Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -11,9 +11,9 @@ import pandas as pd
|
|
| 11 |
from bs4 import BeautifulSoup
|
| 12 |
|
| 13 |
PUBLIC_URL = "https://www.fit-portal.go.jp/PublicInfo"
|
| 14 |
-
OUTDIR = "data_fit"
|
| 15 |
|
| 16 |
-
#
|
| 17 |
|
| 18 |
def normalize_filename(name: str) -> str:
|
| 19 |
name = unicodedata.normalize("NFKC", name)
|
|
@@ -45,7 +45,6 @@ def pick_sheet_name(xls_path: str, preferred: str | None) -> str | None:
|
|
| 45 |
xl = pd.ExcelFile(xls_path)
|
| 46 |
if preferred and preferred in xl.sheet_names:
|
| 47 |
return preferred
|
| 48 |
-
# 代表地番を優先
|
| 49 |
for candidate in ["代表地番", "代表地番のみ", "代表地番シート"]:
|
| 50 |
if candidate in xl.sheet_names:
|
| 51 |
return candidate
|
|
@@ -60,13 +59,9 @@ def collect_pref_links(session: requests.Session) -> list[dict]:
|
|
| 60 |
links = []
|
| 61 |
for a in soup.find_all("a"):
|
| 62 |
if is_pref_link(a):
|
| 63 |
-
links.append({
|
| 64 |
-
"pref": extract_pref_name(a),
|
| 65 |
-
"href": urljoin(PUBLIC_URL, a.get("href")),
|
| 66 |
-
})
|
| 67 |
# 重複除去
|
| 68 |
-
seen = set()
|
| 69 |
-
uniq = []
|
| 70 |
for item in links:
|
| 71 |
key = (item["pref"], item["href"])
|
| 72 |
if key not in seen:
|
|
@@ -76,7 +71,6 @@ def collect_pref_links(session: requests.Session) -> list[dict]:
|
|
| 76 |
|
| 77 |
def download_one(session: requests.Session, url: str, outdir: str, pref: str) -> str:
|
| 78 |
os.makedirs(outdir, exist_ok=True)
|
| 79 |
-
from urllib.parse import urlparse, parse_qs
|
| 80 |
qs = parse_qs(urlparse(url).query)
|
| 81 |
file_id = (qs.get("file", ["unknown"])[0])[:18]
|
| 82 |
with session.get(url, timeout=180, stream=True) as r:
|
|
@@ -84,87 +78,108 @@ def download_one(session: requests.Session, url: str, outdir: str, pref: str) ->
|
|
| 84 |
fname = guess_filename_from_headers(r, f"{pref}_{file_id}.xlsx")
|
| 85 |
path = os.path.join(outdir, fname)
|
| 86 |
with open(path, "wb") as f:
|
| 87 |
-
for chunk in r.iter_content(chunk_size=1 << 15)
|
|
|
|
| 88 |
if chunk:
|
| 89 |
f.write(chunk)
|
| 90 |
return path
|
| 91 |
|
| 92 |
-
# ---- 3
|
| 93 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
"""
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
いずれも先頭にメタ列(都道府県/元ファイル/読込シート)を付加
|
| 98 |
"""
|
| 99 |
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 100 |
if not sheet:
|
| 101 |
return None
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 108 |
else:
