refactor: reorganize client transport and extension internals
- Split client, native, remote, serve, markdown, and SDK internals into focused packages with direct imports. - Move local and remote transport framing/protocol helpers behind clearer module boundaries. - Break up the extension injected DOM logic into a separate content dispatch bundle and dedicated content modules. - Add explicit client handling for passive remote discovery without noisy PQ warnings. - Keep behavior covered with updated unit, integration, and extension tests.
This commit is contained in:
@@ -9,6 +9,11 @@ import type { SessionCommands } from '../commands/session';
|
||||
import type { ControlMessage, ResponseMessage, IncomingMessage, PageRequest, DispatchArgs, Serializable } from '../types';
|
||||
|
||||
const NATIVE_HOST = "com.browsercli.host";
|
||||
const DEBUG_LOG = false;
|
||||
|
||||
function debugLog(...args: Serializable[]) {
|
||||
if (DEBUG_LOG) console.log("[browser-cli]", ...args);
|
||||
}
|
||||
|
||||
export class NativeConnection {
|
||||
private port: chrome.runtime.Port | null = null;
|
||||
@@ -96,7 +101,7 @@ export class NativeConnection {
|
||||
// Send hello so native host knows which profile/alias this is
|
||||
const alias = await getProfileAlias();
|
||||
nativePort.postMessage({ type: "hello", alias });
|
||||
console.log("[browser-cli] Connected to native host as profile:", alias);
|
||||
debugLog("Connected to native host as profile:", alias);
|
||||
} catch (e) {
|
||||
this.port = null;
|
||||
console.error("[browser-cli] Failed to connect:", e);
|
||||
@@ -112,7 +117,7 @@ export class NativeConnection {
|
||||
// on the captured port and bail if it (or this.port) is already gone.
|
||||
const replyPort = this.port;
|
||||
|
||||
console.log("[browser-cli] ←", command, args);
|
||||
debugLog("←", command, args);
|
||||
|
||||
let data: Serializable, error: string | undefined;
|
||||
try {
|
||||
@@ -131,10 +136,10 @@ export class NativeConnection {
|
||||
}
|
||||
|
||||
if (error !== undefined) {
|
||||
console.log("[browser-cli] → ERROR", command, error);
|
||||
debugLog("→ ERROR", command, error);
|
||||
this.sendResponse(replyPort, { id, success: false, error });
|
||||
} else {
|
||||
console.log("[browser-cli] →", command, data);
|
||||
debugLog("→", command, data);
|
||||
this.sendResponse(replyPort, { id, success: true, data });
|
||||
}
|
||||
|
||||
|
||||
@@ -1,5 +1,4 @@
|
||||
import { assertScriptableUrl, executeScript, fetchTabHtml, isBrowserErrorUrl, isErrorPageScriptError, resolveTabUrl } from '../core';
|
||||
import { contentDispatch } from './injected';
|
||||
import { CommandGroup } from '../classes/CommandGroup';
|
||||
import type { CommandEntry } from '../classes/CommandGroup';
|
||||
import type { DomArgs, DomEvalArgs, DomWaitForArgs, DomPollArgs, Serializable } from '../types';
|
||||
@@ -74,9 +73,17 @@ export class DomCommands extends CommandGroup {
|
||||
}
|
||||
assertScriptableUrl(tabUrl, "run DOM commands on");
|
||||
try {
|
||||
await executeScript({
|
||||
target: { tabId: tab.id },
|
||||
files: ["content-dispatch.js"],
|
||||
});
|
||||
const results = await executeScript({
|
||||
target: { tabId: tab.id },
|
||||
func: contentDispatch,
|
||||
func: (name: string, opArgs: DomArgs) => {
|
||||
const dispatch = globalThis.browserCliContentDispatch;
|
||||
if (!dispatch) throw new Error("browser-cli content dispatcher was not injected");
|
||||
return dispatch(name, opArgs);
|
||||
},
|
||||
args: [funcName, args],
|
||||
});
|
||||
return results[0]?.result;
|
||||
|
||||
@@ -1,567 +1,4 @@
|
||||
import type { ContentArgs } from '../types';
|
||||
|
||||
export function contentDispatch(funcName: string, args: ContentArgs) {
|
||||
function domQuery({ selector }: ContentArgs) {
|
||||
return Array.from(document.querySelectorAll(selector)).map(el => ({
|
||||
tag: el.tagName.toLowerCase(),
|
||||
text: el.textContent.trim().slice(0, 200),
|
||||
attrs: Object.fromEntries(Array.from(el.attributes).map(a => [a.name, a.value])),
|
||||
}));
|
||||
}
|
||||
function domClick({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.click();
|
||||
return true;
|
||||
}
|
||||
function domType({ selector, text }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLInputElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.focus();
|
||||
el.value = text;
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
function domAttr({ selector, attr }: ContentArgs) {
|
||||
return Array.from(document.querySelectorAll(selector))
|
||||
.map(el => el.getAttribute(attr))
|
||||
.filter(v => v !== null);
|
||||
}
|
||||
function domText({ selector }: ContentArgs) {
|
||||
return Array.from(document.querySelectorAll(selector))
|
||||
.map(el => el.textContent.trim())
|
||||
.filter(Boolean);
|
||||
}
|
||||
function domExists({ selector }: ContentArgs) {
|
||||
return document.querySelector(selector) !== null;
|
||||
}
|
||||
function domKey({ selector, key }: ContentArgs) {
|
||||
const el = selector ? document.querySelector(selector) : document.activeElement;
|
||||
if (selector && !el) throw new Error(`No element: ${selector}`);
|
||||
const target = el || document.body;
|
||||
["keydown", "keypress", "keyup"].forEach(type => {
|
||||
target.dispatchEvent(new KeyboardEvent(type, { key, bubbles: true, cancelable: true }));
|
||||
});
|
||||
return true;
|
||||
}
|
||||
function domHover({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.dispatchEvent(new MouseEvent("mouseover", { bubbles: true }));
|
||||
el.dispatchEvent(new MouseEvent("mouseenter", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
function domCheck({ selector, checked }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLInputElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.checked = checked;
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
function domClear({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLInputElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.value = "";
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
function domFocus({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.focus();
|
||||
