import gradio as gr
import joblib
import pandas as pd

# Load model and mappings
model = joblib.load('mushroom_classifier.pkl')
mappings = joblib.load('mappings.pkl')

feature_options = {
    'cap-shape': {'b': 'bell', 'c': 'conical', 'x': 'convex', 'f': 'flat', 'k': 'knobbed', 's': 'sunken'},
    'cap-surface': {'f': 'fibrous', 'g': 'grooves', 'y': 'scaly', 's': 'smooth'},
    'cap-color': {'n': 'brown', 'b': 'buff', 'c': 'cinnamon', 'g': 'gray', 'r': 'green', 'p': 'pink', 'u': 'purple', 'e': 'red', 'w': 'white', 'y': 'yellow'},
    'bruises': {'t': 'bruises', 'f': 'no'},
    'odor': {'a': 'almond', 'l': 'anise', 'c': 'creosote', 'y': 'fishy', 'f': 'foul', 'm': 'musty', 'n': 'none', 'p': 'pungent', 's': 'spicy'},
    'gill-attachment': {'a': 'attached', 'd': 'descending', 'f': 'free', 'n': 'notched'},
    'gill-spacing': {'c': 'close', 'w': 'crowded', 'd': 'distant'},
    'gill-size': {'b': 'broad', 'n': 'narrow'},
    'gill-color': {'k': 'black', 'n': 'brown', 'b': 'buff', 'h': 'chocolate', 'g': 'gray', 'r': 'green', 'o': 'orange', 'p': 'pink', 'u': 'purple', 'e': 'red', 'w': 'white', 'y': 'yellow'},
    'stalk-shape': {'e': 'enlarging', 't': 'tapering'},
    'stalk-root': {'b': 'bulbous', 'c': 'club', 'u': 'cup', 'e': 'equal', 'z': 'rhizomorphs', 'r': 'rooted', '?': 'missing'},
    'stalk-surface-above-ring': {'f': 'fibrous', 'y': 'scaly', 'k': 'silky', 's': 'smooth'},
    'stalk-surface-below-ring': {'f': 'fibrous', 'y': 'scaly', 'k': 'silky', 's': 'smooth'},
    'stalk-color-above-ring': {'n': 'brown', 'b': 'buff', 'c': 'cinnamon', 'g': 'gray', 'o': 'orange', 'p': 'pink', 'e': 'red', 'w': 'white', 'y': 'yellow'},
    'stalk-color-below-ring': {'n': 'brown', 'b': 'buff', 'c': 'cinnamon', 'g': 'gray', 'o': 'orange', 'p': 'pink', 'e': 'red', 'w': 'white', 'y': 'yellow'},
    'veil-type': {'p': 'partial', 'u': 'universal'},
    'veil-color': {'n': 'brown', 'o': 'orange', 'w': 'white', 'y': 'yellow'},
    'ring-number': {'n': 'none', 'o': 'one', 't': 'two'},
    'ring-type': {'c': 'cobwebby', 'e': 'evanescent', 'f': 'flaring', 'l': 'large', 'n': 'none', 'p': 'pendant', 's': 'sheathing', 'z': 'zone'},
    'spore-print-color': {'k': 'black', 'n': 'brown', 'b': 'buff', 'h': 'chocolate', 'r': 'green', 'o': 'orange', 'u': 'purple', 'w': 'white', 'y': 'yellow'},
    'population': {'a': 'abundant', 'c': 'clustered', 'n': 'numerous', 's': 'scattered', 'v': 'several', 'y': 'solitary'},
    'habitat': {'g': 'grasses', 'l': 'leaves', 'm': 'meadows', 'p': 'paths', 'u': 'urban', 'w': 'waste', 'd': 'woods'}
}

# def predict_mushroom(features):
#     numerical_features = {feature: feature_options[feature][value] for feature, value in features.items()}
#     input_df = pd.DataFrame([numerical_features])
#     prediction = model.predict(input_df)
#     return 'Poisonous' if prediction[0] == 1 else 'Edible'

def predict_mushroom(*inputs):
    # Build a dictionary of feature: full name selections.
    # The keys are the feature names (e.g., 'cap-shape') and the values are the full names (e.g., 'bell')
    user_full = dict(zip(feature_options.keys(), inputs))
    
    # Convert the full names to letters using feature_options (reverse mapping)
    user_letters = {}
    for feature, full_value in user_full.items():
        # Create reverse mapping: full name -> letter for this feature
        inverse = {v: k for k, v in feature_options[feature].items()}
        if full_value in inverse:
            user_letters[feature] = inverse[full_value]
        else:
            raise ValueError(f"Invalid selection for {feature}: {full_value}")
    
    # Convert letters to numeric values using the mappings from mappings.pkl
    numeric_features = {}
    for feature, letter in user_letters.items():
        if feature in mappings and letter in mappings[feature]:
            numeric_features[feature] = mappings[feature][letter]
        else:
            raise ValueError(f"Mapping not found for feature {feature} with letter {letter}")
    
    # Create a DataFrame for model input
    input_df = pd.DataFrame([numeric_features])
    
    # Load the trained model (assumes the file is in the working directory)
    model = joblib.load('mushroom_classifier.pkl')
    
    # Predict using the model
    prediction = model.predict(input_df)
    
    # Return the human-readable result
    return 'Edible' if prediction[0] == 0 else 'Poisonous'

demo = gr.Interface(
    fn=predict_mushroom,
    inputs=[gr.Dropdown(choices=list(options.values()), label=feature) for feature, options in feature_options.items()],
    outputs="text",
    title="MycoNom - Mushroom Edibility Classifier",
    description="Select the mushroom features to determine if it's edible or poisonous.<br><br>You can train your own version of this model by heading to OPEN-ARC: https://github.com/Infinitode/OPEN-ARC.<br><br>**Disclaimer:** This model is for **educational purposes only** and should not be used for real-life mushroom classification or any decision-making processes related to the consumption of mushrooms. While the model performs well on the provided dataset, it has not been thoroughly validated for real-world scenarios and may not accurately detect poisonous mushrooms in all conditions. Always consult an expert or use trusted resources when identifying mushrooms."
)

demo.launch()