@chukyo-umebo/web_parser 0.0.1

package/README.md

@@ -0,0 +1,56 @@
+# @chukyo-passpal/web_parser
+
+中京大学の各種システム（Albo, Cubics, Manabo）のHTMLを解析し、構造化されたデータとして抽出するためのTypeScriptライブラリです。
+
+## 特徴
+
+- **型安全性**: Zodを使用して解析結果を検証し、TypeScriptの型定義を提供します。
+- **軽量**: `htmlparser2` と `css-select` を使用した高速な解析。
+- **モジュール化**: システムごとにパーサーが分かれています。
+
+## インストール
+
+```bash
+bun add @chukyo-passpal/web_parser
+# または
+npm install @chukyo-passpal/web_parser
+```
+
+## 使い方
+
+```typescript
+import { parseManaboNews } from "@chukyo-passpal/web_parser";
+
+const html = `...`; // 取得したHTML文字列
+const result = parseManaboNews(html);
+
+if (result.success) {
+  console.log(result.data);
+} else {
+  console.error(result.error);
+}
+```
+
+## 対応システム
+
+- **Albo**: アルバイト情報など
+- **Cubics**: 履修登録、時間割など
+- **Manabo**: ポータルサイト（お知らせ、課題、メールなど）
+
+## 開発
+
+このプロジェクトは [Bun](https://bun.sh) を使用しています。
+
+```bash
+# 依存関係のインストール
+bun install
+
+# ビルド
+bun run build
+
+# テスト
+bun test
+
+# リント
+bun run lint
+```

package/dist/cjs/albo/index.js

@@ -0,0 +1,5 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.parseCubicsPtNews = void 0;
+var albo_1 = require("./parser/albo");
+Object.defineProperty(exports, "parseCubicsPtNews", { enumerable: true, get: function () { return albo_1.parseCubicsPtNews; } });

package/dist/cjs/albo/parser/albo.js

@@ -0,0 +1,52 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.parseCubicsPtNews = void 0;
+const dom_1 = require("../../common/dom");
+const albo_1 = require("../types/albo");
+const utils_1 = require("../../common/utils");
+const parseNewsEntry = (row) => {
+    const categoryImg = (0, dom_1.queryOne)("th.category img", row);
+    const statusImg = (0, dom_1.queryOne)("td.status img", row);
+    const headlineAnchor = (0, dom_1.queryOne)("td.headline a", row);
+    if (!headlineAnchor) {
+        return null;
+    }
+    const title = (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)(headlineAnchor));
+    const onclick = (0, dom_1.getAttribute)(headlineAnchor, "onclick") ?? "";
+    const href = (0, dom_1.getAttribute)(headlineAnchor, "href");
+    const link = (0, utils_1.extractFirstQuotedValue)(onclick) ?? href ?? null;
+    return {
+        category: (0, dom_1.getAttribute)(categoryImg, "alt"),
+        status: (0, dom_1.getAttribute)(statusImg, "alt") ?? (0, dom_1.getAttribute)(statusImg, "src"),
+        title,
+        link,
+    };
+};
+const parseCubicsPtNews = (html) => {
+    const document = (0, dom_1.loadDocument)(html);
+    const tabs = (0, dom_1.queryAll)("#tab_element li", document).map((tab) => {
+        const anchor = (0, dom_1.queryOne)("a", tab);
+        const href = (0, dom_1.getAttribute)(anchor, "href") ?? "";
+        const targetId = href.replace(/^#/, "");
+        const title = (0, utils_1.normalizeWhitespace)(anchor ? (0, dom_1.getTextContent)(anchor) : "");
+        const rows = (0, dom_1.queryAll)(`#${targetId} table.tab-newslist tr`, document)
+            .map((row) => parseNewsEntry(row))
+            .filter((entry) => entry !== null);
+        return {
+            id: targetId,
+            title,
+            entries: rows,
+            active: (0, dom_1.elementHasClass)(tab, "selected"),
+        };
+    });
+    const selectedTab = tabs.find((tab) => tab.active);
+    return albo_1.CubicsPtNewsSchema.safeParse({
+        selectedTabId: selectedTab ? selectedTab.id : null,
+        tabs: tabs.map((tab) => ({
+            id: tab.id,
+            title: tab.title,
+            entries: tab.entries,
+        })),
+    });
+};
+exports.parseCubicsPtNews = parseCubicsPtNews;

