Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -13,7 +13,7 @@ from bs4 import BeautifulSoup
|
|
| 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)
|
|
@@ -45,6 +45,7 @@ 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 |
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
|
|
@@ -78,56 +79,66 @@ def download_one(session: requests.Session, url: str, outdir: str, pref: str) ->
|
|
| 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 |
-
# ✅ ここを修正:通常の iter_content ループに
|
| 82 |
for chunk in r.iter_content(chunk_size=1 << 15):
|
| 83 |
if chunk:
|
| 84 |
f.write(chunk)
|
| 85 |
return path
|
| 86 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
def choose_names_from_multiindex(mi: pd.MultiIndex) -> list[str]:
|
| 88 |
"""
|
| 89 |
-
3段ヘッダ(MultiIndex
|
| 90 |
-
|
| 91 |
-
|
|
|
|
| 92 |
"""
|
| 93 |
-
|
| 94 |
-
if x is None:
|
| 95 |
-
return ""
|
| 96 |
-
s = str(x).strip()
|
| 97 |
-
return "" if s.lower() == "nan" else s
|
| 98 |
-
|
| 99 |
-
# 優先で選択
|
| 100 |
-
picked = []
|
| 101 |
for tpl in mi:
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 107 |
|
| 108 |
# 重複解消
|
| 109 |
seen = {}
|
| 110 |
-
|
| 111 |
-
for n in
|
| 112 |
if n not in seen:
|
| 113 |
seen[n] = 0
|
| 114 |
-
|
| 115 |
else:
|
| 116 |
seen[n] += 1
|
| 117 |
-
|
| 118 |
-
return
|
| 119 |
-
|
| 120 |
-
|
| 121 |
-
# ---- 3段ヘッダ → 1枚目のみ採用/他はスキップ行数で読込 ----------------------
|
| 122 |
|
| 123 |
-
|
| 124 |
-
|
|
|
|
|
|
|
|
|
|
| 125 |
|
| 126 |
-
def load_excel_first(xls_path: str, sheet_pref: str | None) -> tuple[pd.DataFrame, list]:
|
| 127 |
"""
|
| 128 |
-
1枚目:
|
| 129 |
-
|
| 130 |
-
|
|
|
|
|
|
|
| 131 |
"""
|
| 132 |
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 133 |
if not sheet:
|
|
@@ -139,37 +150,37 @@ def load_excel_first(xls_path: str, sheet_pref: str | None) -> tuple[pd.DataFram
|
|
| 139 |
header=HEADER_ROWS,
|
| 140 |
dtype=str
|
| 141 |
)
|
| 142 |
-
#
|
| 143 |
df = df.iloc[:, 1:]
|
| 144 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
| 145 |
if isinstance(df.columns, pd.MultiIndex):
|
| 146 |
chosen = choose_names_from_multiindex(df.columns)
|
| 147 |
else:
|
| 148 |
-
#
|
| 149 |
-
raw = []
|
| 150 |
-
for c in df.columns:
|
| 151 |
-
s = "" if c is None else str(c).strip()
|
| 152 |
-
raw.append("" if s.lower() == "nan" else s)
|
| 153 |
-
raw = [r if r else "col" for r in raw]
|
| 154 |
seen = {}
|
| 155 |
chosen = []
|
| 156 |
for n in raw:
|
| 157 |
-
if n in seen:
|
| 158 |
-
seen[n] += 1
|
| 159 |
-
chosen.append(f"{n}.{seen[n]}")
|
| 160 |
-
else:
|
| 161 |
seen[n] = 0
|
| 162 |
chosen.append(n)
|
| 163 |
-
|
|
|
|
|
|
|
| 164 |
df.columns = chosen
|
|
|
|
| 165 |
|
| 166 |
-
|
| 167 |
-
return df, cols
|
| 168 |
-
|
| 169 |
-
