""" sentiment_page.py ================= Halaman Analisis Sentimen — diperbaiki untuk mengatasi semua prediksi negatif. PERBAIKAN vs VERSI LAMA: ───────────────────────────────────────────────────────────────────────────── 1. MENGGUNAKAN sentiment_service.py (HYBRID CLASSIFIER) Tidak lagi memanggil model secara langsung dengan predict_batch sederhana. Sebaliknya menggunakan sentiment_service yang sudah mengimplementasikan: • Lexicon-based override untuk kelas POSITIF (yang tidak ada di model) • Negation handling (tidak bagus ≠ bagus) • Fallback ke model NB untuk negatif vs netral 2. PREPROCESSING SELARAS Memanggil preprocess_for_model() dari sentiment_service agar preprocessing yang digunakan untuk prediksi persis sama dengan yang dipakai saat training. 3. CONFIDENCE SCORE AKURAT Menggunakan confidence gabungan dari lexicon score + model probability, bukan hanya max probability dari model yang bias ke negatif. 4. LABEL MAPPING KONSISTEN Label dikembalikan selalu dalam format: 'Positif', 'Netral', 'Negatif' (kapital huruf pertama) agar konsisten di seluruh tampilan. ───────────────────────────────────────────────────────────────────────────── """ import streamlit as st import pandas as pd import numpy as np import os import re import string from collections import Counter from datetime import datetime, timedelta from html import escape from database import engine, get_tweet_count, get_latest_crawl_time from timezone_utils import ( parse_dt_with_tz, parse_dt_with_source_tz, get_timezone_label, get_timezone_name, ) import plotly.graph_objects as go # ── Import sentiment_service (hybrid classifier) ────────────────────────── from sentiment_service import ( preprocess_for_model, preprocess_untuk_lexicon, _hitung_skor_lexicon, _klasifikasi_hybrid, _load_stopwords, _load_stemmer, _STOPWORDS, _STEMMER, ) from page_modules.table_utils import render_standard_table # ───────────────────────────────────────────────────────────── # Timezone & formatting helpers # ───────────────────────────────────────────────────────────── def parse_dt(series): return parse_dt_with_tz(series, st.session_state.get("user_timezone", "WIB (UTC+7)")) def parse_crawled_dt(series): return parse_dt_with_source_tz( series, st.session_state.get("user_timezone", "WIB (UTC+7)"), os.getenv("APP_TIMEZONE", "Asia/Makassar"), ) def format_dt(value): if value is None or pd.isna(value): return "Belum ada" try: tz_label = get_timezone_label(st.session_state.get("user_timezone", "WIB (UTC+7)")) return f"{value.strftime('%d/%m/%Y %H:%M')} {tz_label}" except Exception: return "Belum ada" def format_now(): now = parse_dt(pd.Series([datetime.utcnow().isoformat()])).iloc[0] return format_dt(now) def user_today(): timezone_choice = st.session_state.get("user_timezone", "WIB (UTC+7)") return pd.Timestamp.now(tz=get_timezone_name(timezone_choice)).date() def _sync_dynamic_period(): mode = st.session_state.get("analysis_mode") today = user_today() configs = { "realtime": (today - timedelta(days=6), today, "Tweet Terkini — 7 Hari Terakhir"), "30days": (today - timedelta(days=29), today, "30 Hari Terakhir"), "captured": (today, today, "Tweet Hari Ini"), } if mode not in configs: return start_day, end_day, mode_display = configs[mode] start_day = pd.Timestamp(start_day).date() end_day = pd.Timestamp(end_day).date() dt_start = datetime.combine(start_day, datetime.min.time()) dt_end = datetime.combine(end_day + timedelta(days=1), datetime.min.time()) st.session_state.filter_start_date = dt_start st.session_state.filter_end_date = dt_end