package/dist/cjs/albo/types/albo.js

@@ -0,0 +1,19 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.CubicsPtNewsSchema = exports.CubicsPtNewsTabSchema = exports.CubicsPtNewsEntrySchema = void 0;
+const zod_1 = require("zod");
+exports.CubicsPtNewsEntrySchema = zod_1.z.object({
+    category: zod_1.z.string().nullable(),
+    status: zod_1.z.string().nullable(),
+    title: zod_1.z.string(),
+    link: zod_1.z.string().nullable(),
+});
+exports.CubicsPtNewsTabSchema = zod_1.z.object({
+    id: zod_1.z.string(),
+    title: zod_1.z.string(),
+    entries: zod_1.z.array(exports.CubicsPtNewsEntrySchema),
+});
+exports.CubicsPtNewsSchema = zod_1.z.object({
+    selectedTabId: zod_1.z.string().nullable(),
+    tabs: zod_1.z.array(exports.CubicsPtNewsTabSchema),
+});

package/dist/cjs/common/dom.js

@@ -0,0 +1,116 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.matches = exports.getChildrenElements = exports.filterElements = exports.isElement = exports.previousSiblingElement = exports.nextSiblingElement = exports.closest = exports.elementHasClass = exports.getInnerHtml = exports.getTextContent = exports.getAttribute = exports.queryOne = exports.queryAll = exports.loadDocument = void 0;
+const htmlparser2_1 = require("htmlparser2");
+const css_select_1 = require("css-select");
+const domutils_1 = require("domutils");
+const toArray = (root) => {
+    if (!root) {
+        return [];
+    }
+    if (Array.isArray(root)) {
+        return root;
+    }
+    if ("children" in root && Array.isArray(root.children)) {
+        return root.children;
+    }
+    return [root];
+};
+const loadDocument = (html) => (0, htmlparser2_1.parseDocument)(html);
+exports.loadDocument = loadDocument;
+const queryAll = (selector, root) => (0, css_select_1.selectAll)(selector, toArray(root));
+exports.queryAll = queryAll;
+const queryOne = (selector, root) => (0, css_select_1.selectOne)(selector, toArray(root));
+exports.queryOne = queryOne;
+const getAttribute = (element, name) => {
+    if (!element) {
+        return null;
+    }
+    const value = (0, domutils_1.getAttributeValue)(element, name);
+    return typeof value === "string" ? value : null;
+};
+exports.getAttribute = getAttribute;
+const getTextContent = (node) => {
+    if (!node) {
+        return "";
+    }
+    return (0, domutils_1.textContent)(node);
+};
+exports.getTextContent = getTextContent;
+const decodeNumericEntities = (value) => value
+    .replace(/&#x([0-9a-fA-F]+);/g, (match, hex) => {
+    const codePoint = Number.parseInt(hex, 16);
+    return Number.isNaN(codePoint) ? match : String.fromCodePoint(codePoint);
+})
+    .replace(/&#([0-9]+);/g, (match, dec) => {
+    const codePoint = Number.parseInt(dec, 10);
+    return Number.isNaN(codePoint) ? match : String.fromCodePoint(codePoint);
+});
+const getInnerHtml = (element) => {
+    if (!element) {
+        return "";
+    }
+    const inner = (0, domutils_1.getInnerHTML)(element);
+    const decoded = decodeNumericEntities(inner);
+    const normalised = decoded.replace(/\r\n/g, "\n").replace(/\r/g, "\n");
+    const withNbsp = normalised.replace(/\u00a0/g, " ");
+    return withNbsp.replace(/\salt(?!\s*=)(?=[\s/>])/g, ' alt=""');
+};
+exports.getInnerHtml = getInnerHtml;
+const elementHasClass = (element, className) => {
+    if (!element) {
+        return false;
+    }
+    const classAttr = (0, exports.getAttribute)(element, "class");
+    if (!classAttr) {
+        return false;
+    }
+    return classAttr.split(/\s+/).includes(className);
+};
+exports.elementHasClass = elementHasClass;
+const closest = (element, selector) => {
+    let current = element;
+    while (current) {
+        if ((0, css_select_1.is)(current, selector)) {
+            return current;
+        }
+        const parent = (0, domutils_1.getParent)(current);
+        current = parent && (0, domutils_1.isTag)(parent) ? parent : null;
+    }
+    return null;
+};
+exports.closest = closest;
+const nextSiblingElement = (element) => {
+    if (!element) {
+        return null;
+    }
+    const sibling = (0, domutils_1.nextElementSibling)(element);
+    return sibling ?? null;
+};
+exports.nextSiblingElement = nextSiblingElement;
+const previousSiblingElement = (element) => {
+    if (!element) {
+        return null;
+    }
+    const sibling = (0, domutils_1.prevElementSibling)(element);
+    return sibling ?? null;
+};
+exports.previousSiblingElement = previousSiblingElement;
+const isElement = (node) => !!node && (0, domutils_1.isTag)(node);
+exports.isElement = isElement;
+const filterElements = (nodes) => nodes.filter((node) => (0, domutils_1.isTag)(node));
+exports.filterElements = filterElements;
+const getChildrenElements = (element) => {
+    if (!element) {
+        return [];
+    }
+    return (0, exports.filterElements)(element.children ?? []);
+};
+exports.getChildrenElements = getChildrenElements;
+const matches = (element, selector) => {
+    if (!element) {
+        return false;
+    }
+    return (0, css_select_1.is)(element, selector);
+};
+exports.matches = matches;