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
df.insert(1, "元ファイル", os.path.basename(xls_path))
|
| 120 |
-
df.insert(2, "読込シート", sheet)
|
| 121 |
-
|
| 122 |
-
# 1枚目は MultiIndex 列をメタ列も含めて設定しておく
|
| 123 |
-
if use_header and isinstance(pd.Index(orig_cols), pd.MultiIndex):
|
| 124 |
-
meta = [
|
| 125 |
-
("meta", "都道府県", ""),
|
| 126 |
-
("meta", "元ファイル", ""),
|
| 127 |
-
("meta", "読込シート", ""),
|
| 128 |
-
]
|
| 129 |
-
df.columns = pd.MultiIndex.from_tuples(meta + list(orig_cols))
|
| 130 |
-
return df
|
| 131 |
-
except Exception as e:
|
| 132 |
-
print(f"[WARN] 読み込み失敗: {xls_path} ({e})")
|
| 133 |
-
return None
|
| 134 |
|
| 135 |
-
# ---- 3段 → 1段 へフラット化 ----
|
| 136 |
def flatten_columns(cols, sep: str = "_") -> list[str]:
|
| 137 |
"""
|
| 138 |
-
MultiIndex 列を '上位_中位_下位'
|
| 139 |
None / NaN / 空白は除去。重複名は .1, .2... を付与。
|
| 140 |
"""
|
| 141 |
-
# 1) 一旦すべて文字列へ
|
| 142 |
def as_str(x):
|
| 143 |
s = "" if x is None else str(x)
|
| 144 |
s = s.strip()
|
| 145 |
return "" if s.lower() == "nan" else s
|
| 146 |
|
| 147 |
-
flat = []
|
| 148 |
if isinstance(cols, pd.MultiIndex):
|
|
|
|
| 149 |
for tpl in cols:
|
| 150 |
-
parts = [as_str(p) for p in tpl]
|
| 151 |
-
parts
|
| 152 |
-
name = sep.join(parts) if parts else "col"
|
| 153 |
-
flat.append(name)
|
| 154 |
else:
|
| 155 |
-
|
| 156 |
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
uniq = []
|
| 160 |
-
for c in flat:
|
| 161 |
if c not in seen:
|
| 162 |
seen[c] = 0
|
| 163 |
-
|
| 164 |
else:
|
| 165 |
seen[c] += 1
|
| 166 |
-
|
| 167 |
-
return
|
| 168 |
|
| 169 |
def zip_paths(paths: list[str], out_zip: str) -> str:
|
| 170 |
with zipfile.ZipFile(out_zip, "w", compression=zipfile.ZIP_DEFLATED) as z:
|
|
@@ -173,34 +188,23 @@ def zip_paths(paths: list[str], out_zip: str) -> str:
|
|
| 173 |
z.write(p, arcname=os.path.basename(p))
|
| 174 |
return out_zip
|
| 175 |
|
| 176 |
-
#
|
| 177 |
|
| 178 |
def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress=gr.Progress(track_tqdm=False)):
|
| 179 |
-
"""
|
| 180 |
-
sheet_name: "代表地番" 等。空欄なら自動
|
| 181 |
-
sleep_sec: ダウンロード間隔
|
| 182 |
-
limit: 先頭N件のみ(テスト用)
|
| 183 |
-
re_download: 既存ファイルがあっても再取得する
|
| 184 |
-
do_flatten: 列名を1段にフラット化する
|
| 185 |
-
sep: フラット化時のセパレータ
|
| 186 |
-
"""
|
| 187 |
progress(0, desc="初期化中…")
|
| 188 |
|
| 189 |
session = requests.Session()
|
| 190 |
session.headers.update({
|
| 191 |
-
"User-Agent": "Mozilla/5.0 (compatible; FITCollector/1.