st.session_state.filter_label = f"{dt_start.strftime('%d/%m/%Y')} s/d {end_day.strftime('%d/%m/%Y')}" st.session_state.mode_display = mode_display st.session_state.filter_date_column = "created_at" # ───────────────────────────────────────────────────────────── # Core prediction function (DIPERBAIKI) # ───────────────────────────────────────────────────────────── def predict_batch_hybrid(texts: list[str]) -> list[tuple[str, float]]: """ Prediksi sentimen menggunakan Hybrid Classifier dari sentiment_service. Pipeline per tweet: 1. preprocess_for_model() → teks untuk TF-IDF + NB 2. preprocess_untuk_lexicon() → teks untuk pengecekan lexicon 3. _hitung_skor_lexicon() → hitung sinyal positif/negatif kuat 4. _klasifikasi_hybrid() → putuskan label + confidence Return: list of (label, confidence) """ results = [] for text in texts: teks_model = preprocess_for_model(text) teks_lexicon = preprocess_untuk_lexicon(text) teks_lower = str(text).lower() skor = _hitung_skor_lexicon(teks_lexicon) label, conf = _klasifikasi_hybrid(teks_model, skor, teks_lower) results.append((label, conf)) return results def preprocess_single(text: str) -> str: """Preprocess satu teks untuk disimpan ke kolom clean_text.""" return preprocess_for_model(text) # ───────────────────────────────────────────────────────────── # Shared UI helpers # ───────────────────────────────────────────────────────────── def _section_header(title, subtitle=""): sub_html = ( f'
{subtitle}
' if subtitle else "" ) st.markdown( f'
' f'
{title}
' f'{sub_html}
', unsafe_allow_html=True, ) def _section_gap(size="md"): heights = {"sm": "1rem", "md": "1.45rem", "lg": "1.9rem"} st.markdown(f'
', unsafe_allow_html=True) def _render_sentiment_styles(): st.markdown(""" """, unsafe_allow_html=True) # ───────────────────────────────────────────────────────────── # Page Header # ───────────────────────────────────────────────────────────── def _render_page_header(): st.markdown("""
📈

Analisis Sentimen

Klasifikasi otomatis tweet menggunakan Hybrid Classifier — Positif · Netral · Negatif

🤖 Hybrid Classifier
""", unsafe_allow_html=True) # ───────────────────────────────────────────────────────────── # Summary Pills # ───────────────────────────────────────────────────────────── def _render_summary_pills(total, pos_n, neu_n, neg_n, pos_p, neu_p, neg_p, filter_label): pills = [ ("stat-1", "📊", "linear-gradient(135deg,#eef2ff,#e0e7ff)", "#3b6cf7", "#1e3a8a", "#c7d2fe", "Total Dianalisis", f"{total:,}", f"Periode {filter_label}"), ("stat-2", "😊", "linear-gradient(135deg,#f0fdf4,#dcfce7)", "#16a34a", "#14532d", "#86efac", "Positif", f"{pos_n:,}", f"{pos_p:.1f}% dari total"), ("stat-3", "😐", "linear-gradient(135deg,#f8fafc,#f1f5f9)", "#64748b", "#334155", "#cbd5e1", "Netral", f"{neu_n:,}", f"{neu_p:.1f}% dari total"), ("stat-4", "😞", "linear-gradient(135deg,#fef2f2,#fee2e2)", "#ef4444", "#7f1d1d", "#fca5a5", "Negatif", f"{neg_n:,}", f"{neg_p:.1f}% dari total"), ] cols = st.columns(4, gap="medium") for col, (anim, icon, bg, color, dark, border, label, val, sub) in zip(cols, pills): fs = "1.6rem" if len(str(val)) <= 6 else "1.2rem" with col: st.markdown( f'
' f'
{icon}
' f'
{label}
' f'
{val}
' f'
{sub}
' f'
', unsafe_allow_html=True, ) # ───────────────────────────────────────────────────────────── # Proportion Bar # ───────────────────────────────────────────────────────────── def _render_proportion_bar(pos_p, neu_p, neg_p): pos_p = round(pos_p, 1) neu_p = round(neu_p, 1) neg_p = round(neg_p, 1) st.markdown( f'
' f'
Proporsi Sentimen Keseluruhan
' f'