def load_excel_other(xls_path: str, sheet_pref: str | None, target_cols: list) -> pd.DataFrame | None:
|
| 170 |
"""
|
| 171 |
-
2枚目以降:
|
| 172 |
-
|
|
|
|
|
|
|
|
|
|
| 173 |
"""
|
| 174 |
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 175 |
if not sheet:
|
|
@@ -182,58 +193,27 @@ def load_excel_other(xls_path: str, sheet_pref: str | None, target_cols: list) -
|
|
| 182 |
skiprows=SKIP_ROWS_OTHERS,
|
| 183 |
dtype=str
|
| 184 |
)
|
| 185 |
-
#
|
| 186 |
df = df.iloc[:, 1:]
|
| 187 |
# 前後空白トリム
|
| 188 |
for c in df.select_dtypes(include=["object"]).columns:
|
| 189 |
df[c] = df[c].str.strip()
|
| 190 |
|
| 191 |
-
#
|
| 192 |
if df.shape[1] != len(target_cols):
|
| 193 |
print(f"[WARN] 列数不一致: file={os.path.basename(xls_path)} "
|
| 194 |
f"read={df.shape[1]} vs target={len(target_cols)} -> 自動調整")
|
| 195 |
if df.shape[1] > len(target_cols):
|
| 196 |
df = df.iloc[:, :len(target_cols)]
|
| 197 |
else:
|
| 198 |
-
#
|
| 199 |
-
for
|
| 200 |
-
df[
|
| 201 |
-
# 列順を並べ替え
|
| 202 |
df = df.iloc[:, :len(target_cols)]
|
| 203 |
|
| 204 |
df.columns = target_cols
|
| 205 |
return df
|
| 206 |
|
| 207 |
-
# ---- 列名フラット化 ---------------------------------------------------------
|
| 208 |
-
|
| 209 |
-
def flatten_columns(cols, sep: str = "_") -> list[str]:
|
| 210 |
-
"""
|
| 211 |
-
MultiIndex 列を '上位_中位_下位' にフラット化。
|
| 212 |
-
None / NaN / 空白は除去。重複名は .1, .2... を付与。
|
| 213 |
-
"""
|
| 214 |
-
def as_str(x):
|
| 215 |
-
s = "" if x is None else str(x)
|
| 216 |
-
s = s.strip()
|
| 217 |
-
return "" if s.lower() == "nan" else s
|
| 218 |
-
|
| 219 |
-
if isinstance(cols, pd.MultiIndex):
|
| 220 |
-
raw = []
|
| 221 |
-
for tpl in cols:
|
| 222 |
-
parts = [as_str(p) for p in tpl if as_str(p)]
|
| 223 |
-
raw.append(sep.join(parts) if parts else "col")
|
| 224 |
-
else:
|
| 225 |
-
raw = [as_str(c) or "col" for c in cols]
|
| 226 |
-
|
| 227 |
-
seen, out = {}, []
|
| 228 |
-
for c in raw:
|
| 229 |
-
if c not in seen:
|
| 230 |
-
seen[c] = 0
|
| 231 |
-
out.append(c)
|
| 232 |
-
else:
|
| 233 |
-
seen[c] += 1
|
| 234 |
-
out.append(f"{c}.{seen[c]}")
|
| 235 |
-
return out
|
| 236 |
-
|
| 237 |
def zip_paths(paths: list[str], out_zip: str) -> str:
|
| 238 |
with zipfile.ZipFile(out_zip, "w", compression=zipfile.ZIP_DEFLATED) as z:
|
| 239 |
for p in paths:
|
|
@@ -241,21 +221,22 @@ def zip_paths(paths: list[str], out_zip: str) -> str:
|
|
| 241 |
z.write(p, arcname=os.path.basename(p))
|
| 242 |
return out_zip
|
| 243 |
|
| 244 |
-
#
|
| 245 |
|
| 246 |
-
def run_job(sheet_name, sleep_sec, limit, re_download,
|
| 247 |
progress(0, desc="初期化中…")
|
| 248 |
|
| 249 |
session = requests.Session()
|
| 250 |
session.headers.update({
|
| 251 |
-
"User-Agent": "Mozilla/5.0 (compatible; FITCollector/1.