return true;
|
||||
}
|
||||
function domSubmit({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
const form = (el.tagName === "FORM" ? el : el.closest("form")) as HTMLFormElement | null;
|
||||
if (!form) throw new Error(`No form found for: ${selector}`);
|
||||
form.submit();
|
||||
return true;
|
||||
}
|
||||
function pageInfo() {
|
||||
const metas: Record<string, string> = {};
|
||||
document.querySelectorAll("meta[name], meta[property]").forEach(m => {
|
||||
const k = m.getAttribute("name") || m.getAttribute("property");
|
||||
if (k) metas[k] = m.getAttribute("content") || "";
|
||||
});
|
||||
return {
|
||||
title: document.title,
|
||||
url: location.href,
|
||||
readyState: document.readyState,
|
||||
lang: document.documentElement.lang || null,
|
||||
meta: metas,
|
||||
};
|
||||
}
|
||||
function domScroll({ selector, x, y }: ContentArgs) {
|
||||
if (selector) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.scrollIntoView({ behavior: "smooth", block: "center" });
|
||||
return true;
|
||||
}
|
||||
window.scrollTo({ top: y || 0, left: x || 0, behavior: "smooth" });
|
||||
return true;
|
||||
}
|
||||
function domSelect({ selector, value }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLSelectElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.value = value;
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
function extractLinks() {
|
||||
const seen = new Set<string>();
|
||||
return Array.from(document.querySelectorAll("a[href]")).reduce<Array<{ text: string; href: string }>>((links, a) => {
|
||||
const href = (a as HTMLAnchorElement).href;
|
||||
if (!href || seen.has(href)) return links;
|
||||
seen.add(href);
|
||||
links.push({
|
||||
text: a.textContent.trim().slice(0, 100),
|
||||
href,
|
||||
});
|
||||
return links;
|
||||
}, []);
|
||||
}
|
||||
function extractImages() {
|
||||
const seen = new Set<string>();
|
||||
return Array.from(document.querySelectorAll("img")).reduce<Array<{ alt: string; src: string }>>((images, img) => {
|
||||
const src =
|
||||
img.src ||
|
||||
img.getAttribute("data-src") ||
|
||||
img.getAttribute("data-lazy-src") ||
|
||||
img.getAttribute("data-original") ||
|
||||
(img.srcset ? img.srcset.split(",")[0].trim().split(" ")[0] : "") ||
|
||||
"";
|
||||
if (!src || seen.has(src)) return images;
|
||||
seen.add(src);
|
||||
const FAKE_ALT = new Set(["true", "false", "null", "undefined", "image", "img"]);
|
||||
const alt = img.alt && !FAKE_ALT.has(img.alt.trim().toLowerCase()) ? img.alt.trim() : "";
|
||||
images.push({ alt, src });
|
||||
return images;
|
||||
}, []);
|
||||
}
|
||||
function extractText() {
|
||||
return document.body.innerText;
|
||||
}
|
||||
function extractJson({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
return JSON.parse(el.textContent);
|
||||
}
|
||||
function extractMarkdown({ selector }: ContentArgs) {
|
||||
const BLOCKS = new Set([
|
||||
"article", "aside", "blockquote", "body", "div", "dl", "fieldset", "figcaption",
|
||||
"figure", "footer", "form", "h1", "h2", "h3", "h4", "h5", "h6", "header", "hr",
|
||||
"li", "main", "nav", "ol", "p", "pre", "section", "table", "tbody", "td", "tfoot",
|
||||
"th", "thead", "tr", "ul"
|
||||
]);
|
||||
const NOISE_SELECTOR = [
|
||||
"script",
|
||||
"style",
|
||||
"noscript",
|
||||
"template",
|
||||
"svg",
|
||||
"canvas",
|
||||
"iframe",
|
||||
"dialog",
|
||||
"button",
|
||||
"input",
|
||||
"textarea",
|
||||
"select",
|
||||
"option",
|
||||
"form",
|
||||
"[hidden]",
|
||||
"[aria-hidden='true']",
|
||||
".sr-only",
|
||||
"[class*='sr-only']",
|
||||
"[class*='file-tile']",
|
||||
"form[data-type='unified-composer']",
|
||||
".composer-btn",
|
||||
"[data-composer-surface='true']",
|
||||
"#thread-bottom-container",
|
||||
"[data-testid*='action-button']",
|
||||
].join(", ");
|
||||
|
||||
function normalizeText(value: string) {
|
||||
return value.replace(/\s+/g, " ").trim();
|
||||
}
|
||||
|
||||
function normalizeInline(value: string) {
|
||||
return value
|
||||
.replace(/[ \t]+\n/g, "\n")
|
||||
.replace(/\n[ \t]+/g, "\n")
|
||||
.replace(/\n{3,}/g, "\n\n")
|
||||
.replace(/[ \t]{2,}/g, " ")
|
||||
.trim();
|
||||
}
|
||||
|
||||
function collapseBlankLines(value: string) {
|
||||
return value
|
||||
.replace(/[ \t]+\n/g, "\n")
|
||||
.replace(/\n{3,}/g, "\n\n")
|
||||
.trim();
|
||||
}
|
||||
|
||||
function escapeMarkdown(text: string) {
|
||||
return text.replace(/([\\`[\]])/g, "\\$1");
|
||||
}
|
||||
|
||||
function escapeTableCell(text: string) {
|
||||
return text.replace(/\|/g, "\\|").replace(/\n+/g, " ").trim();
|
||||
}
|
||||
|
||||
function absoluteUrl(attr: string | null | undefined, fallback?: string) {
|
||||
return attr || fallback || "";
|
||||
}
|
||||
|
||||
function isNoiseElement(node: Node | null): boolean {
|
||||
if (!node || node.nodeType !== Node.ELEMENT_NODE) return false;
|
||||
const el = node as Element;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (["script", "style", "noscript", "template", "svg", "canvas", "iframe", "dialog"].includes(tag)) return true;
|
||||
if (["button", "input", "textarea", "select", "option", "form"].includes(tag)) return true;
|
||||
if (el.hasAttribute("hidden")) return true;
|
||||
if ((el.getAttribute("aria-hidden") || "").toLowerCase() === "true") return true;
|
||||
if (el.matches(".sr-only, [class*='sr-only']")) return true;
|
||||
if (el.matches("[class*='file-tile'], form[data-type='unified-composer'], .composer-btn, [data-composer-surface='true'], #thread-bottom-container")) return true;
|
||||
if (el.matches("[data-testid*='action-button']")) return true;
|
||||
return false;
|
||||
}
|
||||
|
||||
function stripNoise(root: Element): Element {
|
||||
const clone = root.cloneNode(true) as Element;
|
||||
clone.querySelectorAll(NOISE_SELECTOR).forEach(node => node.remove());
|
||||
return clone;
|
||||
}
|
||||
|
||||
function candidateScore(node: Element) {
|
||||
const text = normalizeText((node as HTMLElement).innerText || "");
|
||||
if (!text) return -Infinity;
|
||||
|
||||
const headings = node.querySelectorAll("h1, h2, h3, h4, h5, h6").length;
|
||||
const paragraphs = node.querySelectorAll("p").length;