|
| 192 |
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
|
| 193 |
})
|
| 194 |
|
| 195 |
# 1) リンク収集
|
| 196 |
links = collect_pref_links(session)
|
| 197 |
if not links:
|
| 198 |
-
return ("
|
| 199 |
-
None, None, None, None)
|
| 200 |
-
|
| 201 |
if limit and limit > 0:
|
| 202 |
links = links[:int(limit)]
|
| 203 |
-
|
| 204 |
progress(0.1, desc=f"リンク検出 {len(links)} 件")
|
| 205 |
|
| 206 |
# 2) ダウンロード
|
|
@@ -209,7 +213,6 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 209 |
progress(0.1 + 0.6 * i / max(1, len(links)),
|
| 210 |
desc=f"ダウンロード {i}/{len(links)}: {item['pref']}")
|
| 211 |
try:
|
| 212 |
-
# 既存ファイルを利用(高速化)
|
| 213 |
existing = None
|
| 214 |
if not re_download and os.path.isdir(OUTDIR):
|
| 215 |
for fn in os.listdir(OUTDIR):
|
|
@@ -221,44 +224,42 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 221 |
else:
|
| 222 |
path = download_one(session, item["href"], OUTDIR, item["pref"])
|
| 223 |
time.sleep(float(sleep_sec))
|
| 224 |
-
downloaded.append(
|
| 225 |
except Exception as e:
|
| 226 |
print(f"[WARN] ダウンロード失敗: {item['pref']} {e}")
|
| 227 |
|
| 228 |
if not downloaded:
|
| 229 |
-
return ("
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 230 |
None, None, None, None)
|
| 231 |
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
|
| 235 |
-
|
| 236 |
-
|
| 237 |
-
|
| 238 |
-
|
| 239 |
-
sheet_name if sheet_name else None,
|
| 240 |
-
it["pref"],
|
| 241 |
-
use_header=(i == 1) # 1枚目だけ上3行を列名に
|
| 242 |
-
)
|
| 243 |
if df is not None and len(df) > 0:
|
| 244 |
frames.append(df)
|
|
|
|
|
|
|
| 245 |
|
| 246 |
-
|
| 247 |
-
return ("Excelは取得できましたが、読み込めるデータがありませんでした(シート名の指定を見直してください)。",
|
| 248 |
-
None, None, None, None)
|
| 249 |
-
|
| 250 |
-
# 2枚目以降は列名が無いので、1枚目の列名を適用
|
| 251 |
-
if len(frames) > 1:
|
| 252 |
-
frames[1:] = [f.set_axis(frames[0].columns, axis=1) for f in frames[1:]]
|
| 253 |
-
|
| 254 |
-
# 4) 縦結合
|
| 255 |
combined = pd.concat(frames, ignore_index=True)
|
| 256 |
|
| 257 |
-
#
|
| 258 |
if do_flatten:
|
| 259 |
combined.columns = flatten_columns(combined.columns, sep=sep or "_")
|
| 260 |
|
| 261 |
-
#
|
| 262 |
os.makedirs(OUTDIR, exist_ok=True)
|
| 263 |
out_xlsx = os.path.join(OUTDIR, "combined_fit.xlsx")
|
| 264 |
out_parq = os.path.join(OUTDIR, "combined_fit.parquet")
|
|
@@ -266,16 +267,15 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 266 |
combined.to_excel(w, index=False, sheet_name="combined")
|
| 267 |
combined.to_parquet(out_parq, index=False)
|
| 268 |
|
| 269 |
-
#
|
| 270 |
raw_zip = os.path.join(OUTDIR, "raw_excels.zip")
|
| 271 |
-
zip_paths(
|
| 272 |
|
| 273 |
-
#
|