|
| 252 |
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
|
| 253 |
})
|
| 254 |
|
| 255 |
# 1) リンク収集
|
| 256 |
links = collect_pref_links(session)
|
| 257 |
if not links:
|
| 258 |
-
return ("
|
|
|
|
| 259 |
if limit and limit > 0:
|
| 260 |
links = links[:int(limit)]
|
| 261 |
progress(0.1, desc=f"リンク検出 {len(links)} 件")
|
|
@@ -284,8 +265,8 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 284 |
if not downloaded:
|
| 285 |
return ("ダウンロードに失敗しました。", None, None, None, None)
|
| 286 |
|
| 287 |
-
# 3)
|
| 288 |
-
progress(0.75, desc="1
|
| 289 |
first_path = downloaded[0]
|
| 290 |
try:
|
| 291 |
df0, cols0 = load_excel_first(first_path, sheet_name if sheet_name else None)
|
|
@@ -295,7 +276,7 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 295 |
|
| 296 |
frames = [df0]
|
| 297 |
|
| 298 |
-
# 4)
|
| 299 |
for j, p in enumerate(downloaded[1:], start=2):
|
| 300 |
progress(0.75 + 0.25 * (j - 1) / max(1, len(downloaded) - 1),
|
| 301 |
desc=f"{j}枚目を読み込み")
|
|
@@ -308,11 +289,7 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 308 |
# 5) 縦結合
|
| 309 |
combined = pd.concat(frames, ignore_index=True)
|
| 310 |
|
| 311 |
-
# 6)
|
| 312 |
-
if do_flatten:
|
| 313 |
-
combined.columns = flatten_columns(combined.columns, sep=sep or "_")
|
| 314 |
-
|
| 315 |
-
# 7) 出力
|
| 316 |
os.makedirs(OUTDIR, exist_ok=True)
|
| 317 |
out_xlsx = os.path.join(OUTDIR, "combined_fit.xlsx")
|
| 318 |
out_parq = os.path.join(OUTDIR, "combined_fit.parquet")
|
|
@@ -320,11 +297,11 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 320 |
combined.to_excel(w, index=False, sheet_name="combined")
|
| 321 |
combined.to_parquet(out_parq, index=False)
|
| 322 |
|
| 323 |
-
#
|
| 324 |
raw_zip = os.path.join(OUTDIR, "raw_excels.zip")
|
| 325 |
zip_paths(downloaded, raw_zip)
|
| 326 |
|
| 327 |
-
#
|
| 328 |
preview_csv = os.path.join(OUTDIR, "combined_head.csv")
|
| 329 |
combined.head(1000).to_csv(preview_csv, index=False)
|
| 330 |
|
|
@@ -335,21 +312,22 @@ def run_job(sheet_name, sleep_sec, limit, re_download, do_flatten, sep, progress
|
|
| 335 |
f"・Parquet: combined_fit.parquet\n"
|
| 336 |
f"・Raw ZIP: raw_excels.zip\n"
|
| 337 |
f"・プレビュー: combined_head.csv\n"
|
| 338 |
-
f"
|
| 339 |
)
|
| 340 |
return (msg, out_xlsx, out_parq, raw_zip, preview_csv)
|
| 341 |
|
| 342 |
-
#
|
| 343 |
|
| 344 |
with gr.Blocks(title="FIT 公表(都道府県別Excel)一括取得&結合") as demo:
|
| 345 |
gr.Markdown(
|
| 346 |
"""
|
| 347 |
# FIT 公表(都道府県別Excel)一括取得 & 結合
|
| 348 |
-
|
| 349 |
-
- 1
|
| 350 |
-
-
|
| 351 |
-
-
|
| 352 |
-
-
|
|
|
|
| 353 |
"""
|
| 354 |
)
|
| 355 |
with gr.Row():
|
|
@@ -358,9 +336,6 @@ with gr.Blocks(title="FIT 公表(都道府県別Excel)一括取得&結合"
|
|
| 358 |
with gr.Row():
|
| 359 |
limit = gr.Number(value=None, precision=0, label="先頭N県のみ(テスト用・空欄は全県)")
|
| 360 |
reget = gr.Checkbox(label="既存ファイルがあっても再ダウンロードする", value=False)
|
| 361 |
-
with gr.Accordion("列名オプション", open=True):
|
| 362 |
-