|
||||
const listItems = node.querySelectorAll("li").length;
|
||||
const tables = node.querySelectorAll("table").length;
|
||||
const codeBlocks = node.querySelectorAll("pre, code").length;
|
||||
const images = node.querySelectorAll("img, figure").length;
|
||||
const mainLike = node.matches("main, article, [role='main']") ? 1 : 0;
|
||||
const proseBlocks = node.matches(".markdown, .prose, [data-message-author-role='assistant']") ? 1 : 0;
|
||||
const buttons = node.querySelectorAll("button, input, textarea, select").length;
|
||||
const forms = node.querySelectorAll("form").length;
|
||||
const svgs = node.querySelectorAll("svg, canvas").length;
|
||||
|
||||
return text.length
|
||||
+ (mainLike * 4000)
|
||||
+ (proseBlocks * 5000)
|
||||
+ (headings * 250)
|
||||
+ (paragraphs * 60)
|
||||
+ (listItems * 35)
|
||||
+ (tables * 80)
|
||||
+ (codeBlocks * 60)
|
||||
+ (images * 25)
|
||||
- (buttons * 120)
|
||||
- (forms * 200)
|
||||
- (svgs * 40);
|
||||
}
|
||||
|
||||
function pickRoot() {
|
||||
if (selector) {
|
||||
const matched = document.querySelector(selector);
|
||||
if (!matched) throw new Error(`No element: ${selector}`);
|
||||
return matched;
|
||||
}
|
||||
|
||||
const candidates = Array.from(document.querySelectorAll(
|
||||
"main, article, [role='main'], section, .markdown, .prose, [data-message-author-role]"
|
||||
))
|
||||
.filter(node => normalizeText((node as HTMLElement).innerText || "").length > 0);
|
||||
if (!candidates.length) return document.body;
|
||||
candidates.sort((a, b) => candidateScore(b) - candidateScore(a));
|
||||
return candidates[0];
|
||||
}
|
||||
|
||||
function inlineText(node: Node): string {
|
||||
if (node.nodeType === Node.TEXT_NODE) {
|
||||
return escapeMarkdown(node.textContent || "");
|
||||
}
|
||||
if (node.nodeType !== Node.ELEMENT_NODE) return "";
|
||||
if (isNoiseElement(node)) return "";
|
||||
|
||||
const el = node as HTMLElement;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (tag === "br") return "\n";
|
||||
if (tag === "img") {
|
||||
const img = el as HTMLImageElement;
|
||||
const src = absoluteUrl(img.getAttribute("src"), img.src);
|
||||
if (!src) return "";
|
||||
const alt = normalizeText(img.getAttribute("alt") || "");
|
||||
return alt ? `` : ``;
|
||||
}
|
||||
if (tag === "a") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
const href = absoluteUrl(el.getAttribute("href"), (el as HTMLAnchorElement).href);
|
||||
if (!href) return text;
|
||||
return `[${text || href}](${href})`;
|
||||
}
|
||||
if (tag === "code") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
return text ? `\`${text.replace(/`/g, "\\`")}\`` : "";
|
||||
}
|
||||
if (tag === "strong" || tag === "b") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
return text ? `**${text}**` : "";
|
||||
}
|
||||
if (tag === "em" || tag === "i") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
return text ? `*${text}*` : "";
|
||||
}
|
||||
|
||||
const chunks: string[] = [];
|
||||
for (const child of el.childNodes) {
|
||||
const rendered = inlineText(child);
|
||||
if (!rendered) continue;
|
||||
chunks.push(rendered);
|
||||
if (child.nodeType === Node.ELEMENT_NODE && BLOCKS.has((child as Element).tagName.toLowerCase())) {
|
||||
chunks.push("\n");
|
||||
}
|
||||
}
|
||||
return chunks.join("");
|
||||
}
|
||||
|
||||
function textBlock(node: Node): string {
|
||||
return collapseBlankLines(normalizeInline(Array.from(node.childNodes).map(inlineText).join("")));
|
||||
}
|
||||
|
||||
function preserveNodeText(node: Node): string {
|
||||
if (node.nodeType === Node.TEXT_NODE) {
|
||||
return node.textContent || "";
|
||||
}
|
||||
if (node.nodeType !== Node.ELEMENT_NODE) return "";
|
||||
|
||||
const el = node as HTMLElement;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (tag === "br") return "\n";
|
||||
|
||||
const parts: string[] = [];
|
||||
for (const child of el.childNodes) {
|
||||
const rendered = preserveNodeText(child);
|
||||
if (!rendered) continue;
|
||||
parts.push(rendered);
|
||||
}
|
||||
|
||||
if (["div", "p", "li"].includes(tag)) {
|
||||
return `${parts.join("")}\n`;
|
||||
}
|
||||
return parts.join("");
|
||||
}
|
||||
|
||||
function repairFlattenedDiagram(text: string): string {
|
||||
if (text.includes("\n")) return text;
|
||||
const markerCount = (text.match(/[│▼├└]/g) || []).length;
|
||||
if (markerCount < 2) return text;
|
||||
|
||||
let repaired = text;
|
||||
repaired = repaired.replace(/\s{2,}([│▼])/g, "\n $1");
|
||||
repaired = repaired.replace(/([│▼])\s{2,}/g, "$1\n");
|
||||
repaired = repaired.replace(/([│▼])(?=[^\s\n│▼├└])/g, "$1\n");
|
||||
repaired = repaired.replace(/(?<=[^\s\n])([├└])/g, "\n$1");
|
||||
repaired = repaired.replace(/([^\s\n])(\()/g, "$1\n$2");
|
||||
return repaired
|
||||
.split("\n")
|
||||
.map(line => line.replace(/\s+$/, ""))
|
||||
.filter(line => line.trim())
|
||||
.join("\n");
|
||||
}
|
||||
|
||||
function convertDashListsToBranches(lines: string[]): string[] {
|
||||
const converted: string[] = [];
|
||||
let index = 0;
|
||||
while (index < lines.length) {
|
||||
const match = lines[index].match(/^(\s*)-\s+(.*)$/);
|
||||
if (!match) {
|
||||
converted.push(lines[index]);
|
||||
index += 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
const indent = match[1];
|
||||
const items = [];
|
||||
while (index < lines.length) {
|
||||
const nextMatch = lines[index].match(new RegExp(`^${indent.replace(/[.*+?^${}()|[\]\\]/g, "\\$&")}-\\s+(.*)$`));
|
||||
if (!nextMatch) break;
|
||||
items.push(nextMatch[1]);
|
||||
index += 1;
|
||||
}
|
||||
|
||||
items.forEach((item, itemIndex) => {
|
||||
const branch = itemIndex === items.length - 1 ? "└" : "├";
|
||||
converted.push(`${indent}${branch} ${item}`);
|
||||
});
|
||||
}
|
||||
return converted;
|
||||
}
|
||||
|
||||
function normalizeCodeBlock(text: string): string {
|
||||
let lines = text.replace(/\r\n?/g, "\n").split("\n").map(line => line.replace(/\s+$/, ""));
|
||||
while (lines.length && !lines[0].trim()) lines.shift();
|
||||
while (lines.length && !lines[lines.length - 1].trim()) lines.pop();
|
||||
|
||||