| 274 |
preview_csv = os.path.join(OUTDIR, "combined_head.csv")
|
| 275 |
combined.head(1000).to_csv(preview_csv, index=False)
|
| 276 |
|
| 277 |
progress(1.0, desc=f"完了({len(combined):,} 行)")
|
| 278 |
-
|
| 279 |
msg = (
|
| 280 |
f"✅ 結合完了: 行数 = {len(combined):,}\n"
|
| 281 |
f"・Excel: combined_fit.xlsx\n"
|
|
@@ -284,17 +284,19 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 284 |
f"・プレビュー: combined_head.csv\n"
|
| 285 |
f"・列名フラット化: {'ON' if do_flatten else 'OFF'}(区切り: '{sep or '_'}')"
|
| 286 |
)
|
| 287 |
-
|
| 288 |
return (msg, out_xlsx, out_parq, raw_zip, preview_csv)
|
| 289 |
|
| 290 |
-
#
|
| 291 |
|
| 292 |
with gr.Blocks(title="FIT 公表(都道府県別Excel)一括取得&結合") as demo:
|
| 293 |
gr.Markdown(
|
| 294 |
"""
|
| 295 |
# FIT 公表(都道府県別Excel)一括取得 & 結合
|
| 296 |
-
|
| 297 |
-
-
|
|
|
|
|
|
|
|
|
|
| 298 |
"""
|
| 299 |
)
|
| 300 |
with gr.Row():
|
|
|
|
| 11 |
from bs4 import BeautifulSoup
|
| 12 |
|
| 13 |
PUBLIC_URL = "https://www.fit-portal.go.jp/PublicInfo"
|
| 14 |
+
OUTDIR = "data_fit"
|
| 15 |
|
| 16 |
+
# ---- ユーティリティ ---------------------------------------------------------
|
| 17 |
|
| 18 |
def normalize_filename(name: str) -> str:
|
| 19 |
name = unicodedata.normalize("NFKC", name)
|
|
|
|
| 45 |
xl = pd.ExcelFile(xls_path)
|
| 46 |
if preferred and preferred in xl.sheet_names:
|
| 47 |
return preferred
|
|
|
|
| 48 |
for candidate in ["代表地番", "代表地番のみ", "代表地番シート"]:
|
| 49 |
if candidate in xl.sheet_names:
|
| 50 |
return candidate
|
|
|
|
| 59 |
links = []
|
| 60 |
for a in soup.find_all("a"):
|
| 61 |
if is_pref_link(a):
|
| 62 |
+
links.append({"pref": extract_pref_name(a), "href": urljoin(PUBLIC_URL, a.get("href"))})
|
|
|
|
|
|
|
|
|
|
| 63 |
# 重複除去
|
| 64 |
+
seen, uniq = set(), []
|
|
|
|
| 65 |
for item in links:
|
| 66 |
key = (item["pref"], item["href"])
|
| 67 |
if key not in seen:
|
|
|
|
| 71 |
|
| 72 |
def download_one(session: requests.Session, url: str, outdir: str, pref: str) -> str:
|
| 73 |
os.makedirs(outdir, exist_ok=True)
|
|
|
|
| 74 |
qs = parse_qs(urlparse(url).query)
|
| 75 |
file_id = (qs.get("file", ["unknown"])[0])[:18]
|
| 76 |
with session.get(url, timeout=180, stream=True) as r:
|
|
|
|
| 78 |
fname = guess_filename_from_headers(r, f"{pref}_{file_id}.xlsx")
|
| 79 |
path = os.path.join(outdir, fname)
|
| 80 |
with open(path, "wb") as f:
|
| 81 |
+
for chunk in r.iter_iterable = r.iter_content(chunk_size=1 << 15)
|
| 82 |
+
for chunk in iter_iterable:
|
| 83 |
if chunk:
|
| 84 |
f.write(chunk)
|
| 85 |
return path
|
| 86 |
|
| 87 |
+
# ---- 3段ヘッダ → 1枚目のみ採用/他はスキップ行数で読込 ----------------------
|
| 88 |
+
|
| 89 |
+