do_flatten = gr.Checkbox(label="列名を1段にフラット化(推奨)", value=True)
|
| 363 |
-
sep = gr.Textbox(label="フラット化セパレータ", value="_", placeholder="例)_, /, | など")
|
| 364 |
|
| 365 |
run_btn = gr.Button("実行", variant="primary")
|
| 366 |
out_msg = gr.Markdown()
|
|
@@ -371,7 +346,7 @@ with gr.Blocks(title="FIT 公表(都道府県別Excel)一括取得&結合"
|
|
| 371 |
|
| 372 |
run_btn.click(
|
| 373 |
fn=run_job,
|
| 374 |
-
inputs=[sheet, sleep, limit, reget
|
| 375 |
outputs=[out_msg, out_xlsx, out_parq, out_zip, out_preview]
|
| 376 |
)
|
| 377 |
|
|
|
|
| 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 |
+
# 一般的に「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
|
|
|
|
| 79 |
fname = guess_filename_from_headers(r, f"{pref}_{file_id}.xlsx")
|
| 80 |
path = os.path.join(outdir, fname)
|
| 81 |
with open(path, "wb") as f:
|
|
|
|
| 82 |
for chunk in r.iter_content(chunk_size=1 << 15):
|
| 83 |
if chunk:
|
| 84 |
f.write(chunk)
|
| 85 |
return path
|
| 86 |
+
|
| 87 |
+
# -------------------- 列名選択: 小分類 > 中分類 > 大分類 --------------------
|
| 88 |
+
|
| 89 |
+
def _clean_cell(x) -> str:
|
| 90 |
+
if x is None:
|
| 91 |
+
return ""
|
| 92 |
+
s = str(x).strip()
|
| 93 |
+
if s.lower() == "nan":
|
| 94 |
+
return ""
|
| 95 |
+
return s
|
| 96 |
+
|
| 97 |
def choose_names_from_multiindex(mi: pd.MultiIndex) -> list[str]:
|
| 98 |
"""
|
| 99 |
+
3段ヘッダ(MultiIndex)から列名を選ぶ。
|
| 100 |
+
ルール: 小分類(第3段)に値があればそれ、無ければ中分類(第2段)、
|
| 101 |
+
それも無ければ大分類(第1段)。すべて空なら 'col'。
|
| 102 |
+
最後に重複を .1, .2… で解消。
|
| 103 |
"""
|
| 104 |
+
names = []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
for tpl in mi:
|
| 106 |
+
# tpl は (大, 中, 小) 想定
|
| 107 |
+
if len(tpl) < 3:
|
| 108 |
+
# 念のため不足時の安全対策
|
| 109 |
+
a = _clean_cell(tpl[0]) if len(tpl) >= 1 else ""
|
| 110 |
+
b = _clean_cell(tpl[1]) if len(tpl) >= 2 else ""
|
| 111 |
+
c = ""
|
| 112 |
+
else:
|
| 113 |
+
a, b, c = (_clean_cell(tpl[0]), _clean_cell(tpl[1]), _clean_cell(tpl[2]))
|
| 114 |
+
name = c or b or a or "col"
|
| 115 |
+
names.append(name)
|
| 116 |
|
| 117 |
# 重複解消
|
| 118 |
seen = {}
|
| 119 |
+
out = []
|
| 120 |
+
for n in names:
|
| 121 |
if n not in seen:
|
| 122 |
seen[n] = 0
|
| 123 |
+
out.append(n)
|
| 124 |
else:
|
| 125 |
seen[n] += 1
|
| 126 |
+
out.append(f"{n}.{seen[n]}")
|
| 127 |
+
return out
|
|
|
|
|
|
|
|
|
|
| 128 |
|
| 129 |
+
# -------------------- 読み込みルール --------------------
|
| 130 |
+
# 0行目は削除し、1/2/3行目をヘッダ(= header=[1,2,3])
|
| 131 |
+
HEADER_ROWS = [1, 2, 3]
|
| 132 |
+
# 2枚目以降は 0〜3行目をスキップ(= skiprows=4)、header=None でデータのみ
|
| 133 |
+
SKIP_ROWS_OTHERS = 4
|
| 134 |
|
| 135 |
+
def load_excel_first(xls_path: str, sheet_pref: str | None) -> tuple[pd.DataFrame, list[str]]:
|
| 136 |
"""
|
| 137 |
+
1枚目:
|
| 138 |
+
- header=[1,2,3] で3段ヘッダを読み込み(0行目は自動的に使われない)
|
| 139 |
+
- 左端の列を削除
|
| 140 |
+
- MultiIndex から列名を「小>中>大」の優先で単一行に変換
|
| 141 |
+
戻り値: (df, chosen_names)
|
| 142 |
"""
|
| 143 |
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 144 |