const flattened = repairFlattenedDiagram(lines.join("\n"));
|
||||
lines = flattened ? flattened.split("\n") : [];
|
||||
lines = lines.map(line => {
|
||||
const trimmed = line.trim();
|
||||
if ((trimmed === "│" || trimmed === "▼") && !/^\s+[│▼]\s*$/.test(line)) {
|
||||
return ` ${trimmed}`;
|
||||
}
|
||||
return line;
|
||||
});
|
||||
lines = convertDashListsToBranches(lines);
|
||||
return lines.join("\n");
|
||||
}
|
||||
|
||||
function tableToMarkdown(table: Element) {
|
||||
const rows = Array.from(table.querySelectorAll("tr"))
|
||||
.map(row => Array.from(row.children)
|
||||
.filter(cell => cell.tagName === "TD" || cell.tagName === "TH")
|
||||
.map(cell => escapeTableCell(textBlock(cell)))
|
||||
)
|
||||
.filter(cells => cells.length > 0);
|
||||
if (!rows.length) return "";
|
||||
|
||||
const widths = rows.reduce((max, row) => Math.max(max, row.length), 0);
|
||||
const normalizedRows = rows.map(row => {
|
||||
const next = row.slice();
|
||||
while (next.length < widths) next.push("");
|
||||
return next;
|
||||
});
|
||||
|
||||
let headers = normalizedRows[0];
|
||||
let bodyRows = normalizedRows.slice(1);
|
||||
const firstRowIsBlank = headers.every(cell => !cell.trim());
|
||||
if (firstRowIsBlank && normalizedRows.length > 1) {
|
||||
headers = normalizedRows[1];
|
||||
bodyRows = normalizedRows.slice(2);
|
||||
}
|
||||
|
||||
const firstRow = table.querySelector("tr");
|
||||
const thead = table.querySelector("thead");
|
||||
const firstRowHasTh = firstRow && Array.from(firstRow.children).some(cell => cell.tagName === "TH");
|
||||
if (!(thead || firstRowHasTh || firstRowIsBlank)) {
|
||||
headers = new Array(widths).fill("");
|
||||
bodyRows = normalizedRows;
|
||||
}
|
||||
|
||||
const separator = new Array(widths).fill("---");
|
||||
const lines = [
|
||||
`| ${headers.join(" | ")} |`,
|
||||
`| ${separator.join(" | ")} |`,
|
||||
];
|
||||
for (const row of bodyRows) {
|
||||
lines.push(`| ${row.join(" | ")} |`);
|
||||
}
|
||||
return lines.join("\n");
|
||||
}
|
||||
|
||||
function listToMarkdown(list: Element, depth = 0): string {
|
||||
const ordered = list.tagName.toLowerCase() === "ol";
|
||||
const items: string[] = [];
|
||||
const children = Array.from(list.children).filter(child => child.tagName === "LI");
|
||||
children.forEach((item, index) => {
|
||||
const marker = ordered ? `${index + 1}. ` : "- ";
|
||||
const indent = " ".repeat(depth);
|
||||
const nested: string[] = [];
|
||||
const content: string[] = [];
|
||||
|
||||
for (const child of item.childNodes) {
|
||||
const childEl = child as Element;
|
||||
if (child.nodeType === Node.ELEMENT_NODE && (childEl.tagName === "UL" || childEl.tagName === "OL")) {
|
||||
nested.push(listToMarkdown(childEl, depth + 1));
|
||||
} else {
|
||||
content.push(inlineText(child));
|
||||
}
|
||||
}
|
||||
|
||||
const line = collapseBlankLines(normalizeInline(content.join("")));
|
||||
if (line) {
|
||||
const lineParts = line.split("\n");
|
||||
items.push(`${indent}${marker}${lineParts[0]}`);
|
||||
const continuationIndent = `${indent}${" ".repeat(marker.length)}`;
|
||||
lineParts.slice(1).forEach(part => items.push(`${continuationIndent}${part}`));
|
||||
}
|
||||
nested.filter(Boolean).forEach(block => items.push(block));
|
||||
});
|
||||
return items.join("\n");
|
||||
}
|
||||
|
||||
function blockToMarkdown(node: Node): string {
|
||||
if (node.nodeType === Node.TEXT_NODE) {
|
||||
return normalizeText(node.textContent || "");
|
||||
}
|
||||
if (node.nodeType !== Node.ELEMENT_NODE) return "";
|
||||
if (isNoiseElement(node)) return "";
|
||||
|
||||
const el = node as HTMLElement;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (tag === "table") return tableToMarkdown(el);
|
||||
if (tag === "ul" || tag === "ol") return listToMarkdown(el);
|
||||
if (el.matches(".cm-editor[data-is-code-block-view='true']")) {
|
||||
const lines = Array.from(el.querySelectorAll(".cm-line")).map(line => {
|
||||
const text = preserveNodeText(line);
|
||||
return text === "\n" ? "" : text.replace(/\n$/, "");
|
||||
});
|
||||
const code = normalizeCodeBlock(lines.join("\n"));
|
||||
return code ? `\`\`\`\n${code}\n\`\`\`` : "";
|
||||
}
|
||||
if (tag === "pre") {
|
||||
const code = normalizeCodeBlock(preserveNodeText(el));
|
||||
return code ? `\`\`\`\n${code}\n\`\`\`` : "";
|
||||
}
|
||||
if (tag === "blockquote") {
|
||||
const content = collapseBlankLines(Array.from(el.childNodes).map(blockToMarkdown).join("\n\n"));
|
||||
return content
|
||||
.split("\n")
|
||||
.map(line => line ? `> ${line}` : ">")
|
||||
.join("\n");
|
||||
}
|
||||
if (/^h[1-6]$/.test(tag)) {
|
||||
const level = Number(tag.slice(1));
|
||||
const text = textBlock(el);
|
||||
return text ? `${"#".repeat(level)} ${text}` : "";
|
||||
}
|
||||
if (tag === "p" || tag === "figcaption") {
|
||||
return textBlock(el);
|
||||
}
|
||||
if (tag === "hr") {
|
||||
return "---";
|
||||
}
|
||||
if (tag === "img") {
|
||||
return inlineText(el);
|
||||
}
|
||||
|
||||
const childBlocks = Array.from(el.childNodes)
|
||||
.map(child => blockToMarkdown(child))
|
||||
.filter(Boolean);
|
||||
if (childBlocks.length) return collapseBlankLines(childBlocks.join("\n\n"));
|
||||
|
||||
return textBlock(node);
|
||||
}
|
||||
|
||||
const root = stripNoise(pickRoot());
|
||||
const markdown = blockToMarkdown(root);
|
||||
return collapseBlankLines(markdown);
|
||||
}
|
||||
|
||||
const fns = { domQuery, domClick, domType, domAttr, domText, domExists,
|
||||
domScroll, domSelect, domKey, domHover, domCheck, domClear, domFocus, domSubmit,
|
||||
pageInfo,
|
||||
extractLinks, extractImages, extractText, extractJson, extractMarkdown };
|
||||
const fn = fns[funcName as keyof typeof fns];
|
||||
if (!fn) throw new Error(`Unknown content function: ${funcName}`);
|
||||
return fn(args);
|
||||
}
|
||||
|
||||
// ── Session ───────────────────────────────────────────────────────────────────
|
||||
// In-page DOM handlers live under ../content and are bundled into
|
||||
// extension/content-dispatch.js. Keep this file as a compatibility export for
|
||||
// tests or external imports; production DOM commands inject the content bundle.