HEADER_ROWS = [1, 2, 3] # 0行目は削除、1/2/3行目を列名として結合
|
| 90 |
+
SKIP_ROWS_OTHERS = 4 # 2枚目以降は 0〜3 行目をスキップ
|
| 91 |
+
|
| 92 |
+
def load_excel_first(xls_path: str, sheet_pref: str | None) -> tuple[pd.DataFrame, list]:
|
| 93 |
+
"""
|
| 94 |
+
1枚目: 0行目は使わず、1/2/3行目を列名にする(MultiIndex)。
|
| 95 |
+
さらに「一番左の列」を削除して返す。
|
| 96 |
+
戻り値: (df, columns_multiindex)
|
| 97 |
+
"""
|
| 98 |
+
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 99 |
+
if not sheet:
|
| 100 |
+
raise RuntimeError("シートが見つかりません")
|
| 101 |
+
df = pd.read_excel(
|
| 102 |
+
xls_path,
|
| 103 |
+
sheet_name=sheet,
|
| 104 |
+
engine="openpyxl",
|
| 105 |
+
header=HEADER_ROWS,
|
| 106 |
+
dtype=str
|
| 107 |
+
)
|
| 108 |
+
# 左端の余計な列を削除
|
| 109 |
+
df = df.iloc[:, 1:]
|
| 110 |
+
# 前後空白トリム
|
| 111 |
+
for c in df.select_dtypes(include=["object"]).columns:
|
| 112 |
+
df[c] = df[c].str.strip()
|
| 113 |
+
cols = list(df.columns) # MultiIndex のまま保持
|
| 114 |
+
return df, cols
|
| 115 |
+
|
| 116 |
+
def load_excel_other(xls_path: str, sheet_pref: str | None, target_cols: list) -> pd.DataFrame | None:
|
| 117 |
"""
|
| 118 |
+
2枚目以降: 3行目までスキップしてデータのみ読み込み。
|
| 119 |
+
左端列を削除後、1枚目の列(MultiIndex)を適用。
|
|
|
|
| 120 |
"""
|
| 121 |
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 122 |
if not sheet:
|
| 123 |
return None
|
| 124 |
+
df = pd.read_excel(
|
| 125 |
+
xls_path,
|
| 126 |
+
sheet_name=sheet,
|
| 127 |
+
engine="openpyxl",
|
| 128 |
+
header=None,
|
| 129 |
+
skiprows=SKIP_ROWS_OTHERS,
|
| 130 |
+
dtype=str
|
| 131 |
+
)
|
| 132 |
+
# 左端の余計な列を削除
|
| 133 |
+
df = df.iloc[:, 1:]
|
| 134 |
+
# 前後空白トリム
|
| 135 |
+
for c in df.select_dtypes(include=["object"]).columns:
|
| 136 |
+
df[c] = df[c].str.strip()
|
| 137 |
+
|
| 138 |
+
# 列数が合わない場合は合わせられる範囲で調整(警告付き)
|
| 139 |
+
if df.shape[1] != len(target_cols):
|
| 140 |
+
print(f"[WARN] 列数不一致: file={os.path.basename(xls_path)} "
|
| 141 |
+
f"read={df.shape[1]} vs target={len(target_cols)} -> 自動調整")
|
| 142 |
+
if df.shape[1] > len(target_cols):
|
| 143 |
+
df = df.iloc[:, :len(target_cols)]
|
| 144 |
else:
|
| 145 |
+
# 足りない場合は欠損列を追加
|
| 146 |
+
for _ in range(len(target_cols) - df.shape[1]):
|
| 147 |
+
df[pd.util.hash_pandas_object(df).name or f"_pad_{_}"] = None
|
| 148 |
+
# 列順を並べ替え
|
| 149 |
+
df = df.iloc[:, :len(target_cols)]
|
| 150 |
+
|
| 151 |
+
df.columns = target_cols
|
| 152 |
+
return df
|
| 153 |
+
|
| 154 |
+
# ---- 列名フラット化 ---------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 155 |
|
|
|
|
| 156 |
def flatten_columns(cols, sep: str = "_") -> list[str]:
|
| 157 |
"""
|
| 158 |
+
MultiIndex 列を '上位_中位_下位' にフラット化。
|
| 159 |