package/dist/cjs/common/utils.js

@@ -0,0 +1,13 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.extractFirstQuotedValue = exports.normalizeWhitespace = void 0;
+const normalizeWhitespace = (value) => value.replace(/\s+/g, " ").trim();
+exports.normalizeWhitespace = normalizeWhitespace;
+const extractFirstQuotedValue = (raw) => {
+    if (!raw) {
+        return null;
+    }
+    const match = raw.match(/'([^']+)'/);
+    return match ? match[1] ?? null : null;
+};
+exports.extractFirstQuotedValue = extractFirstQuotedValue;

package/dist/cjs/cubics/index.js

@@ -0,0 +1,5 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.parseCubicsAsTimetable = void 0;
+var cubics_1 = require("./parser/cubics");
+Object.defineProperty(exports, "parseCubicsAsTimetable", { enumerable: true, get: function () { return cubics_1.parseCubicsAsTimetable; } });

package/dist/cjs/cubics/parser/cubics.js

@@ -0,0 +1,94 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.parseCubicsAsTimetable = void 0;
+const dom_1 = require("../../common/dom");
+const cubics_1 = require("../types/cubics");
+const utils_1 = require("../../common/utils");
+const parseCubicsAsTimetable = (html) => {
+    const document = (0, dom_1.loadDocument)(html);
+    const studentTable = (0, dom_1.queryOne)("table.output", document);
+    const studentRows = studentTable ? (0, dom_1.queryAll)("tr", studentTable) : [];
+    const row1Cells = studentRows[0] ? (0, dom_1.queryAll)("td", studentRows[0]) : [];
+    const row2Cells = studentRows[1] ? (0, dom_1.queryAll)("td", studentRows[1]) : [];
+    const row3Cells = studentRows[2] ? (0, dom_1.queryAll)("td", studentRows[2]) : [];
+    const row4Cells = studentRows[3] ? (0, dom_1.queryAll)("td", studentRows[3]) : [];
+    const getCellText = (cells, index) => (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)(cells[index] ?? null));
+    const student = {
+        id: getCellText(row1Cells, 0),
+        name: getCellText(row1Cells, 1),
+        division: getCellText(row2Cells, 0),
+        affiliation: getCellText(row2Cells, 1),
+        status: getCellText(row2Cells, 2),
+        className: getCellText(row2Cells, 3),
+        faculty: getCellText(row3Cells, 0),
+        department: getCellText(row3Cells, 1),
+        course: getCellText(row3Cells, 2),
+        address: getCellText(row4Cells, 0),
+    };
+    const periodInput = (0, dom_1.queryOne)("input[name='lblSpcfProd']", document);
+    const fallbackPeriod = (0, dom_1.queryOne)("div.searcharea ul li", document);
+    const periodRangeRaw = (0, dom_1.getAttribute)(periodInput, "value") ?? (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)(fallbackPeriod));
+    const periodRange = (0, utils_1.normalizeWhitespace)(periodRangeRaw);
+    const curriculumTable = (0, dom_1.queryOne)("table.output_curriculum", document);
+    const headerRow = curriculumTable ? (0, dom_1.queryOne)("tr", curriculumTable) : null;
+    const headerCells = headerRow ? (0, dom_1.queryAll)("th", headerRow).slice(1) : [];
+    const days = headerCells.map((element) => {
+        const text = (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)(element));
+        const segments = text.split(" ").filter((segment) => segment.length > 0);
+        return {
+            label: segments[0] ?? "",
+            date: segments[1] ?? "",
+        };
+    });
+    const periodRows = curriculumTable ? (0, dom_1.queryAll)("tr", curriculumTable).slice(1) : [];
+    const periods = periodRows.map((row) => {
+        const periodLabel = (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)((0, dom_1.queryOne)("th", row)));
+        const cells = (0, dom_1.queryAll)("td", row);
+        const slots = cells.map((cell) => {
+            const lessonCode = (0, dom_1.getAttribute)((0, dom_1.queryOne)('input[name$=".hdnLsnCd1"]', cell), "value");