|
||||
export { contentDispatch } from '../content/dispatch';
|
||||
|
||||
@@ -51,7 +51,16 @@ export class NavigationCommands extends CommandGroup {
|
||||
|
||||
private async navTo({ tabId, url }: NavToArgs) {
|
||||
const tab = await chrome.tabs.update(tabId, { url });
|
||||
return { id: tab.id, url: tab.url || url };
|
||||
const deadline = Date.now() + 1000;
|
||||
while (tabId && Date.now() < deadline) {
|
||||
const current = await chrome.tabs.get(tabId);
|
||||
const currentUrl = current.url || current.pendingUrl || "";
|
||||
if (currentUrl === url || currentUrl.startsWith(url)) {
|
||||
return { id: current.id, url: currentUrl };
|
||||
}
|
||||
await new Promise(r => setTimeout(r, 50));
|
||||
}
|
||||
return { id: tab.id, url: tab.url || tab.pendingUrl || url };
|
||||
}
|
||||
|
||||
private async navReload({ tabId }: NavTabArgs, bypassCache: boolean) {
|
||||
|
||||
@@ -23,12 +23,15 @@ export class TabsMutationCommands extends CommandGroup {
|
||||
return runLargeOperation("tabs.close", async () => {
|
||||
let toClose: number[] = [];
|
||||
if (duplicates) {
|
||||
const all = await chrome.tabs.query({});
|
||||
const seen = new Set();
|
||||
for (const t of all) {
|
||||
if (!t.url) continue;
|
||||
if (seen.has(t.url)) toClose.push(t.id);
|
||||
else seen.add(t.url);
|
||||
const windows = await chrome.windows.getAll({ populate: true });
|
||||
const seen = new Set<string>();
|
||||
for (const w of windows) {
|
||||
for (const t of w.tabs || []) {
|
||||
const url = t.url || t.pendingUrl;
|
||||
if (!url || t.id == null) continue;
|
||||
if (seen.has(url)) toClose.push(t.id);
|
||||
else seen.add(url);
|
||||
}
|
||||
}
|
||||
} else if (inactive) {
|
||||
const all = await chrome.tabs.query({});
|
||||
|
||||
@@ -0,0 +1,7 @@
|
||||
import { contentDispatch } from './content/dispatch';
|
||||
|
||||
declare global {
|
||||
var browserCliContentDispatch: typeof contentDispatch | undefined;
|
||||
}
|
||||
|
||||
globalThis.browserCliContentDispatch = contentDispatch;
|
||||
@@ -0,0 +1,51 @@
|
||||
import type { ContentArgs } from '../types';
|
||||
import {
|
||||
domAttr,
|
||||
domCheck,
|
||||
domClear,
|
||||
domClick,
|
||||
domExists,
|
||||
domFocus,
|
||||
domHover,
|
||||
domKey,
|
||||
domQuery,
|
||||
domScroll,
|
||||
domSelect,
|
||||
domSubmit,
|
||||
domText,
|
||||
domType,
|
||||
} from './dom-ops';
|
||||
import { extractImages, extractJson, extractLinks, extractText } from './extract';
|
||||
import { extractMarkdown } from './markdown';
|
||||
import { pageInfo } from './page-info';
|
||||
|
||||
const contentFunctions = {
|
||||
domQuery,
|
||||
domClick,
|
||||
domType,
|
||||
domAttr,
|
||||
domText,
|
||||
domExists,
|
||||
domScroll,
|
||||
domSelect,
|
||||
domKey,
|
||||
domHover,
|
||||
domCheck,
|
||||
domClear,
|
||||
domFocus,
|
||||
domSubmit,
|
||||
pageInfo,
|
||||
extractLinks,
|
||||
extractImages,
|
||||
extractText,
|
||||
extractJson,
|
||||
extractMarkdown,
|
||||
};
|
||||
|
||||
export type ContentFunctionName = keyof typeof contentFunctions;
|
||||
|
||||
export function contentDispatch(funcName: string, args: ContentArgs = {}) {
|
||||
const fn = contentFunctions[funcName as ContentFunctionName];
|
||||
if (!fn) throw new Error(`Unknown content function: ${funcName}`);
|
||||
return fn(args);
|
||||
}
|
||||
@@ -0,0 +1,114 @@
|
||||
import type { ContentArgs } from '../types';
|
||||
|
||||
export function domQuery({ selector }: ContentArgs) {
|
||||
return Array.from(document.querySelectorAll(selector)).map(el => ({
|
||||
tag: el.tagName.toLowerCase(),
|
||||
text: el.textContent.trim().slice(0, 200),
|
||||
attrs: Object.fromEntries(Array.from(el.attributes).map(a => [a.name, a.value])),
|
||||
}));
|
||||
}
|
||||
|
||||
export function domClick({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.click();
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domType({ selector, text }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLInputElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.focus();
|
||||
el.value = text;
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domAttr({ selector, attr }: ContentArgs) {
|
||||
return Array.from(document.querySelectorAll(selector))
|
||||
.map(el => el.getAttribute(attr))
|
||||
.filter(v => v !== null);
|
||||
}
|
||||
|
||||
export function domText({ selector }: ContentArgs) {
|
||||
return Array.from(document.querySelectorAll(selector))
|
||||
.map(el => el.textContent.trim())
|
||||
.filter(Boolean);
|
||||
}
|
||||
|
||||
export function domExists({ selector }: ContentArgs) {
|
||||
return document.querySelector(selector) !== null;
|
||||
}
|
||||
|
||||
export function domKey({ selector, key }: ContentArgs) {
|
||||
const el = selector ? document.querySelector(selector) : document.activeElement;
|
||||
if (selector && !el) throw new Error(`No element: ${selector}`);
|
||||
const target = el || document.body;
|
||||
["keydown", "keypress", "keyup"].forEach(type => {
|
||||
target.dispatchEvent(new KeyboardEvent(type, { key, bubbles: true, cancelable: true }));
|
||||
});
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domHover({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.dispatchEvent(new MouseEvent("mouseover", { bubbles: true }));
|
||||
el.dispatchEvent(new MouseEvent("mouseenter", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domCheck({ selector, checked }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLInputElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.checked = checked;
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domClear({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLInputElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.value = "";
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domFocus({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.focus();
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domSubmit({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
const form = (el.tagName === "FORM" ? el : el.closest("form")) as HTMLFormElement | null;
|
||||
if (!form) throw new Error(`No form found for: ${selector}`);
|
||||
form.submit();
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domScroll({ selector, x, y }: ContentArgs) {
|
||||
if (selector) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.scrollIntoView({ behavior: "smooth", block: "center" });