if not sheet:
|
|
|
|
| 150 |
header=HEADER_ROWS,
|
| 151 |
dtype=str
|
| 152 |
)
|
| 153 |
+
# 左端の列を削除
|
| 154 |
df = df.iloc[:, 1:]
|
| 155 |
+
# 前後空白トリム
|
| 156 |
+
for c in df.select_dtypes(include=["object"]).columns:
|
| 157 |
+
df[c] = df[c].str.strip()
|
| 158 |
+
|
| 159 |
+
# 列名を選択
|
| 160 |
if isinstance(df.columns, pd.MultiIndex):
|
| 161 |
chosen = choose_names_from_multiindex(df.columns)
|
| 162 |
else:
|
| 163 |
+
# 念のため単層だった場合もクリーニング&重複解消
|
| 164 |
+
raw = [_clean_cell(c) or "col" for c in df.columns]
|
|
|
|
|
|
|
|
|
|
|
|
|
| 165 |
seen = {}
|
| 166 |
chosen = []
|
| 167 |
for n in raw:
|
| 168 |
+
if n not in seen:
|
|
|
|
|
|
|
|
|
|
| 169 |
seen[n] = 0
|
| 170 |
chosen.append(n)
|
| 171 |
+
else:
|
| 172 |
+
seen[n] += 1
|
| 173 |
+
chosen.append(f"{n}.{seen[n]}")
|
| 174 |
df.columns = chosen
|
| 175 |
+
return df, chosen
|
| 176 |
|
| 177 |
+
def load_excel_other(xls_path: str, sheet_pref: str | None, target_cols: list[str]) -> pd.DataFrame | None:
|
|
|
|
|
|
|
|
|
|
| 178 |
"""
|
| 179 |
+
2枚目以降:
|
| 180 |
+
- skiprows=4, header=None でデータのみ
|
| 181 |
+
- 左端の列を削除
|
| 182 |
+
- 列数が合わなければ切り詰め/ダミー列追加で合わせる
|
| 183 |
+
- 列名を 1枚目の chosen に置換
|
| 184 |
"""
|
| 185 |
sheet = pick_sheet_name(xls_path, sheet_pref)
|
| 186 |
if not sheet:
|
|
|
|
| 193 |
skiprows=SKIP_ROWS_OTHERS,
|
| 194 |
dtype=str
|
| 195 |
)
|
| 196 |
+
# 左端の列を削除
|
| 197 |
df = df.iloc[:, 1:]
|
| 198 |
# 前後空白トリム
|
| 199 |
for c in df.select_dtypes(include=["object"]).columns:
|
| 200 |
df[c] = df[c].str.strip()
|
| 201 |
|
| 202 |
+
# 列数調整
|
| 203 |
if df.shape[1] != len(target_cols):
|
| 204 |
print(f"[WARN] 列数不一致: file={os.path.basename(xls_path)} "
|
| 205 |
f"read={df.shape[1]} vs target={len(target_cols)} -> 自動調整")
|
| 206 |
if df.shape[1] > len(target_cols):
|
| 207 |
df = df.iloc[:, :len(target_cols)]
|
| 208 |
else:
|
| 209 |
+
# 足りないときは None 列を追加
|
| 210 |
+
for k in range(len(target_cols) - df.shape[1]):
|
| 211 |
+
df[f"_pad_{k}"] = None
|
|
|
|
| 212 |
df = df.iloc[:, :len(target_cols)]
|
| 213 |
|
| 214 |
df.columns = target_cols
|
| 215 |
return df
|
| 216 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 217 |
def zip_paths(paths: list[str], out_zip: str) -> str:
|
| 218 |
with zipfile.ZipFile(out_zip, "w", compression=zipfile.ZIP_DEFLATED) as z:
|
| 219 |
for p in paths:
|
|
|
|
| 221 |
z.write(p, arcname=os.path.basename(p))
|
| 222 |
return out_zip
|
| 223 |
|
| 224 |
+
# -------------------- メイン実行(Gradioから呼ぶ) --------------------
|
| 225 |
|
| 226 |
+
def run_job(sheet_name, sleep_sec, limit, re_download, progress=gr.Progress(track_tqdm=False)):
|
| 227 |
progress(0, desc="初期化中…")
|
| 228 |
|
| 229 |
session = requests.Session()
|
| 230 |
session.headers.update({
|
| 231 |
+
"User-Agent": "Mozilla/5.0 (compatible; FITCollector/1.3; +https://huggingface.co/spaces)",
|
| 232 |
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
|
| 233 |
})
|
| 234 |
|
| 235 |