+            const lessonYear = (0, dom_1.getAttribute)((0, dom_1.queryOne)('input[name$=".hdnLsnOpcFcy1"]', cell), "value");
+            const termCode = (0, dom_1.getAttribute)((0, dom_1.queryOne)('input[name$=".hdnTacTrmCd1"]', cell), "value");
+            const timetableCode = (0, dom_1.getAttribute)((0, dom_1.queryOne)('input[name$=".hdnTmtxCd"]', cell), "value");
+            const optionDate = (0, dom_1.getAttribute)((0, dom_1.queryOne)('input[name$=".hdnOpcDt"]', cell), "value");
+            const detailAnchor = (0, dom_1.queryAll)("a", cell).find((anchor) => ((0, dom_1.getAttribute)(anchor, "onclick") ?? "").includes("execPopupWindowOpen"));
+            const detailOnclick = detailAnchor ? (0, dom_1.getAttribute)(detailAnchor, "onclick") ?? "" : "";
+            const infoAnchors = (0, dom_1.queryAll)("a.chiphelp", cell);
+            const classroom = (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)(infoAnchors[0] ?? null));
+            const subject = (0, utils_1.normalizeWhitespace)((0, dom_1.getTextContent)(infoAnchors[1] ?? null));
+            const slot = {
+                classroom: classroom.length ? classroom : null,
+                subject: subject.length ? subject : null,
+                detailUrl: (0, utils_1.extractFirstQuotedValue)(detailOnclick),
+                lessonCode,
+                lessonYear,
+                termCode,
+                timetableCode,
+                optionDate,
+            };
+            if (!slot.classroom && !slot.subject && !slot.detailUrl && !slot.lessonCode) {
+                return {
+                    classroom: null,
+                    subject: null,
+                    detailUrl: null,
+                    lessonCode: null,
+                    lessonYear: null,
+                    termCode: null,
+                    timetableCode: null,
+                    optionDate: null,
+                };
+            }
+            return slot;
+        });
+        return {
+            periodLabel,
+            slots,
+        };
+    });
+    return cubics_1.CubicsAsTimetableSchema.safeParse({
+        student,
+        periodRange,
+        days,
+        periods,
+    });
+};
+exports.parseCubicsAsTimetable = parseCubicsAsTimetable;

package/dist/cjs/cubics/types/cubics.js

@@ -0,0 +1,38 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.CubicsAsTimetableSchema = exports.CubicsAsTimetablePeriodSchema = exports.CubicsAsTimetableSlotSchema = void 0;
+const zod_1 = require("zod");
+exports.CubicsAsTimetableSlotSchema = zod_1.z.object({
+    classroom: zod_1.z.string().nullable(),
+    subject: zod_1.z.string().nullable(),
+    detailUrl: zod_1.z.string().nullable(),
+    lessonCode: zod_1.z.string().nullable(),
+    lessonYear: zod_1.z.string().nullable(),
+    termCode: zod_1.z.string().nullable(),
+    timetableCode: zod_1.z.string().nullable(),
+    optionDate: zod_1.z.string().nullable(),
+});
+exports.CubicsAsTimetablePeriodSchema = zod_1.z.object({
+    periodLabel: zod_1.z.string(),
+    slots: zod_1.z.array(exports.CubicsAsTimetableSlotSchema),
+});
+exports.CubicsAsTimetableSchema = zod_1.z.object({
+    student: zod_1.z.object({
+        id: zod_1.z.string(),
+        name: zod_1.z.string(),
+        division: zod_1.z.string(),
+        affiliation: zod_1.z.string(),
+        status: zod_1.z.string(),
+        className: zod_1.z.string(),
+        faculty: zod_1.z.string(),
+        department: zod_1.z.string(),
+        course: zod_1.z.string(),
+        address: zod_1.z.string(),
+    }),
+    periodRange: zod_1.z.string(),
+    days: zod_1.z.array(zod_1.z.object({
+        label: zod_1.z.string(),
+        date: zod_1.z.string(),
+    })),
+    periods: zod_1.z.array(exports.CubicsAsTimetablePeriodSchema),
+});