|
||||
return true;
|
||||
}
|
||||
window.scrollTo({ top: y || 0, left: x || 0, behavior: "smooth" });
|
||||
return true;
|
||||
}
|
||||
|
||||
export function domSelect({ selector, value }: ContentArgs) {
|
||||
const el = document.querySelector(selector) as HTMLSelectElement | null;
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
el.value = value;
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
return true;
|
||||
}
|
||||
|
||||
@@ -0,0 +1,45 @@
|
||||
import type { ContentArgs } from '../types';
|
||||
|
||||
export function extractLinks() {
|
||||
const seen = new Set<string>();
|
||||
return Array.from(document.querySelectorAll("a[href]")).reduce<Array<{ text: string; href: string }>>((links, a) => {
|
||||
const href = (a as HTMLAnchorElement).href;
|
||||
if (!href || seen.has(href)) return links;
|
||||
seen.add(href);
|
||||
links.push({
|
||||
text: a.textContent.trim().slice(0, 100),
|
||||
href,
|
||||
});
|
||||
return links;
|
||||
}, []);
|
||||
}
|
||||
|
||||
export function extractImages() {
|
||||
const seen = new Set<string>();
|
||||
return Array.from(document.querySelectorAll("img")).reduce<Array<{ alt: string; src: string }>>((images, img) => {
|
||||
const src =
|
||||
img.src ||
|
||||
img.getAttribute("data-src") ||
|
||||
img.getAttribute("data-lazy-src") ||
|
||||
img.getAttribute("data-original") ||
|
||||
(img.srcset ? img.srcset.split(",")[0].trim().split(" ")[0] : "") ||
|
||||
"";
|
||||
if (!src || seen.has(src)) return images;
|
||||
seen.add(src);
|
||||
const FAKE_ALT = new Set(["true", "false", "null", "undefined", "image", "img"]);
|
||||
const alt = img.alt && !FAKE_ALT.has(img.alt.trim().toLowerCase()) ? img.alt.trim() : "";
|
||||
images.push({ alt, src });
|
||||
return images;
|
||||
}, []);
|
||||
}
|
||||
|
||||
export function extractText() {
|
||||
return document.body.innerText;
|
||||
}
|
||||
|
||||
export function extractJson({ selector }: ContentArgs) {
|
||||
const el = document.querySelector(selector);
|
||||
if (!el) throw new Error(`No element: ${selector}`);
|
||||
return JSON.parse(el.textContent);
|
||||
}
|
||||
|
||||
@@ -0,0 +1,404 @@
|
||||
import type { ContentArgs } from '../types';
|
||||
|
||||
export function extractMarkdown({ selector }: ContentArgs) {
|
||||
const BLOCKS = new Set([
|
||||
"article", "aside", "blockquote", "body", "div", "dl", "fieldset", "figcaption",
|
||||
"figure", "footer", "form", "h1", "h2", "h3", "h4", "h5", "h6", "header", "hr",
|
||||
"li", "main", "nav", "ol", "p", "pre", "section", "table", "tbody", "td", "tfoot",
|
||||
"th", "thead", "tr", "ul"
|
||||
]);
|
||||
const NOISE_SELECTOR = [
|
||||
"script",
|
||||
"style",
|
||||
"noscript",
|
||||
"template",
|
||||
"svg",
|
||||
"canvas",
|
||||
"iframe",
|
||||
"dialog",
|
||||
"button",
|
||||
"input",
|
||||
"textarea",
|
||||
"select",
|
||||
"option",
|
||||
"form",
|
||||
"[hidden]",
|
||||
"[aria-hidden='true']",
|
||||
".sr-only",
|
||||
"[class*='sr-only']",
|
||||
"[class*='file-tile']",
|
||||
"form[data-type='unified-composer']",
|
||||
".composer-btn",
|
||||
"[data-composer-surface='true']",
|
||||
"#thread-bottom-container",
|
||||
"[data-testid*='action-button']",
|
||||
].join(", ");
|
||||
|
||||
function normalizeText(value: string) {
|
||||
return value.replace(/\s+/g, " ").trim();
|
||||
}
|
||||
|
||||
function normalizeInline(value: string) {
|
||||
return value
|
||||
.replace(/[ \t]+\n/g, "\n")
|
||||
.replace(/\n[ \t]+/g, "\n")
|
||||
.replace(/\n{3,}/g, "\n\n")
|
||||
.replace(/[ \t]{2,}/g, " ")
|
||||
.trim();
|
||||
}
|
||||
|
||||
function collapseBlankLines(value: string) {
|
||||
return value
|
||||
.replace(/[ \t]+\n/g, "\n")
|
||||
.replace(/\n{3,}/g, "\n\n")
|
||||
.trim();
|
||||
}
|
||||
|
||||
function escapeMarkdown(text: string) {
|
||||
return text.replace(/([\\`[\]])/g, "\\$1");
|
||||
}
|
||||
|
||||
function escapeTableCell(text: string) {
|
||||
return text.replace(/\|/g, "\\|").replace(/\n+/g, " ").trim();
|
||||
}
|
||||
|
||||
function absoluteUrl(attr: string | null | undefined, fallback?: string) {
|
||||
return attr || fallback || "";
|
||||
}
|
||||
|
||||
function isNoiseElement(node: Node | null): boolean {
|
||||
if (!node || node.nodeType !== Node.ELEMENT_NODE) return false;
|
||||
const el = node as Element;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (["script", "style", "noscript", "template", "svg", "canvas", "iframe", "dialog"].includes(tag)) return true;
|
||||
if (["button", "input", "textarea", "select", "option", "form"].includes(tag)) return true;
|
||||
if (el.hasAttribute("hidden")) return true;
|
||||
if ((el.getAttribute("aria-hidden") || "").toLowerCase() === "true") return true;
|
||||
if (el.matches(".sr-only, [class*='sr-only']")) return true;
|
||||
if (el.matches("[class*='file-tile'], form[data-type='unified-composer'], .composer-btn, [data-composer-surface='true'], #thread-bottom-container")) return true;
|
||||
if (el.matches("[data-testid*='action-button']")) return true;
|
||||
return false;
|
||||
}
|
||||
|
||||
function stripNoise(root: Element): Element {
|
||||
const clone = root.cloneNode(true) as Element;
|
||||
clone.querySelectorAll(NOISE_SELECTOR).forEach(node => node.remove());
|
||||
return clone;
|
||||
}
|
||||
|
||||
function candidateScore(node: Element) {
|
||||
const text = normalizeText((node as HTMLElement).innerText || "");
|
||||
if (!text) return -Infinity;
|
||||
|
||||
const headings = node.querySelectorAll("h1, h2, h3, h4, h5, h6").length;
|
||||
const paragraphs = node.querySelectorAll("p").length;
|
||||
const listItems = node.querySelectorAll("li").length;
|
||||
const tables = node.querySelectorAll("table").length;
|
||||
const codeBlocks = node.querySelectorAll("pre, code").length;
|
||||
const images = node.querySelectorAll("img, figure").length;
|
||||
const mainLike = node.matches("main, article, [role='main']") ? 1 : 0;
|
||||
const proseBlocks = node.matches(".markdown, .prose, [data-message-author-role='assistant']") ? 1 : 0;
|
||||
const buttons = node.querySelectorAll("button, input, textarea, select").length;
|
||||
const forms = node.querySelectorAll("form").length;
|
||||
const svgs = node.querySelectorAll("svg, canvas").length;
|
||||
|
||||
return text.length
|
||||