' f'
' f'{"😊 " + str(pos_p) + "%" if pos_p >= 8 else ""}
' f'
' f'{"😐 " + str(neu_p) + "%" if neu_p >= 8 else ""}
' f'
' f'{"😞 " + str(neg_p) + "%" if neg_p >= 8 else ""}
' f'
' f'
' f'● Positif {pos_p}%' f'● Netral {neu_p}%' f'● Negatif {neg_p}%' f'
', unsafe_allow_html=True, ) # ───────────────────────────────────────────────────────────── # Donut + Bar chart # ───────────────────────────────────────────────────────────── def _render_donut_chart(pos_n, neu_n, neg_n, total, filter_label): labels, values, colors = [], [], [] color_map = {"Positif": "#16a34a", "Netral": "#94a3b8", "Negatif": "#ef4444"} for label, val in [("Positif", pos_n), ("Netral", neu_n), ("Negatif", neg_n)]: if val > 0: labels.append(label) values.append(val) colors.append(color_map[label]) dominant = max([("Positif", pos_n), ("Netral", neu_n), ("Negatif", neg_n)], key=lambda x: x[1]) dom_pct = round(dominant[1] / total * 100) if total > 0 else 0 dom_emoji = {"Positif": "😊", "Netral": "😐", "Negatif": "😞"}.get(dominant[0], "📊") fig = go.Figure(data=[go.Pie( labels=labels, values=values, hole=0.62, marker=dict(colors=colors, line=dict(color="white", width=4)), textinfo="label+percent", textfont=dict(size=12), hovertemplate="%{label}
%{value:,} tweet — %{percent}", direction="clockwise", sort=False, )]) fig.add_annotation( text=f"{dom_pct}%
{dom_emoji} {dominant[0]}", x=0.5, y=0.5, showarrow=False, align="center", font=dict(size=18, color="#0f172a"), ) fig.update_layout( height=290, margin=dict(l=10, r=10, t=10, b=10), showlegend=True, legend=dict(orientation="h", y=-0.1, x=0.5, xanchor="center", font=dict(size=11)), paper_bgcolor="rgba(0,0,0,0)", ) st.plotly_chart(fig, width="stretch", config={"displayModeBar": False}) return dominant def _render_bar_chart(pos_n, neu_n, neg_n, total): categories = ["Positif 😊", "Netral 😐", "Negatif 😞"] values = [pos_n, neu_n, neg_n] colors = ["#16a34a", "#94a3b8", "#ef4444"] percentages = [(v / total * 100) if total > 0 else 0 for v in values] fig = go.Figure() for cat, val, color, pct in zip(categories, values, colors, percentages): fig.add_trace(go.Bar( x=[cat], y=[val], marker=dict(color=color, opacity=0.88, cornerradius=8), text=[f"{val:,}"], textposition="outside", textfont=dict(size=13, color="#0f172a"), width=0.5, hovertemplate=f"{cat}
{val:,} tweet ({pct:.1f}%)", )) fig.update_layout( height=290, margin=dict(l=0, r=0, t=30, b=0), paper_bgcolor="rgba(0,0,0,0)", plot_bgcolor="rgba(0,0,0,0)", showlegend=False, xaxis=dict(showgrid=False, tickfont=dict(size=11, color="#64748b")), yaxis=dict(showgrid=True, gridcolor="rgba(226,232,240,0.8)", griddash="dot", tickfont=dict(size=10, color="#94a3b8")), ) st.plotly_chart(fig, width="stretch", config={"displayModeBar": False}) # ───────────────────────────────────────────────────────────── # Trend chart # ───────────────────────────────────────────────────────────── def _render_trend_chart(fdf, filter_label, dt_start, dt_end): _section_header( "📅 Tren Sentimen dari Hari ke Hari", f"Periode {filter_label} · Diupdate: {format_now()}" ) df_tl = fdf.copy() df_tl["date"] = pd.to_datetime(df_tl["created_at"], errors="coerce").dt.date actual_tl = df_tl.groupby(["date", "sentiment"]).size().reset_index(name="count") date_range = pd.date_range(pd.Timestamp(dt_start).date(), pd.Timestamp(dt_end).date(), freq="D") base_dates = pd.DataFrame({"date": date_range.date}) fig = go.Figure() config_lines = [ ("Positif", "#16a34a", "rgba(22,163,74,0.08)"), ("Netral", "#94a3b8", "rgba(148,163,184,0.06)"), ("Negatif", "#ef4444", "rgba(239,68,68,0.08)"), ] for sent, color, fill in config_lines: data = base_dates.merge( actual_tl[actual_tl["sentiment"] == sent][["date", "count"]], on="date", how="left", ) data["count"] = data["count"].fillna(0).astype(int) fig.add_trace(go.Scatter( x=data["date"], y=data["count"], name=sent, mode="lines+markers", line=dict(color=color, width=2.5, shape="spline", smoothing=0.8), marker=dict(size=6, color="white", line=dict(color=color, width=2.5)), fill="tozeroy", fillcolor=fill, hovertemplate=f"{sent}
