files/04_topicos/etapa_08_analisar_macrotemas.py

from __future__ import annotations

import sys
from pathlib import Path

ROOT_DIR = Path(__file__).resolve().parents[1]
if str(ROOT_DIR) not in sys.path:
    sys.path.insert(0, str(ROOT_DIR))

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

from pipeline_00_config import (
    MACROTHEME_TEMPORAL_PDF,
    MACROTHEME_TEMPORAL_PNG,
    MACROTHEMES,
    STUDY_END,
    STUDY_START,
    TOPICS_TO_REMOVE,
    TOPIC_ALIGNED_COMMENTS_CSV,
    ensure_directories,
    parse_utc_datetime,
    read_csv_with_fallback,
)


def map_macrotheme(topic_id: int) -> str | None:
    for macrotheme, topic_ids in MACROTHEMES.items():
        if topic_id in topic_ids:
            return macrotheme
    return None


def main() -> None:
    ensure_directories()

    aligned = read_csv_with_fallback(TOPIC_ALIGNED_COMMENTS_CSV, low_memory=False)
    aligned["topic_id"] = pd.to_numeric(aligned["topic_id"], errors="coerce")
    aligned["published_at"] = parse_utc_datetime(aligned["published_at"])
    aligned = aligned.dropna(subset=["topic_id", "published_at"]).copy()
    aligned["topic_id"] = aligned["topic_id"].astype(int)

    excluded_topics = set(TOPICS_TO_REMOVE + [-1])
    aligned = aligned.loc[~aligned["topic_id"].isin(excluded_topics)].copy()
    aligned["macrotheme"] = aligned["topic_id"].apply(map_macrotheme)
    aligned = aligned.dropna(subset=["macrotheme"]).copy()
    aligned["mes_ano"] = aligned["published_at"].dt.to_period("M")

    counts = aligned.groupby(["mes_ano", "macrotheme"]).size().reset_index(name="count")
    pivot = counts.pivot_table(index="mes_ano", columns="macrotheme", values="count", fill_value=0)
    full_range = pd.period_range(start=STUDY_START, end=STUDY_END, freq="M")
    pivot = pivot.reindex(full_range, fill_value=0)
    proportions = pivot.div(pivot.sum(axis=1).replace(0, 1), axis=0)

    plt.figure(figsize=(12, 6))
    for macrotheme in proportions.columns:
        plt.plot(proportions.index.astype(str), proportions[macrotheme], label=macrotheme)

    tick_positions = np.arange(0, len(proportions.index), 6)
    plt.xticks(tick_positions, proportions.index.astype(str)[tick_positions], rotation=45, ha="right")
    plt.title("Evolucao temporal dos macrotemas")
    plt.xlabel("Mes")
    plt.ylabel("Proporcao de comentarios")
    plt.legend(ncol=2, frameon=False)
    plt.tight_layout()
    plt.savefig(MACROTHEME_TEMPORAL_PNG, dpi=300)
    plt.savefig(MACROTHEME_TEMPORAL_PDF)
    plt.close()

    print(f"Grafico salvo em: {MACROTHEME_TEMPORAL_PNG}")


if __name__ == "__main__":
    main()