Webシラバスからデータをスクレイピングして整形

main.py

@@ -0,0 +1,133 @@
+import requests as req
+from bs4 import BeautifulSoup
+from tkinter import *
+from tkinter import ttk
+import time
+
+FETCH_SLEEP = 0
+room = ["特別教室", "階段教室", "ネットワーク情報センター", "第1講義室", "第2講義室", "第3講義室", "第4講義室",
+        "第5講義室", "第6講義室", "専攻科講義室A", "専攻科講義室B",
+        "専攻科ゼミ室A", "共通ゼミ室A", "マルチメディア講義室A", "マルチメディア講義室B", "多目的室A,B(講義棟C)",
+        "M製図室", "E実験室", "D実験室",
+        "実習工場", "J計算機演習室", "J回路実験室", "コミュニティールームB", "コミュミニティルームD", "CAD室",
+        "校内外フィールド", "C都市創造実験室・レクチャー室", "水理実験室",
+        "図書館フリー閲覧室", "学友会研修室1", "学友会館研修室2/3", "学友会館研修室4", "物理実験室", "化学実験室",
+        "ものづくり実習室", ""]
+departmentName = [
+    {
+        "name": "機械工学科",
+        "alphabet": 'M'
+    },
+    {
+        "name": "電気電子工学科",
+        "alphabet": 'E'
+    },
+    {
+        "name": "電子制御工学科",
+        "alphabet": 'D'
+    },
+    {
+        "name": "情報工学科",
+        "alphabet": 'J'
+    },
+    {
+        "name": "環境都市工学科",
+        "alphabet": 'C'
+    }
+]
+subjectData: list[list[dict]] = [[], [], [], [], []]
+
+
+def update_room(idx, select):
+    subjectData[idx][0]["room"] = select
+
+
+def generation_progress_bar(num):
+    result = "["
+    done = int(30 * num)
+    for count in range(done):
+        result += "#"
+    for count in range(30 - done):
+        result += "-"
+    result += "] "
+
+    if done == 30:
+        result += "done"
+    else:
+        result += str(num * 100 * 100 // 1 / 100) + "%"
+    return result
+
+
+print("------------------------------")
+print("KMM-timetable-crawler v0.0.0\ndeveloped by kokastar")
+print("------------------------------")
+
+print("年度を指定してください", end=":")
+year: str = input()
+
+subjectElements = []
+allSubjectSiz = 0
+print("\nfetching syllabus department data...")
+print(generation_progress_bar(0), end="")
+for i in range(5):
+    page = req.get("https://syllabus.kosen-k.go.jp/Pages/PublicSubjects?school_id=14&department_id=1" + str(
+        i + 1) + "&year=" + year + "&lang=ja")
+    soup = BeautifulSoup(page.text, "html.parser")
+    mcc_show = soup.find_all(class_="mcc-show")
+    subjectElements.append(mcc_show)
+    allSubjectSiz += len(mcc_show)
+    print("\r" + generation_progress_bar((i + 1) / 5), end="")
+    time.sleep(FETCH_SLEEP)
+
+fetchedSubject = 0
+print("\nfetching subjects data...")
+print(generation_progress_bar(0), end="")
+for i in range(5):
+    for j in range(len(subjectElements[i])):
+        try:
+            url = "https://syllabus.kosen-k.go.jp" + subjectElements[i][j].attrs['href']
+            page = req.get(url)
+            soup = BeautifulSoup(page.text, "html.parser")
+            tdData = soup.find_all("td")
+            subjectName = soup.find("h1").string
+            teachers = []
+            for k in soup.find_all(id="Teachers"):
+                teachers.append(k.string)
+            subjectData[i].append({"subjectName": subjectName, "subjectId": str(i + 1) + tdData[10].string,
+                                   "professor": teachers, "url": url, "room": ""})
+        except KeyError:
+            a = 0
+            # subjectPageData.append()
+        fetchedSubject += 1
+        print("\r" + generation_progress_bar(fetchedSubject / allSubjectSiz), end="")
+        time.sleep(FETCH_SLEEP)
+
+for i in range(len(subjectData)):
+    grade = int(subjectData[i][0]["subjectId"]) // 10
+    department = int(subjectData[i][0]["subjectId"]) % 10 - 1
+    className = str(grade) + "年" + departmentName[department]["name"] + "教室"
+    room[len(room) - 1] = className
+    update_room(i, className)
+
+    root = Tk()
+    root.title(className + " " + subjectData[i][0]["subjectName"])
+    # Frame
+    frame = ttk.Frame(root, padding=10)
+    frame.grid()
+    # Combobox
+    v = StringVar()
+    cb = ttk.Combobox(
+        frame, textvariable=v,
+        values=room, width=30, )
+    cb.set(room[len(room) - 1])
+    cb.grid(row=0, column=0)
+    # Button
+    button1 = ttk.Button(
+        frame, text='OK',
+        command=lambda: update_room(i, v.get()))
+    button2 = ttk.Button(
+        frame, text='close',
+        command=root.destroy)
+    button1.grid(row=0, column=1)
+    button2.grid(row=0, column=2)
+    root.mainloop()

requirement.txt

@@ -0,0 +1,7 @@
+beautifulsoup4==4.12.2
+certifi==2023.7.22
+charset-normalizer==3.3.0
+idna==3.4
+requests==2.31.0
+soupsieve==2.5
+urllib3==2.0.6