%{{x}}: %{{y}} tweet", )) fig.update_layout( height=290, margin=dict(l=0, r=0, t=10, b=0), paper_bgcolor="rgba(0,0,0,0)", plot_bgcolor="rgba(0,0,0,0)", legend=dict(orientation="h", y=1.1, x=0, font=dict(size=11)), xaxis=dict(showgrid=True, gridcolor="rgba(226,232,240,0.6)", tickfont=dict(size=10, color="#94a3b8")), yaxis=dict(showgrid=True, gridcolor="rgba(226,232,240,0.6)", griddash="dot", tickfont=dict(size=10, color="#94a3b8")), hovermode="x unified", ) with st.container(border=True, key="sentiment_trend_panel"): st.plotly_chart(fig, width="stretch", config={"displayModeBar": False}) # ───────────────────────────────────────────────────────────── # Confidence histogram # ───────────────────────────────────────────────────────────── # def _render_confidence_chart(fdf): # _section_header( # "🎯 Distribusi Keyakinan Model", # "Seberapa yakin classifier dalam mengklasifikasikan setiap tweet" # ) # fig = go.Figure() # sent_cfg = [ # ("Positif", "#16a34a", "rgba(22,163,74,0.7)"), # ("Netral", "#94a3b8", "rgba(148,163,184,0.7)"), # ("Negatif", "#ef4444", "rgba(239,68,68,0.7)"), # ] # for sent, color, fill_color in sent_cfg: # sub = fdf[fdf["sentiment"] == sent]["confidence"] # if sub.empty: # continue # fig.add_trace(go.Histogram( # x=sub, name=sent, nbinsx=20, # marker=dict(color=fill_color, line=dict(color=color, width=1)), # opacity=0.85, # hovertemplate=f"{sent}
Keyakinan: %{{x:.0%}}
Jumlah: %{{y}}", # )) # fig.update_layout( # height=260, margin=dict(l=0, r=0, t=10, b=0), barmode="overlay", # paper_bgcolor="rgba(0,0,0,0)", plot_bgcolor="rgba(0,0,0,0)", # legend=dict(orientation="h", y=1.1, x=0, font=dict(size=11)), # xaxis=dict(tickformat=".0%", title="Tingkat Keyakinan", # tickfont=dict(size=10, color="#94a3b8"), # showgrid=True, gridcolor="rgba(226,232,240,0.6)"), # yaxis=dict(title="Jumlah Tweet", # tickfont=dict(size=10, color="#94a3b8"), # showgrid=True, gridcolor="rgba(226,232,240,0.6)", griddash="dot"), # ) # with st.container(border=True, key="sentiment_conf_panel"): # st.plotly_chart(fig, width="stretch", config={"displayModeBar": False}) # ───────────────────────────────────────────────────────────── # Word frequency per sentimen # ───────────────────────────────────────────────────────────── def _render_word_freq_per_sentiment(fdf): _section_header( "📊 Kata Dominan per Sentimen", "Top 15 kata paling sering muncul di masing-masing kelompok sentimen" ) stop_extra = {"ongkos", "kirim", "gratis", "komdigi", "ongkir"} sent_cfg = [ ("Positif", "#16a34a", "#1e3a8a"), ("Netral", "#64748b", "#334155"), ("Negatif", "#ef4444", "#7f1d1d"), ] cols = st.columns(3, gap="medium") for col, (sent, color, dark) in zip(cols, sent_cfg): with col: sub = fdf[fdf["sentiment"] == sent] all_words = " ".join(sub["clean_text"].fillna("")).split() filtered = [w for w in all_words if len(w) > 2 and w not in stop_extra] wf = Counter(filtered).most_common(15) st.markdown( f'
' f'
' f'
' f'{"😊" if sent=="Positif" else "😐" if sent=="Netral" else "😞"}
' f'
{sent}
' f'
' f'{len(sub):,} tweet
' f'