None / NaN / 空白は除去。重複名は .1, .2... を付与。
|
| 160 |
"""
|
|
|
|
| 161 |
def as_str(x):
|
| 162 |
s = "" if x is None else str(x)
|
| 163 |
s = s.strip()
|
| 164 |
return "" if s.lower() == "nan" else s
|
| 165 |
|
|
|
|
| 166 |
if isinstance(cols, pd.MultiIndex):
|
| 167 |
+
raw = []
|
| 168 |
for tpl in cols:
|
| 169 |
+
parts = [as_str(p) for p in tpl if as_str(p)]
|
| 170 |
+
raw.append(sep.join(parts) if parts else "col")
|
|
|
|
|
|
|
| 171 |
else:
|
| 172 |
+
raw = [as_str(c) or "col" for c in cols]
|
| 173 |
|
| 174 |
+
seen, out = {}, []
|
| 175 |
+
for c in raw:
|
|
|
|
|
|
|
| 176 |
if c not in seen:
|
| 177 |
seen[c] = 0
|
| 178 |
+
out.append(c)
|
| 179 |
else:
|
| 180 |
seen[c] += 1
|
| 181 |
+
out.append(f"{c}.{seen[c]}")
|
| 182 |
+
return out
|
| 183 |
|
| 184 |
def zip_paths(paths: list[str], out_zip: str) -> str:
|
| 185 |
with zipfile.ZipFile(out_zip, "w", compression=zipfile.ZIP_DEFLATED) as z:
|
|
|
|
| 188 |
z.write(p, arcname=os.path.basename(p))
|
| 189 |
return out_zip
|
| 190 |
|
| 191 |
+
# ---- メイン(Gradioから呼び出し) -------------------------------------------
|
| 192 |
|
| 193 |
def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress=gr.Progress(track_tqdm=False)):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 194 |
progress(0, desc="初期化中…")
|
| 195 |
|
| 196 |
session = requests.Session()
|
| 197 |
session.headers.update({
|
| 198 |
+
"User-Agent": "Mozilla/5.0 (compatible; FITCollector/1.2; +https://huggingface.co/spaces)",
|
| 199 |
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
|
| 200 |
})
|
| 201 |
|
| 202 |
# 1) リンク収集
|
| 203 |
links = collect_pref_links(session)
|
| 204 |
if not links:
|
| 205 |
+
return ("都道府県ファイルのリンク検出に失敗しました。", None, None, None, None)
|
|
|
|
|
|
|
| 206 |
if limit and limit > 0:
|
| 207 |
links = links[:int(limit)]
|
|
|
|
| 208 |
progress(0.1, desc=f"リンク検出 {len(links)} 件")
|
| 209 |
|
| 210 |
# 2) ダウンロード
|
|
|
|
| 213 |
progress(0.1 + 0.6 * i / max(1, len(links)),
|
| 214 |
desc=f"ダウンロード {i}/{len(links)}: {item['pref']}")
|
| 215 |
try:
|
|
|
|
| 216 |
existing = None
|
| 217 |
if not re_download and os.path.isdir(OUTDIR):
|
| 218 |
for fn in os.listdir(OUTDIR):
|
|
|
|
| 224 |
else:
|
| 225 |
path = download_one(session, item["href"], OUTDIR, item["pref"])
|
| 226 |
time.sleep(float(sleep_sec))
|
| 227 |
+
downloaded.append(path)
|
| 228 |
except Exception as e:
|
| 229 |
print(f"[WARN] ダウンロード失敗: {item['pref']} {e}")
|
| 230 |
|
| 231 |
if not downloaded:
|
| 232 |
+
return ("ダウンロードに失敗しました。", None, None, None, None)
|
| 233 |
+
|
| 234 |
+
# 3) 読み込み・1枚目
|
| 235 |
+
progress(0.75, desc="1枚目を読み込み(列名生成)")
|
| 236 |