package/dist/cjs/index.js

@@ -0,0 +1,19 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __exportStar = (this && this.__exportStar) || function(m, exports) {
+    for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+__exportStar(require("./albo/index"), exports);
+__exportStar(require("./cubics/index"), exports);
+__exportStar(require("./manabo/index"), exports);

package/dist/cjs/manabo/index.js

@@ -0,0 +1,23 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.parseManaboTimetable = exports.parseManaboNews = exports.parseManaboMailMember = exports.parseManaboMailSend = exports.parseManaboMailView = exports.parseManaboSentMail = exports.parseManaboReceivedMail = exports.parseManaboEntryResponse = exports.parseManaboEntryForm = exports.parseManaboClassQuizResult = exports.parseManaboClassSyllabus = exports.parseManaboClassNews = exports.parseManaboClassEntry = exports.parseManaboClassContent = exports.parseManaboClassDirectory = void 0;
+var manaboClass_1 = require("./parser/manaboClass");
+Object.defineProperty(exports, "parseManaboClassDirectory", { enumerable: true, get: function () { return manaboClass_1.parseManaboClassDirectory; } });
+Object.defineProperty(exports, "parseManaboClassContent", { enumerable: true, get: function () { return manaboClass_1.parseManaboClassContent; } });
+Object.defineProperty(exports, "parseManaboClassEntry", { enumerable: true, get: function () { return manaboClass_1.parseManaboClassEntry; } });
+Object.defineProperty(exports, "parseManaboClassNews", { enumerable: true, get: function () { return manaboClass_1.parseManaboClassNews; } });
+Object.defineProperty(exports, "parseManaboClassSyllabus", { enumerable: true, get: function () { return manaboClass_1.parseManaboClassSyllabus; } });
+Object.defineProperty(exports, "parseManaboClassQuizResult", { enumerable: true, get: function () { return manaboClass_1.parseManaboClassQuizResult; } });
+var manaboEntry_1 = require("./parser/manaboEntry");
+Object.defineProperty(exports, "parseManaboEntryForm", { enumerable: true, get: function () { return manaboEntry_1.parseManaboEntryForm; } });
+Object.defineProperty(exports, "parseManaboEntryResponse", { enumerable: true, get: function () { return manaboEntry_1.parseManaboEntryResponse; } });
+var manaboMail_1 = require("./parser/manaboMail");
+Object.defineProperty(exports, "parseManaboReceivedMail", { enumerable: true, get: function () { return manaboMail_1.parseManaboReceivedMail; } });
+Object.defineProperty(exports, "parseManaboSentMail", { enumerable: true, get: function () { return manaboMail_1.parseManaboSentMail; } });
+Object.defineProperty(exports, "parseManaboMailView", { enumerable: true, get: function () { return manaboMail_1.parseManaboMailView; } });
+Object.defineProperty(exports, "parseManaboMailSend", { enumerable: true, get: function () { return manaboMail_1.parseManaboMailSend; } });
+Object.defineProperty(exports, "parseManaboMailMember", { enumerable: true, get: function () { return manaboMail_1.parseManaboMailMember; } });
+var manaboNews_1 = require("./parser/manaboNews");
+Object.defineProperty(exports, "parseManaboNews", { enumerable: true, get: function () { return manaboNews_1.parseManaboNews; } });
+var manaboTimetable_1 = require("./parser/manaboTimetable");
+Object.defineProperty(exports, "parseManaboTimetable", { enumerable: true, get: function () { return manaboTimetable_1.parseManaboTimetable; } });