', unsafe_allow_html=True, ) if not wf: st.info("Belum ada data") st.markdown('
', unsafe_allow_html=True) continue words_list = [w[0] for w in wf] counts_list = [w[1] for w in wf] max_c = max(counts_list) if counts_list else 1 fig = go.Figure(data=[go.Bar( y=words_list[::-1], x=counts_list[::-1], orientation="h", marker=dict( color=[f"rgba({int(color[1:3],16)},{int(color[3:5],16)},{int(color[5:7],16)},{0.35+0.65*(c/max_c):.2f})" for c in counts_list[::-1]], line=dict(width=0), cornerradius=4, ), text=[str(c) for c in counts_list[::-1]], textposition="outside", textfont=dict(size=9, color="#475569"), hovertemplate="%{y}
%{x} kali", )]) fig.update_layout( height=360, margin=dict(l=0, r=40, t=4, b=4), paper_bgcolor="rgba(0,0,0,0)", plot_bgcolor="rgba(0,0,0,0)", showlegend=False, xaxis=dict(showgrid=True, gridcolor="rgba(203,213,225,0.6)", tickfont=dict(size=8, color="#94a3b8"), fixedrange=True), yaxis=dict(showgrid=False, tickfont=dict(size=9, color="#334155"), fixedrange=True), ) st.plotly_chart(fig, width="stretch", config={"displayModeBar": False}) st.markdown('', unsafe_allow_html=True) # ───────────────────────────────────────────────────────────── # Word cloud # ───────────────────────────────────────────────────────────── def _render_wordcloud(fdf): _section_header( "☁️ Word Cloud per Sentimen", "Kata-kata populer dari tweet yang sudah melalui preprocessing" ) try: from wordcloud import WordCloud import matplotlib.pyplot as plt stop_wc = {"ongkos", "kirim", "gratis", "komdigi", "ongkir"} wc1, wc2, wc3 = st.columns(3, gap="medium") cfg = [ (wc1, "Positif", "Greens", "😊 Positif", "#f0fdf4", "#14532d"), (wc2, "Netral", "Blues", "😐 Netral", "#f8fafc", "#334155"), (wc3, "Negatif", "Reds", "😞 Negatif", "#fef2f2", "#7f1d1d"), ] for col, sent, cmap, title, bg, tc in cfg: with col: st.markdown( f'
' f'{title}
', unsafe_allow_html=True, ) sub = fdf[fdf["sentiment"] == sent] words = [w for w in " ".join(sub["clean_text"].fillna("")).split() if len(w) > 2 and w not in stop_wc] if words: wc = WordCloud( width=420, height=260, background_color="white", colormap=cmap, max_words=60, relative_scaling=0.5, collocations=False, ).generate(" ".join(words)) fig, ax = plt.subplots(figsize=(5, 3.1)) ax.imshow(wc, interpolation="bilinear") ax.axis("off") plt.tight_layout(pad=0) st.pyplot(fig, clear_figure=True) else: st.info("Data kata tidak cukup") except ImportError: st.warning("Install `wordcloud` dan `matplotlib` terlebih dahulu.") # ───────────────────────────────────────────────────────────── # Tweet table # ───────────────────────────────────────────────────────────── def _render_tweet_table(fdf, filter_label): _section_header( "📋 Tabel Analisis Sentimen", f"Total {len(fdf):,} tweet · {filter_label}" ) with st.container(border=True, key="sentiment_table_controls"): tf1, tf2, tf3 = st.columns([3, 1.6, 1.7], gap="medium", vertical_alignment="bottom") with tf1: search = st.text_input( "Search tweet", placeholder="Ketik kata kunci di isi tweet...", key="sentiment_table_search", ) with tf2: sf = st.selectbox( "Filter sentimen", ["Semua", "Positif 😊", "Netral 😐", "Negatif 😞"], key="sentiment_table_filter", ) with tf3: sort_by = st.selectbox( "Urutkan", ["Terbaru dulu", "Terlama dulu", "Keyakinan tertinggi"], key="sentiment_table_sort", ) _section_gap("sm") tdf = fdf.copy() if "crawled_at" not in tdf.columns: tdf["crawled_at"] = pd.NaT if search: tdf = tdf[tdf["text"].str.contains(search, case=False, na=False)] sf_map = {"Positif 😊": "Positif", "Netral 😐": "Netral", "Negatif 😞": "Negatif"} if sf != "Semua": tdf = tdf[tdf["sentiment"] == sf_map.get(sf, sf)] if sort_by == "Terbaru dulu": tdf = tdf.sort_values("created_at", ascending=False) elif sort_by == "Terlama dulu": tdf = tdf.sort_values("created_at", ascending=True) else: tdf = tdf.sort_values("confidence", ascending=False) # out = tdf[["created_at", "crawled_at", "text", "clean_text", "sentiment", "confidence"]].copy() out = tdf[["created_at", "crawled_at", "text", "clean_text", "sentiment"]].copy() out["created_at"] = out["created_at"].apply(format_dt) out["crawled_at"] = out["crawled_at"].apply(format_dt) # out["confidence"] = out["confidence"].apply(lambda x: f"{x:.0%}") out["sentiment"] = out["sentiment"].map({ "Positif": "😊 Positif", "Netral": "😐 Netral", "Negatif": "😞 Negatif", }).fillna(out["sentiment"]) # out.columns = ["Tanggal Tweet", "Masuk Database", "Tweet Asli", "Tweet Bersih", "Sentimen", "Keyakinan"] out.columns = ["Tanggal Tweet", "Masuk Database", "Tweet Asli", "Tweet Bersih", "Sentimen"] render_standard_table( out, height=400, min_width=1220, badge_columns=["Sentimen"], nowrap=["Tanggal Tweet", "Masuk Database", "Sentimen", "Keyakinan"], wide_columns=["Tweet Asli", "Tweet Bersih"], column_widths={ "Tanggal Tweet": "170px", "Masuk Database": "170px", "Tweet Asli": "360px", "Tweet Bersih": "360px", "Sentimen": "130px", "Keyakinan": "110px", }, ) st.caption(f"Menampilkan {len(tdf):,} tweet") return tdf # ───────────────────────────────────────────────────────────── # Insight & Recommendation # ───────────────────────────────────────────────────────────── def _render_insight_panel(dominant, pos_n, neu_n, neg_n, total, filter_label): dom_name = dominant[0] dom_pct = dominant[1] / total * 100 if total else 0 neg_pct = neg_n / total * 100 if total else 0 pos_pct = pos_n / total * 100 if total else 0 neu_pct = neu_n / total * 100 if total else 0 _section_header( "💡 Insight & Rekomendasi Tindakan", f"Berdasarkan analisis {total:,} tweet · {filter_label}" ) dom_color = {"Positif": "#16a34a", "Netral": "#64748b", "Negatif": "#ef4444"}.get(dom_name, "#3b6cf7") dom_bg = {"Positif": "#f0fdf4", "Netral": "#f8fafc", "Negatif": "#fef2f2"}.get(dom_name, "#eef2ff") dom_emoji = {"Positif": "😊", "Netral": "😐", "Negatif": "😞"}.get(dom_name, "📊") st.markdown( f'
' f'
' f'
{dom_emoji}
' f'
' f'
Sentimen Dominan
' f'
' f'{dom_name} · {dom_pct:.1f}%
' f'
' f'
' f'Positif: {pos_pct:.1f}% · ' f'Netral: {neu_pct:.1f}% · ' f'Negatif: {neg_pct:.1f}%' f'
', unsafe_allow_html=True, ) rows = [] if neg_pct >= 40: rows.append(("🔴 URGENT", "#fef2f2", "#7f1d1d", "#ef4444", "Tanggapi Keluhan Publik", "Sentimen negatif tinggi — ketidakpuasan signifikan", "Buat klarifikasi resmi dan buka ruang dialog publik", "Humas / Tim Kebijakan")) if neg_pct >= 20: rows.append(("🟠 TINGGI", "#fff7ed", "#7c2d12", "#ea580c", "Tinjau Ulang Kebijakan", f"Sentimen negatif mencapai {neg_pct:.1f}%", "Evaluasi poin kebijakan yang paling banyak dikeluhkan", "Tim Kebijakan")) if neu_pct >= 30: rows.append(("🟡 SEDANG", "#fefce8", "#713f12", "#ca8a04", "Tingkatkan Sosialisasi", f"Sentimen netral {neu_pct:.1f}% — publik belum berpihak", "Perbanyak konten edukatif dan FAQ resmi", "Tim Komunikasi")) if pos_pct >= 20: rows.append(("🟢 INFO", "#f0fdf4", "#14532d", "#16a34a", "Pertahankan Momentum Positif", f"Sentimen positif {pos_pct:.1f}%", "Perkuat narasi positif via kanal resmi secara konsisten", "Tim Media Sosial")) rows.append(("🔵 RUTIN", "#eff6ff", "#1e3a8a", "#3b6cf7", "Pemantauan Berkelanjutan", "Opini publik dapat berubah sewaktu-waktu", "Pantau sentimen harian dan buat laporan berkala", "Tim Analis Data")) df_rek = pd.DataFrame(rows, columns=["Prioritas", "bg", "tc", "bc", "Tindakan", "Dasar Analisis", "Rekomendasi", "Penanggung Jawab"]) for _, row in df_rek.iterrows(): st.markdown( f'