+
first_path = downloaded[0]
|
| 237 |
+
try:
|
| 238 |
+
df0, cols0 = load_excel_first(first_path, sheet_name if sheet_name else None)
|
| 239 |
+
except Exception as e:
|
| 240 |
+
return (f"1枚目の読み込みに失敗しました: {os.path.basename(first_path)} / {e}",
|
| 241 |
None, None, None, None)
|
| 242 |
|
| 243 |
+
frames = [df0]
|
| 244 |
+
|
| 245 |
+
# 4) 読み込み・2枚目以降
|
| 246 |
+
for j, p in enumerate(downloaded[1:], start=2):
|
| 247 |
+
progress(0.75 + 0.25 * (j - 1) / max(1, len(downloaded) - 1),
|
| 248 |
+
desc=f"{j}枚目を読み込み")
|
| 249 |
+
df = load_excel_other(p, sheet_name if sheet_name else None, cols0)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 250 |
if df is not None and len(df) > 0:
|
| 251 |
frames.append(df)
|
| 252 |
+
else:
|
| 253 |
+
print(f"[WARN] 読み込みスキップ: {os.path.basename(p)}")
|
| 254 |
|
| 255 |
+
# 5) 縦結合
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 256 |
combined = pd.concat(frames, ignore_index=True)
|
| 257 |
|
| 258 |
+
# 6) 列名のフラット化(既定ON)
|
| 259 |
if do_flatten:
|
| 260 |
combined.columns = flatten_columns(combined.columns, sep=sep or "_")
|
| 261 |
|
| 262 |
+
# 7) 出力
|
| 263 |
os.makedirs(OUTDIR, exist_ok=True)
|
| 264 |
out_xlsx = os.path.join(OUTDIR, "combined_fit.xlsx")
|
| 265 |
out_parq = os.path.join(OUTDIR, "combined_fit.parquet")
|
|
|
|
| 267 |
combined.to_excel(w, index=False, sheet_name="combined")
|
| 268 |
combined.to_parquet(out_parq, index=False)
|
| 269 |
|
| 270 |
+
# 8) ZIP(取得ファイル一式)
|
| 271 |
raw_zip = os.path.join(OUTDIR, "raw_excels.zip")
|
| 272 |
+
zip_paths(downloaded, raw_zip)
|
| 273 |
|
| 274 |
+
# 9) プレビュー
|
| 275 |
preview_csv = os.path.join(OUTDIR, "combined_head.csv")
|
| 276 |
combined.head(1000).to_csv(preview_csv, index=False)
|
| 277 |
|
| 278 |
progress(1.0, desc=f"完了({len(combined):,} 行)")
|
|
|
|
| 279 |
msg = (
|
| 280 |
f"✅ 結合完了: 行数 = {len(combined):,}\n"
|
| 281 |
f"・Excel: combined_fit.xlsx\n"
|
|
|
|
| 284 |
f"・プレビュー: combined_head.csv\n"
|
| 285 |
f"・列名フラット化: {'ON' if do_flatten else 'OFF'}(区切り: '{sep or '_'}')"
|
| 286 |
)
|
|
|
|
| 287 |
return (msg, out_xlsx, out_parq, raw_zip, preview_csv)
|
| 288 |
|
| 289 |
+
# ---- Gradio UI -------------------------------------------------------------
|
| 290 |
|
| 291 |
with gr.Blocks(title="FIT 公表(都道府県別Excel)一括取得&結合") as demo:
|
| 292 |
gr.Markdown(
|
| 293 |
"""
|
| 294 |
# FIT 公表(都道府県別Excel)一括取得 & 結合
|
| 295 |
+
**要件に沿った処理**:
|
| 296 |
+
- 1枚目のみ「0行目を削除」「1/2/3行目を結合して列名」。
|
| 297 |
+
- 2枚目以降は「3行目までスキップ」してデータのみ。
|
| 298 |
+
- すべてのファイルで**左端の列を削除**。
|
| 299 |
+
- ファイル/シート名などのメタ列は**付与しません**。
|
| 300 |
"""
|
| 301 |
)
|
| 302 |
with gr.Row():
|