection_method: anova/markov_blanket]
+    E --> E10[data: nested dictionary]
+    
+    E10 --> F[Data Features]
+    F --> F1[age: True/False]
+    F --> F2[sex: True/False]
+    F --> F3[bmi: True]
+    F --> F4[ethnicity: True/False]
+    F --> F5[bloods: True/False]
+    F --> F6[diagnostic_order: True/False]
+    F --> F7[drug_order: True/False]
+    F --> F8[annotation_n: True/False]
+    F --> F9[meta_sp_annotation_n: True/False]
+    F --> F10[annotation_mrc_n: True/False]
+    F --> F11[meta_sp_annotation_mrc_n: True/False]
+    F --> F12[vte_status: True]
+    F --> F13[hosp_site: True]
+    F --> F14[Other features: False]
+    
+    D --> G[c_prod Function Definition]
+    G --> H[Generate Cartesian Product]
+    H --> I[Create settings_list]
+    I --> J[Print Full Size]
+    J --> K[Shuffle List]
+    K --> L[Sample n Items]
+    L --> M[Create Iterator]
+    
+    subgraph "c_prod Function Logic"
+        N[Input: Dictionary or List]
+        N --> O{Is List?}
+        O -->|Yes| P[Yield Items Recursively]
+        O -->|No| Q[Generate Product of Values]
+        Q --> R[Yield Dictionary with Keys]
+    end
+    
+    subgraph "Key Dependencies"
+        S[itertools as it]
+        T[random]
+        U[ml_grid.util.global_params]
+    end
+    
+    subgraph "Class Attributes"
+        V[self.global_params]
+        W[self.verbose]
+        X[self.sample_n]
+        Y[self.grid]
+        Z[self.settings_list]
+        AA[self.settings_list_iterator]
+    end
+    
+    style A fill:#e1f5fe
+    style D fill:#f3e5f5
+    style G fill:#fff3e0
+    style M fill:#e8f5e8

assets/data_pipeline.svg

@@ -0,0 +1,59 @@
+graph TD
+    A[HyperparameterSearch.__init__] --> B{Validate ml_grid_object}
+    B -->|None| C[Raise ValueError]
+    B -->|Valid| D{Check Algorithm Type}
+    
+    D -->|Module| E[Raise ValueError: Expected classifier instance]
+    D -->|Valid Classifier| F[Algorithm Validation]
+    
+    F --> G{Check Custom Classifiers}
+    G -->|KNNWrapper<br/>h2o_classifier_class<br/>kerasClassifier_class| H[Valid Custom Classifier]
+    G -->|Other| I{Check sklearn classifier}
+    I -->|is_classifier = True| J[Valid sklearn Classifier]
+    I -->|False| K{Has fit & predict methods?}
+    K -->|Yes| L[Valid Generic Classifier]
+    K -->|No| M[Raise ValueError: Invalid classifier]
+    
+    H --> N[Configure Warnings]
+    J --> N
+    L --> N
+    
+    N --> O{Method contains keras/xgb/catboost?}
+    O -->|Yes| P[Configure GPU]
+    O -->|No| Q[Skip GPU Config]
+    
+    P --> R[run_search Method]
+    Q --> R
+    
+    R --> S{Check bayessearch flag}
+    S -->|True| T{Algorithm is KNNWrapper or kerasClassifier?}
+    S -->|False| U[Validate Parameters]
+    
+    T -->|Yes| V[Set grid_n_jobs = 1]
+    T -->|No| W[Use default grid_n_jobs]
+    
+    V --> X[Use parameter_space as-is]
+    W --> X
+    U --> Y[Get validated parameters]
+    
+    X --> Z{bayessearch = True?}
+    Y --> AA{random_search = True?}
+    
+    Z -->|True| BB[BayesSearchCV]
+    AA -->|True| CC[Calculate n_iter<br/>min max_iter, parameter_grid_size * sub_sample_pct / 100]
+    AA -->|False| DD[GridSearchCV]
+    
+    CC --> EE[RandomizedSearchCV]
+    
+    BB --> FF[grid.fit X_train, y_train]
+    EE --> FF
+    DD --> FF
+    
+    FF --> GG[Return grid.best_estimator_]
+    
+    style A fill:#e1f5fe
+    style R fill:#e8f5e8
+    style GG fill:#fff3e0
+    style C fill:#ffebee
+    style E fill:#ffebee
+    style M fill:#ffebee

assets/grid_param_space.mmd

@@ -0,0 +1,38 @@
+flowchart TD
+    A[Input DataFrame] --> B[Drop Completely Empty Columns]
+    B --> C[Separate Features X and Target y]
+    C --> D[Train-Test Split]
+    D --> E[Train-Validation Split]
+    
+    E --> F[Identify Column Types]
+    F --> G{Are there Numeric Columns?}
+    
+    G -->|Yes| H[Process Each Numeric Column]
+    G -->|No| M[Skip Numeric Imputation]
+    
+    H --> I{Is Column Completely Empty?}
+    I -->|Yes| J[Fill with 0]
+    I -->|No| K[Apply Mean Imputation]
+    
+    J --> L[Continue to Next Column]
+    K --> L
+    L --> N{More Numeric Columns?}
+    N -->|Yes| H
+    N -->|No| M
+    
+    M --> O[Combine Train/Val/Test Sets]
+    O --> P[Reset Indices]
+    P --> Q[Combine Features and Target]
+    Q --> R[Verify No NaN Values]
+    R --> S[Return Final DataFrame]
+    
+    T[save_missing_percentage Function] --> U[Calculate Missing Percentages]
+    U --> V[Convert to Dictionary]
+    V --> W[Save to Pickle File]
+    W --> X[Return Dictionary]
+    
+    style A fill:#e1f5fe
+    style S fill:#c8e6c9
+    style T fill:#fff3e0
+    style X fill:#ffecb3
+    style R fill:#ffcdd2