' f'
' f'
' f'{row.Prioritas}
' f'
' f'
' f'{row.Tindakan}
' f'
' f'📌 {row["Dasar Analisis"]}
' f'
' f'✅ {row.Rekomendasi}
' f'
' f'
' f'' f'👤 {row["Penanggung Jawab"]}
' f'
', unsafe_allow_html=True, ) export_df = df_rek[["Prioritas", "Tindakan", "Dasar Analisis", "Rekomendasi", "Penanggung Jawab"]] return export_df # ───────────────────────────────────────────────────────────── # Main show() # ───────────────────────────────────────────────────────────── def show(): _render_sentiment_styles() _render_page_header() if "analysis_mode" not in st.session_state: st.warning("⚠️ Silakan pilih mode tampilan di halaman Ambil Data Twitter terlebih dahulu.") return _sync_dynamic_period() start_date = st.session_state.get("filter_start_date") end_date = st.session_state.get("filter_end_date") filter_label = st.session_state.get("filter_label", "-") mode_display = st.session_state.get("mode_display", "-") if start_date is None or end_date is None: st.warning("⚠️ Silakan buka halaman Ambil Data Twitter terlebih dahulu.") return mode_meta = { "realtime": ("#16a34a", "📡"), "30days": ("#3b6cf7", "📅"), "captured": ("#0284c7", "📆"), "custom": ("#d97706", "🔍"), } mode_color, mode_icon = mode_meta.get(st.session_state.analysis_mode, ("#3b6cf7", "📊")) st.markdown( f'
' f'{mode_icon}' f'
' f'
{mode_display}
' f'
' f'Periode: {filter_label}
' f'
', unsafe_allow_html=True, ) # ── Load data ───────────────────────────────────────────── try: df_all = pd.read_sql("SELECT * FROM tweets ORDER BY created_at DESC", engine) if df_all.empty: st.warning("⚠️ Belum ada data tweet.") return df_all["created_at"] = parse_dt(df_all["created_at"]) if "crawled_at" in df_all.columns: df_all["crawled_at"] = parse_crawled_dt(df_all["crawled_at"]) except Exception as e: st.error(f"❌ Gagal membaca database: {e}") return s_dt = pd.Timestamp(start_date) e_dt = pd.Timestamp(end_date) df = df_all[(df_all["created_at"] >= s_dt) & (df_all["created_at"] < e_dt)].copy() if df.empty: st.warning(f"⚠️ Tidak ada tweet dalam periode {filter_label}.") return # ── Caching ─────────────────────────────────────────────── total_tweets_in_db = get_tweet_count() latest_crawl_marker = get_latest_crawl_time() or "no-crawl" data_marker = (total_tweets_in_db, latest_crawl_marker) cache_key = ( f"sent_{st.session_state.analysis_mode}_" f"{start_date}_{end_date}_{total_tweets_in_db}_{latest_crawl_marker}" ) for old_key in list(st.session_state.keys()): if old_key.startswith("sent_") and old_key != cache_key: del st.session_state[old_key] force_refresh = data_marker != st.session_state.get("_last_sentiment_data_marker") if cache_key not in st.session_state or force_refresh: if force_refresh: st.info("🔄 Menyegarkan prediksi sentimen dengan data terbaru...") with st.spinner("🔍 Preprocessing & prediksi sentimen (Hybrid Classifier)..."): # Preprocess semua tweet df["clean_text"] = df["text"].apply(preprocess_single) dfc = df[df["clean_text"].str.strip().str.len() > 0].copy() # Prediksi menggunakan hybrid classifier results = predict_batch_hybrid(dfc["text"].tolist()) if results: sentiments, confidences = zip(*results) else: sentiments, confidences = [], [] dfc["sentiment"] = list(sentiments) dfc["confidence"] = list(confidences) st.session_state[cache_key] = dfc st.session_state["_last_sentiment_data_marker"] = data_marker df_s = st.session_state[cache_key] # ── Keyword filter ──────────────────────────────────────── _section_header("🎯 Filter Kata Kunci", "Kosongkan untuk melihat semua tweet") with st.container(border=True, key="sentiment_keyword_panel"): kw = st.text_input( "Kata kunci", placeholder="Contoh: ongkir, kurir, komdigi...", key="sentiment_keyword_search", ) st.markdown( f'
