Build w/ love

requirements.txt

@@ -1,3 +1,4 @@
 altair
 pandas
-streamlit
+streamlit
+requests

src/streamlit_app.py

@@ -1,40 +1,165 @@
-import altair as alt
-import numpy as np
+import time
+import requests
 import pandas as pd
 import streamlit as st
 
-"""
-# Welcome to Streamlit!
-
-Edit `/streamlit_app.py` to customize this app to your heart's desire :heart:.
-If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
-forums](https://discuss.streamlit.io).
-
-In the meantime, below is an example of what you can do with just a few lines of code:
-"""
-
-num_points = st.slider("Number of points in spiral", 1, 10000, 1100)
-num_turns = st.slider("Number of turns in spiral", 1, 300, 31)
-
-indices = np.linspace(0, 1, num_points)
-theta = 2 * np.pi * num_turns * indices
-radius = indices
-
-x = radius * np.cos(theta)
-y = radius * np.sin(theta)
-
-df = pd.DataFrame({
-    "x": x,
-    "y": y,
-    "idx": indices,
-    "rand": np.random.randn(num_points),
-})
-
-st.altair_chart(alt.Chart(df, height=700, width=700)
-    .mark_point(filled=True)
-    .encode(
-        x=alt.X("x", axis=None),
-        y=alt.Y("y", axis=None),
-        color=alt.Color("idx", legend=None, scale=alt.Scale()),
-        size=alt.Size("rand", legend=None, scale=alt.Scale(range=[1, 150])),
-    ))
+API_URL = "https://taic.moda.gov.tw/api/v1/dataset.search.export"
+
+st.set_page_config(page_title="TAIC Pulse", layout="wide")
+
+st.title("臺灣主權AI訓練語料庫 Explorer")
+st.caption("⚡ 即時資料：本頁面會在啟動/手動刷新時從來源 API 抓取最新 JSON，並提供互動式篩選與檢視。（不會背景持續輪詢）")
+
+# ---------------------------
+# Sidebar
+# ---------------------------
+st.sidebar.header("資料設定")
+timeout_sec = st.sidebar.slider("API timeout（秒）", 5, 60, 20)
+st.sidebar.caption("提示：本工具不會持續打 API，只在首次載入或你按下刷新時抓一次。")
+
+# ---------------------------
+# Fetch + cache (one-time per session unless refreshed)
+# ---------------------------
+
+
+@st.cache_data(show_spinner=False)
+def fetch_json_once(timeout: int):
+    r = requests.get(API_URL, timeout=timeout)
+    r.raise_for_status()
+    return r.json()
+
+
+def load_data():
+    with st.spinner("從來源 API 抓取資料中..."):
+        data = fetch_json_once(timeout_sec)
+    fetched_at = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime())
+    return data, fetched_at
+
+
+# Buttons
+colA, colB, colC = st.columns([1, 1, 2])
+with colA:
+    refresh = st.button("🔄 手動刷新（重新抓取）")
+with colB:
+    clear_cache = st.button("🧹 清除快取")
+
+if clear_cache:
+    st.cache_data.clear()
+    st.toast("已清除快取", icon="🧹")
+
+# 用 session_state 記錄「這個 session 已載入過」
+if "payload" not in st.session_state or refresh:
+    if refresh:
+        # 只清掉這個 function 的 cache（簡單作法是全清）
+        st.cache_data.clear()
+    payload, fetched_at = load_data()
+    st.session_state["payload"] = payload
+    st.session_state["fetched_at"] = fetched_at
+
+data = st.session_state["payload"]
+fetched_at = st.session_state["fetched_at"]
+
+with colC:
+    st.metric("最後更新時間", fetched_at)
+
+st.divider()
+
+# ---------------------------
+# Normalize JSON to dataframe
+# ---------------------------
+items = data if isinstance(data, list) else data.get("data", data)
+df = pd.json_normalize(items)
+
+# ---------------------------
+# Dynamic filters
+# ---------------------------
+st.sidebar.header("篩選條件（選單）")
+
+# 優先嘗試常見欄位
+common_names = {"category", "theme", "publisher",
+                "organization", "org", "format", "license", "city"}
+candidate_fields = [c for c in df.columns if c.lower() in common_names]
+
+if not candidate_fields:
+    candidate_fields = st.sidebar.multiselect(
+        "選擇要做成下拉選單的欄位",
+        options=df.columns.tolist(),
+        default=df.columns[:2].tolist() if len(
+            df.columns) >= 2 else df.columns.tolist()
+    )
+
+filters = {}
+for field in candidate_fields:
+    values = sorted(
+        [v for v in df[field].dropna().astype(str).unique().tolist()])
+    if not values:
+        continue
+    choice = st.sidebar.selectbox(f"{field} 篩選", ["(全部)"] + values, index=0)
+    if choice != "(全部)":
+        filters[field] = choice
+
+filtered = df.copy()
+for k, v in filters.items():
+    filtered = filtered[filtered[k].astype(str) == v]
+
+q = st.sidebar.text_input("全文關鍵字（contains）", "")
+if q.strip():
+    mask = filtered.astype(str).apply(
+        lambda row: row.str.contains(q, case=False, na=False)
+    ).any(axis=1)
+    filtered = filtered[mask]
+
+# ---------------------------
+# Preview
+# ---------------------------
+st.subheader("資料預覽")
+st.write(f"共 {len(filtered):,} 筆（原始 {len(df):,} 筆）")
+st.dataframe(filtered, use_container_width=True)
+
+# ---------------------------
+# Download UX (prepare first, then download)
+# ---------------------------
+st.subheader("下載")
+
+# 用 session_state 暫存已生成的檔案 bytes，避免重算
+if "csv_bytes" not in st.session_state:
+    st.session_state["csv_bytes"] = None
+    st.session_state["csv_name"] = None
+
+prep_col1, prep_col2 = st.columns([1, 2])
+
+with prep_col1:
+    prepare = st.button("📦 準備下載 CSV（含進度）")
+
+with prep_col2:
+    st.caption("下載按鈕會在「準備完成」後出現。")
+
+if prepare:
+    # 進度條是「體感進度」：用幾個步驟讓 UX 更順
+    progress = st.progress(0)
+    with st.spinner("正在準備檔案..."):
+        progress.progress(10)
+        time.sleep(0.15)
+
+        progress.progress(35)
+        # 生成 CSV bytes（這步是主要成本）
+        csv_bytes = filtered.to_csv(index=False).encode("utf-8-sig")
+        time.sleep(0.1)
+
+        progress.progress(75)
+        # 你也可以在這裡加上壓縮、欄位整理等
+        time.sleep(0.1)
+
+        progress.progress(100)
+
+    st.session_state["csv_bytes"] = csv_bytes
+    st.session_state["csv_name"] = f"taic_pulse_filtered_{time.strftime('%Y%m%d_%H%M%S')}.csv"
+    st.success("✅ 檔案已準備完成，請使用下方按鈕下載。")
+
+if st.session_state["csv_bytes"] is not None:
+    st.download_button(
+        "⬇️ 下載篩選後 CSV",
+        data=st.session_state["csv_bytes"],
+        file_name=st.session_state["csv_name"],
+        mime="text/csv",
+    )