modules/dictionary.py

import os
import pandas as pd
from rapidfuzz import process, fuzz

# ============================================================
# DATA LOCATION
# ============================================================

BASE_DIR = os.path.dirname(
    os.path.dirname(
        os.path.abspath(__file__)
    )
)

DICTIONARY_FILE = os.path.join(
    BASE_DIR,
    "data",
    "kiembu_dictionary.csv"
)

# ============================================================
# LOAD DICTIONARY
# ============================================================

def load_dictionary():

    try:

        # First attempt: tab-separated
        df = pd.read_csv(
            DICTIONARY_FILE,
            sep="\t",
            encoding="utf-8"
        )

        # If only one column was loaded,
        # try comma-separated instead
        if len(df.columns) < 2:

            df = pd.read_csv(
                DICTIONARY_FILE,
                sep=",",
                encoding="utf-8"
            )

        df.columns = (
            df.columns
            .str.strip()
        )

        df.fillna("", inplace=True)

        print("Dictionary loaded successfully")
        print("Columns:", df.columns.tolist())
        print("Records:", len(df))

        return df

    except Exception as e:

        print(
            f"Dictionary loading error: {e}"
        )

        return pd.DataFrame()

# ============================================================
# GLOBAL DATAFRAME
# ============================================================

dictionary_df = load_dictionary()

# ============================================================
# TRANSLATION FUNCTION
# ============================================================

def translate_word(
    word,
    direction="English → Kiembu",
    df=None
):

    if df is None:
        df = dictionary_df

    if df.empty:
        return "Dictionary data unavailable."

    if not word:
        return "Please enter a word."

    query = word.strip()

    if not query:
        return "Please enter a word."

    query_lower = query.lower()

    # ========================================================
    # DETERMINE DIRECTION
    # ========================================================

    if direction == "English → Kiembu":

        source_col = "English"
        target_col = "Kiembu"

    else:

        source_col = "Kiembu"
        target_col = "English"

    # ========================================================
    # CHECK COLUMNS
    # ========================================================

    if source_col not in df.columns:

        return (
            f"Column '{source_col}' "
            f"not found in dictionary."
        )

    if target_col not in df.columns:

        return (
            f"Column '{target_col}' "
            f"not found in dictionary."
        )

    # ========================================================
    # EXACT MATCH
    # ========================================================

    exact_match = df[
        df[source_col]
        .astype(str)
        .str.strip()
        .str.lower()
        == query_lower
    ]

    if not exact_match.empty:

        source_word = str(
            exact_match.iloc[0][source_col]
        ).strip()

        translated_word = str(
            exact_match.iloc[0][target_col]
        ).strip()

        return (
            f"{source_word}\n\n"
            f"➜ {translated_word}"
        )

    # ========================================================
    # FUZZY MATCH
    # ========================================================

    choices = (
        df[source_col]
        .astype(str)
        .str.strip()
        .tolist()
    )

    best_match = process.extractOne(
        query,
        choices,
        scorer=fuzz.WRatio
    )

    if not best_match:

        return "No translation found."

    matched_word = best_match[0]
    score = best_match[1]

    if score < 75:

        return (
            "No close translation found.\n\n"
            f"Best score: {score}"
        )

    row = df[
        df[source_col]
        .astype(str)
        .str.strip()
        == matched_word
    ]

    if row.empty:

        return "No translation found."

    translated_word = str(
        row.iloc[0][target_col]
    ).strip()

    return (
        f"Suggested Match: {matched_word}\n\n"
        f"➜ {translated_word}"
    )

# ============================================================
# OPTIONAL TEST
# ============================================================

if __name__ == "__main__":

    print(
        translate_word(
            "house",
            "English → Kiembu"
        )
    )

    print(
        translate_word(
            "Nyomba",
            "Kiembu → English"
        )
    )