仙台市コロナ陽性患者情報2021/03/19以降

2021/03/19以降の一覧PDF

requirements.txt

https://www.city.sendai.jp/kenkoanzen-kansen/shise/koho/kisha/r3/documents/0319besshi.pdf
https://www.city.sendai.jp/kenkoanzen-kansen/shise/koho/kisha/r3/documents/0320_besshi.pdf
https://www.city.sendai.jp/kenkoanzen-kansen/shise/koho/kisha/r3/documents/0321besshi.pdf

ダウンロード

!wget -i requirements.txt -P download -w 1 –random-wait

import pdfplumber
import pandas as pd

import datetime
import pathlib

dfs = []

for p in pathlib.Path("download").glob("*.pdf"):

    with pdfplumber.open(p) as pdf:

        for page in pdf.pages:

            table = page.extract_table()

            df_tmp = pd.DataFrame(table[1:], columns=table[0]).rename(
                columns={"№": "No"}
            )
            df_tmp.columns = df_tmp.columns.str.replace("\n", "")
            df_tmp.set_index("No", inplace=True)

            dfs.append(df_tmp)

df = pd.concat(dfs).sort_index()

for col in df.select_dtypes(include=object).columns:
    df[col] = (
        df[col].str.strip().str.normalize("NFKC").str.replace("\n", "", regex=True)
    )

# 日付変換
dt_now = datetime.datetime.now()


def str2date(s: pd.Series) -> pd.Series:

    df = (
        s.str.extract("(\d{1,2})/(\d{1,2})")
        .rename(columns={0: "month", 1: "day"})
        .fillna(0)
        .astype(int)
    )

    df["year"] = dt_now.year

    tmp = pd.to_datetime(df, errors="coerce")

    df["year"] = df["year"].mask(tmp > dt_now, df["year"] - 1)

    return pd.to_datetime(df, errors="coerce")


df["発症日YMD"] = str2date(df["発症日"])
df["確認日YMD"] = str2date(df["陽性判明"])

df.to_csv("sendai2.csv", encoding="utf_8_sig")