' f'' f'Mode {escape(str(mode_display))}' f'' f'Periode {escape(str(filter_label))}' f'
', unsafe_allow_html=True, ) _section_gap("sm") fdf = ( df_s[df_s["text"].str.contains(kw, case=False, na=False)].copy() if kw else df_s.copy() ) if fdf.empty: st.warning("⚠️ Tidak ada tweet yang cocok dengan kata kunci tersebut.") return sc = fdf["sentiment"].value_counts() total = len(fdf) pos_n = int(sc.get("Positif", 0)) neu_n = int(sc.get("Netral", 0)) neg_n = int(sc.get("Negatif", 0)) pos_p = pos_n / total * 100 neu_p = neu_n / total * 100 neg_p = neg_n / total * 100 # ── Ringkasan ───────────────────────────────────────────── _section_header( "📌 Ringkasan Sentimen", f"Berdasarkan tanggal asli tweet · {filter_label}" ) _section_gap("sm") _render_summary_pills(total, pos_n, neu_n, neg_n, pos_p, neu_p, neg_p, filter_label) _section_gap("sm") _render_proportion_bar(pos_p, neu_p, neg_p) _section_gap("lg") # ── Donut + Bar ─────────────────────────────────────────── col_left, col_right = st.columns(2, gap="medium") with col_left: _section_header("🔵 Sebaran Sentimen", f"Periode {filter_label}") with st.container(border=True, key="sentiment_chart_panel"): dominant = _render_donut_chart(pos_n, neu_n, neg_n, total, filter_label) with col_right: _section_header("📊 Perbandingan Jumlah per Sentimen", f"Periode {filter_label}") with st.container(border=True, key="sentiment_bar_panel"): _render_bar_chart(pos_n, neu_n, neg_n, total) _section_gap("lg") # ── Trend ───────────────────────────────────────────────── _render_trend_chart(fdf, filter_label, start_date, end_date) _section_gap("lg") # ── Confidence ──────────────────────────────────────────── # _render_confidence_chart(fdf) # _section_gap("lg") # ── Word freq ───────────────────────────────────────────── _render_word_freq_per_sentiment(fdf) _section_gap("lg") # ── Word cloud ──────────────────────────────────────────── _render_wordcloud(fdf) _section_gap("lg") # ── Tweet table ─────────────────────────────────────────── tdf = _render_tweet_table(fdf, filter_label) _section_gap("lg") # ── Insight ─────────────────────────────────────────────── df_rek = _render_insight_panel(dominant, pos_n, neu_n, neg_n, total, filter_label) _section_gap("lg") # ── Download ────────────────────────────────────────────── _section_header("📥 Unduh Hasil Analisis") with st.container(border=True, key="sentiment_download_panel"): d1, d2, d3 = st.columns(3, gap="medium", vertical_alignment="bottom") with d1: st.download_button( "📥 Semua Hasil Prediksi", fdf.to_csv(index=False).encode("utf-8"), f"hasil_prediksi_{datetime.now().strftime('%Y%m%d_%H%M')}.csv", "text/csv", width="stretch", ) with d2: summary = pd.DataFrame({ "Sentimen": ["Positif", "Netral", "Negatif"], "Jumlah": [pos_n, neu_n, neg_n], "Persen": [f"{pos_p:.2f}%", f"{neu_p:.2f}%", f"{neg_p:.2f}%"], }) st.download_button( "📈 Ringkasan Sentimen", summary.to_csv(index=False).encode("utf-8"), f"ringkasan_{datetime.now().strftime('%Y%m%d_%H%M')}.csv", "text/csv", width="stretch", ) with d3: st.download_button( "🎯 Rekomendasi Tindakan", df_rek.to_csv(index=False).encode("utf-8"), f"rekomendasi_{datetime.now().strftime('%Y%m%d_%H%M')}.csv", "text/csv", width="stretch", )