# 1) リンク収集
|
| 236 |
links = collect_pref_links(session)
|
| 237 |
if not links:
|
| 238 |
+
return ("都道府県ファイルのリンク検出に失敗しました。ページ構成の変更/一時的な制限の可能性があります。",
|
| 239 |
+
None, None, None, None)
|
| 240 |
if limit and limit > 0:
|
| 241 |
links = links[:int(limit)]
|
| 242 |
progress(0.1, desc=f"リンク検出 {len(links)} 件")
|
|
|
|
| 265 |
if not downloaded:
|
| 266 |
return ("ダウンロードに失敗しました。", None, None, None, None)
|
| 267 |
|
| 268 |
+
# 3) 読み込み(1枚目で列名確定)
|
| 269 |
+
progress(0.75, desc="1枚目を読み込み(列名を確定)")
|
| 270 |
first_path = downloaded[0]
|
| 271 |
try:
|
| 272 |
df0, cols0 = load_excel_first(first_path, sheet_name if sheet_name else None)
|
|
|
|
| 276 |
|
| 277 |
frames = [df0]
|
| 278 |
|
| 279 |
+
# 4) 読み込み(2枚目以降)
|
| 280 |
for j, p in enumerate(downloaded[1:], start=2):
|
| 281 |
progress(0.75 + 0.25 * (j - 1) / max(1, len(downloaded) - 1),
|
| 282 |
desc=f"{j}枚目を読み込み")
|
|
|
|
| 289 |
# 5) 縦結合
|
| 290 |
combined = pd.concat(frames, ignore_index=True)
|
| 291 |
|
| 292 |
+
# 6) 出力
|
|
|
|
|
|
|
|
|
|
|
|
|
| 293 |
os.makedirs(OUTDIR, exist_ok=True)
|
| 294 |
out_xlsx = os.path.join(OUTDIR, "combined_fit.xlsx")
|
| 295 |
out_parq = os.path.join(OUTDIR, "combined_fit.parquet")
|
|
|
|
| 297 |
combined.to_excel(w, index=False, sheet_name="combined")
|
| 298 |
combined.to_parquet(out_parq, index=False)
|
| 299 |
|
| 300 |
+
# 7) ZIP(取得ファイル一式)
|
| 301 |
raw_zip = os.path.join(OUTDIR, "raw_excels.zip")
|
| 302 |
zip_paths(downloaded, raw_zip)
|
| 303 |
|
| 304 |
+
# 8) プレビュー
|
| 305 |
preview_csv = os.path.join(OUTDIR, "combined_head.csv")
|
| 306 |
combined.head(1000).to_csv(preview_csv, index=False)
|
| 307 |
|
|
|
|
| 312 |
f"・Parquet: combined_fit.parquet\n"
|
| 313 |
f"・Raw ZIP: raw_excels.zip\n"
|
| 314 |
f"・プレビュー: combined_head.csv\n"
|
| 315 |
+
f"・列名は『小分類>中分類>大分類』の優先で単一行化(結合は不実施)"
|
| 316 |
)
|
| 317 |
return (msg, out_xlsx, out_parq, raw_zip, preview_csv)
|
| 318 |
|
| 319 |
+
# -------------------- Gradio UI --------------------
|
| 320 |
|
| 321 |
with gr.Blocks(title="FIT 公表(都道府県別Excel)一括取得&結合") as demo:
|
| 322 |
gr.Markdown(
|
| 323 |
"""
|
| 324 |
# FIT 公表(都道府県別Excel)一括取得 & 結合
|
| 325 |
+
**列名ポリシー**:
|
| 326 |
+
- 1枚目: 0行目を使わず、1/2/3行目をヘッダとして読み込み(3段)。
|
| 327 |
+
- 列名は **小分類に値があれば小分類、無ければ中分類のみ**(結合しません)。
|
| 328 |
+
- 2枚目以降: 0〜3行目をスキップし、データのみ読み込み。
|
| 329 |
+
- すべてのファイルで **左端の列は削除**。
|
| 330 |
+
- ファイル名/シート名などのメタ列は付与しません。
|
| 331 |
"""
|
| 332 |
)
|
| 333 |
with gr.Row():
|
|
|
|
| 336 |
with gr.Row():
|
| 337 |
limit = gr.Number(value=None, precision=0, label="先頭N県のみ(テスト用・空欄は全県)")
|
| 338 |
reget = gr.Checkbox(label="既存ファイルがあっても再ダウンロードする", value=False)
|
|
|
|
|
|
|
|
|
|
| 339 |
|
| 340 |
run_btn = gr.Button("実行", variant="primary")
|
| 341 |
out_msg = gr.Markdown()
|
|
|
|
| 346 |
|
| 347 |
run_btn.click(
|
| 348 |
fn=run_job,
|
| 349 |
+
inputs=[sheet, sleep, limit, reget],
|
| 350 |
outputs=[out_msg, out_xlsx, out_parq, out_zip, out_preview]
|
| 351 |
)
|
| 352 |
|