+ (mainLike * 4000)
|
||||
+ (proseBlocks * 5000)
|
||||
+ (headings * 250)
|
||||
+ (paragraphs * 60)
|
||||
+ (listItems * 35)
|
||||
+ (tables * 80)
|
||||
+ (codeBlocks * 60)
|
||||
+ (images * 25)
|
||||
- (buttons * 120)
|
||||
- (forms * 200)
|
||||
- (svgs * 40);
|
||||
}
|
||||
|
||||
function pickRoot() {
|
||||
if (selector) {
|
||||
const matched = document.querySelector(selector);
|
||||
if (!matched) throw new Error(`No element: ${selector}`);
|
||||
return matched;
|
||||
}
|
||||
|
||||
const candidates = Array.from(document.querySelectorAll(
|
||||
"main, article, [role='main'], section, .markdown, .prose, [data-message-author-role]"
|
||||
))
|
||||
.filter(node => normalizeText((node as HTMLElement).innerText || "").length > 0);
|
||||
if (!candidates.length) return document.body;
|
||||
candidates.sort((a, b) => candidateScore(b) - candidateScore(a));
|
||||
return candidates[0];
|
||||
}
|
||||
|
||||
function inlineText(node: Node): string {
|
||||
if (node.nodeType === Node.TEXT_NODE) {
|
||||
return escapeMarkdown(node.textContent || "");
|
||||
}
|
||||
if (node.nodeType !== Node.ELEMENT_NODE) return "";
|
||||
if (isNoiseElement(node)) return "";
|
||||
|
||||
const el = node as HTMLElement;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (tag === "br") return "\n";
|
||||
if (tag === "img") {
|
||||
const img = el as HTMLImageElement;
|
||||
const src = absoluteUrl(img.getAttribute("src"), img.src);
|
||||
if (!src) return "";
|
||||
const alt = normalizeText(img.getAttribute("alt") || "");
|
||||
return alt ? `` : ``;
|
||||
}
|
||||
if (tag === "a") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
const href = absoluteUrl(el.getAttribute("href"), (el as HTMLAnchorElement).href);
|
||||
if (!href) return text;
|
||||
return `[${text || href}](${href})`;
|
||||
}
|
||||
if (tag === "code") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
return text ? `\`${text.replace(/`/g, "\\`")}\`` : "";
|
||||
}
|
||||
if (tag === "strong" || tag === "b") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
return text ? `**${text}**` : "";
|
||||
}
|
||||
if (tag === "em" || tag === "i") {
|
||||
const text = normalizeInline(Array.from(el.childNodes).map(inlineText).join(""));
|
||||
return text ? `*${text}*` : "";
|
||||
}
|
||||
|
||||
const chunks: string[] = [];
|
||||
for (const child of el.childNodes) {
|
||||
const rendered = inlineText(child);
|
||||
if (!rendered) continue;
|
||||
chunks.push(rendered);
|
||||
if (child.nodeType === Node.ELEMENT_NODE && BLOCKS.has((child as Element).tagName.toLowerCase())) {
|
||||
chunks.push("\n");
|
||||
}
|
||||
}
|
||||
return chunks.join("");
|
||||
}
|
||||
|
||||
function textBlock(node: Node): string {
|
||||
return collapseBlankLines(normalizeInline(Array.from(node.childNodes).map(inlineText).join("")));
|
||||
}
|
||||
|
||||
function preserveNodeText(node: Node): string {
|
||||
if (node.nodeType === Node.TEXT_NODE) {
|
||||
return node.textContent || "";
|
||||
}
|
||||
if (node.nodeType !== Node.ELEMENT_NODE) return "";
|
||||
|
||||
const el = node as HTMLElement;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (tag === "br") return "\n";
|
||||
|
||||
const parts: string[] = [];
|
||||
for (const child of el.childNodes) {
|
||||
const rendered = preserveNodeText(child);
|
||||
if (!rendered) continue;
|
||||
parts.push(rendered);
|
||||
}
|
||||
|
||||
if (["div", "p", "li"].includes(tag)) {
|
||||
return `${parts.join("")}\n`;
|
||||
}
|
||||
return parts.join("");
|
||||
}
|
||||
|
||||
function repairFlattenedDiagram(text: string): string {
|
||||
if (text.includes("\n")) return text;
|
||||
const markerCount = (text.match(/[│▼├└]/g) || []).length;
|
||||
if (markerCount < 2) return text;
|
||||
|
||||
let repaired = text;
|
||||
repaired = repaired.replace(/\s{2,}([│▼])/g, "\n $1");
|
||||
repaired = repaired.replace(/([│▼])\s{2,}/g, "$1\n");
|
||||
repaired = repaired.replace(/([│▼])(?=[^\s\n│▼├└])/g, "$1\n");
|
||||
repaired = repaired.replace(/(?<=[^\s\n])([├└])/g, "\n$1");
|
||||
repaired = repaired.replace(/([^\s\n])(\()/g, "$1\n$2");
|
||||
return repaired
|
||||
.split("\n")
|
||||
.map(line => line.replace(/\s+$/, ""))
|
||||
.filter(line => line.trim())
|
||||
.join("\n");
|
||||
}
|
||||
|
||||
function convertDashListsToBranches(lines: string[]): string[] {
|
||||
const converted: string[] = [];
|
||||
let index = 0;
|
||||
while (index < lines.length) {
|
||||
const match = lines[index].match(/^(\s*)-\s+(.*)$/);
|
||||
if (!match) {
|
||||
converted.push(lines[index]);
|
||||
index += 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
const indent = match[1];
|
||||
const items = [];
|
||||
while (index < lines.length) {
|
||||
const nextMatch = lines[index].match(new RegExp(`^${indent.replace(/[.*+?^${}()|[\]\\]/g, "\\$&")}-\\s+(.*)$`));
|
||||
if (!nextMatch) break;
|
||||
items.push(nextMatch[1]);
|
||||
index += 1;
|
||||
}
|
||||
|
||||
items.forEach((item, itemIndex) => {
|
||||
const branch = itemIndex === items.length - 1 ? "└" : "├";
|
||||
converted.push(`${indent}${branch} ${item}`);
|
||||
});
|
||||
}
|
||||
return converted;
|
||||
}
|
||||
|
||||
function normalizeCodeBlock(text: string): string {
|
||||
let lines = text.replace(/\r\n?/g, "\n").split("\n").map(line => line.replace(/\s+$/, ""));
|
||||
while (lines.length && !lines[0].trim()) lines.shift();
|
||||
while (lines.length && !lines[lines.length - 1].trim()) lines.pop();
|
||||
|
||||
const flattened = repairFlattenedDiagram(lines.join("\n"));
|
||||
lines = flattened ? flattened.split("\n") : [];
|
||||
lines = lines.map(line => {
|
||||
const trimmed = line.trim();
|
||||
if ((trimmed === "│" || trimmed === "▼") && !/^\s+[│▼]\s*$/.test(line)) {
|
||||
return ` ${trimmed}`;
|
||||
}
|
||||
return line;
|
||||
});
|
||||
lines = convertDashListsToBranches(lines);
|
||||
return lines.join("\n");
|
||||
}
|
||||
|
||||
function tableToMarkdown(table: Element) {
|
||||
const rows = Array.from(table.querySelectorAll("tr"))
|
||||
.map(row => Array.from(row.children)
|
||||
.filter(cell => cell.tagName === "TD" || cell.tagName === "TH")
|
||||
.map(cell => escapeTableCell(textBlock(cell)))
|
||||
)
|
||||
.filter(cells => cells.length > 0);
|
||||
if (!rows.length) return "";
|
||||
|
||||
const widths = rows.reduce((max, row) => Math.max(max, row.length), 0);
|
||||
const normalizedRows = rows.map(row => {
|
||||
const next = row.slice();
|
||||
while (next.length < widths) next.push("");
|
||||
return next;
|
||||
});
|
||||
|
||||
let headers = normalizedRows[0];
|
||||
let bodyRows = normalizedRows.slice(1);
|
||||
const firstRowIsBlank = headers.every(cell => !cell.trim());
|
||||
if (firstRowIsBlank && normalizedRows.length > 1) {
|
||||
headers = normalizedRows[1];
|
||||
bodyRows = normalizedRows.slice(2);
|
||||
}
|
||||
|
||||
const firstRow = table.querySelector("tr");
|
||||
const thead = table.querySelector("thead");
|
||||
const firstRowHasTh = firstRow && Array.from(firstRow.children).some(cell => cell.tagName === "TH");
|
||||
if (!(thead || firstRowHasTh || firstRowIsBlank)) {
|
||||
headers = new Array(widths).fill("");
|
||||
bodyRows = normalizedRows;
|
||||
}
|
||||
|
||||
const separator = new Array(widths).fill("---");
|
||||
const lines = [
|
||||
`| ${headers.join(" | ")} |`,
|
||||
`| ${separator.join(" | ")} |`,
|
||||
];
|
||||
for (const row of bodyRows) {
|
||||
lines.push(`| ${row.join(" | ")} |`);
|
||||
}
|
||||
return lines.join("\n");
|
||||
}
|
||||
|
||||
function listToMarkdown(list: Element, depth = 0): string {
|
||||
const ordered = list.tagName.toLowerCase() === "ol";
|
||||
const items: string[] = [];
|
||||
const children = Array.from(list.children).filter(child => child.tagName === "LI");
|
||||
children.forEach((item, index) => {
|
||||
const marker = ordered ? `${index + 1}. ` : "- ";
|
||||
const indent = " ".repeat(depth);
|
||||
const nested: string[] = [];
|
||||
const content: string[] = [];
|
||||
|
||||
for (const child of item.childNodes) {
|
||||
const childEl = child as Element;
|
||||
if (child.nodeType === Node.ELEMENT_NODE && (childEl.tagName === "UL" || childEl.tagName === "OL")) {
|
||||
nested.push(listToMarkdown(childEl, depth + 1));
|
||||
} else {
|
||||
content.push(inlineText(child));
|
||||
}
|
||||
}
|
||||
|
||||
const line = collapseBlankLines(normalizeInline(content.join("")));
|
||||
if (line) {
|
||||
const lineParts = line.split("\n");
|
||||
items.push(`${indent}${marker}${lineParts[0]}`);
|
||||
const continuationIndent = `${indent}${" ".repeat(marker.length)}`;
|
||||
lineParts.slice(1).forEach(part => items.push(`${continuationIndent}${part}`));
|
||||
}
|
||||
nested.filter(Boolean).forEach(block => items.push(block));
|
||||
});
|
||||
return items.join("\n");
|
||||
}
|
||||
|
||||
function blockToMarkdown(node: Node): string {
|
||||
if (node.nodeType === Node.TEXT_NODE) {
|
||||
return normalizeText(node.textContent || "");
|
||||
}
|
||||
if (node.nodeType !== Node.ELEMENT_NODE) return "";
|
||||
if (isNoiseElement(node)) return "";
|
||||
|
||||
const el = node as HTMLElement;
|
||||
const tag = el.tagName.toLowerCase();
|
||||
if (tag === "table") return tableToMarkdown(el);
|
||||
if (tag === "ul" || tag === "ol") return listToMarkdown(el);
|
||||
if (el.matches(".cm-editor[data-is-code-block-view='true']")) {
|
||||
const lines = Array.from(el.querySelectorAll(".cm-line")).map(line => {
|
||||
const text = preserveNodeText(line);
|
||||
return text === "\n" ? "" : text.replace(/\n$/, "");
|
||||
});
|
||||
const code = normalizeCodeBlock(lines.join("\n"));
|
||||
return code ? `\`\`\`\n${code}\n\`\`\`` : "";
|
||||
}
|
||||
if (tag === "pre") {
|
||||
const code = normalizeCodeBlock(preserveNodeText(el));
|
||||
return code ? `\`\`\`\n${code}\n\`\`\`` : "";
|
||||
}
|
||||
if (tag === "blockquote") {
|
||||
const content = collapseBlankLines(Array.from(el.childNodes).map(blockToMarkdown).join("\n\n"));
|
||||
return content
|
||||
.split("\n")
|
||||
.map(line => line ? `> ${line}` : ">")
|
||||
.join("\n");
|
||||
}
|
||||
if (/^h[1-6]$/.test(tag)) {
|
||||
const level = Number(tag.slice(1));
|
||||
const text = textBlock(el);
|
||||
return text ? `${"#".repeat(level)} ${text}` : "";
|
||||
}
|
||||
if (tag === "p" || tag === "figcaption") {
|
||||
return textBlock(el);
|
||||
}
|
||||
if (tag === "hr") {
|
||||
return "---";
|
||||
}
|
||||
if (tag === "img") {
|
||||
return inlineText(el);
|
||||
}
|
||||
|
||||
const childBlocks = Array.from(el.childNodes)
|
||||
.map(child => blockToMarkdown(child))
|
||||
.filter(Boolean);
|
||||
if (childBlocks.length) return collapseBlankLines(childBlocks.join("\n\n"));
|
||||
|
||||
return textBlock(node);
|
||||
}
|
||||
|
||||
const root = stripNoise(pickRoot());
|
||||
const markdown = blockToMarkdown(root);
|
||||
return collapseBlankLines(markdown);
|
||||
}
|
||||
@@ -0,0 +1,14 @@
|
||||
export function pageInfo() {
|
||||
const metas: Record<string, string> = {};
|
||||
document.querySelectorAll("meta[name], meta[property]").forEach(m => {
|
||||
const k = m.getAttribute("name") || m.getAttribute("property");
|
||||
if (k) metas[k] = m.getAttribute("content") || "";
|
||||
});
|
||||
return {
|
||||
title: document.title,
|
||||
url: location.href,
|
||||
readyState: document.readyState,
|
||||
lang: document.documentElement.lang || null,
|
||||
meta: metas,
|
||||
};
|
||||
}
|
||||
@@ -9,6 +9,11 @@ export const LARGE_OPERATION_BATCH_SIZE = 25;
|
||||
export const LARGE_OPERATION_PAUSE_MS = 25;
|
||||
export const GENTLE_OPERATION_BATCH_SIZE = 8;
|
||||
export const GENTLE_OPERATION_PAUSE_MS = 100;
|
||||
const DEBUG_LARGE_OPERATIONS = false;
|
||||
|
||||
function debugLargeOperation(message: string) {
|
||||
if (DEBUG_LARGE_OPERATIONS) console.log(message);
|
||||
}
|
||||
|
||||
export async function hasAudibleTabs() {
|
||||
const audibleTabs = await chrome.tabs.query({ audible: true });
|
||||
@@ -19,11 +24,11 @@ let largeOperationQueue: Promise<void> = Promise.resolve();
|
||||
|
||||
export async function runLargeOperation<T>(name: string, fn: () => Promise<T>): Promise<T> {
|
||||
const run = largeOperationQueue.then(async () => {
|
||||
console.log(`[browser-cli] large operation start: ${name}`);
|
||||
debugLargeOperation(`[browser-cli] large operation start: ${name}`);
|
||||
try {
|
||||
return await fn();
|
||||
} finally {
|
||||
console.log(`[browser-cli] large operation done: ${name}`);
|
||||
debugLargeOperation(`[browser-cli] large operation done: ${name}`);
|
||||
}
|
||||
});
|
||||
largeOperationQueue = run.then(() => {}, () => {});
|
||||
|
||||
Reference in New Issue
Block a user