@y11i-3d/chrome-recording 1.0.0 → 1.1.0

package/README.md

@@ -25,13 +25,14 @@ Or run directly with `npx`:
 npx @y11i-3d/chrome-recording <subcommand> [options]
 ```
 
-After installation, a PowerShell script (`get_chrome_info.ps1`) is automatically copied to `%LOCALAPPDATA%\y11i-3d\chrome-recording\` on Windows. This script is required by the `record` command.
+After installation, a PowerShell script (`chrome.ps1`) is automatically copied to `%LOCALAPPDATA%\y11i-3d\chrome-recording\` on Windows. This script is required by the `record` and `resize` commands.
 
 ## Subcommands
 
 | Subcommand | Description                                             |
 | ---------- | ------------------------------------------------------- |
 | `record`   | Capture Chrome browser content as a video or screenshot |
+| `resize`   | Set the Chrome viewport size                            |
 | `concat`   | Concatenate recorded mp4 files into one                 |
 
 ## record
@@ -53,6 +54,8 @@ By default, recording starts immediately and stops when you press Enter. The out
 | `-d, --duration <seconds>`           | Stop recording automatically after N seconds                                            |
 | `-q, --quality <value>`              | Quality: plain number for CRF (e.g. `23`), number+unit for bitrate (e.g. `4000k`, `4M`) |
 | `-f, --fps <fps>`                    | Frame rate                                                                              |
+| `-n, --no-pointer`                   | Hide the mouse pointer                                                                  |
+| `-t, --title <text>`                 | Select a window whose page title contains this text                                     |
 | `-v, --verbose`                      | Show ffmpeg output                                                                      |
 
 ### Examples
@@ -74,6 +77,41 @@ chrome-recording record -m -c 40:0:0:0
 chrome-recording record -o output.mp4 -q 4M
 ```
 
+## resize
+
+Set the viewport size through the Chrome DevTools Protocol. Chrome asks you to
+approve the connection each time this command runs.
+
+```bash
+chrome-recording resize <width>x<height> [options]
+```
+
+The viewport is the area the page is rendered into, and it is what `record`
+captures. `--padding` enlarges the surrounding content area without changing the
+viewport, which keeps the rounded window corners away from the captured region.
+Padding is added to the right and the bottom.
+
+| Option                           | Description                                        |
+| -------------------------------- | -------------------------------------------------- |
+| `-p, --padding <width>x<height>` | Padding kept outside the viewport (default: `0x0`) |
+| `-t, --title <text>`             | Select a page whose title contains this text       |
+
+The setting survives after the command exits, but it is lost when the page is
+reloaded.
+
+### Examples
+
+```bash
+# Set the viewport to 1920x1080
+chrome-recording resize 1920x1080
+
+# Keep 4px below the viewport so the rounded corners are not captured
+chrome-recording resize 1920x1080 -p 0x4
+
+# Target a specific page
+chrome-recording resize 1280x720 -t "Example"
+```
+
 ## concat
 
 Concatenate all mp4 files in a directory into a single file (stream copy, no re-encode).

package/dist/chrome.d.ts

@@ -0,0 +1,5 @@
+export declare function runChromeScript(args: string[]): string[];
+export declare function parseSize(value: string, label: string): {
+    width: number;
+    height: number;
+};

package/dist/chrome.js

@@ -0,0 +1,33 @@
+import { execSync } from "child_process";
+const SCRIPT_NAME = "chrome.ps1";
+function scriptPathWin() {
+    const localAppDataWin = execSync("powershell.exe -Command '$env:LOCALAPPDATA'")
+        .toString()
+        .trim();
+    return `${localAppDataWin}\\y11i-3d\\chrome-recording\\${SCRIPT_NAME}`;
+}
+export function runChromeScript(args) {
+    const quoted = args.map((arg) => `"${arg.replace(/"/g, '""')}"`).join(" ");
+    const command = `powershell.exe -ExecutionPolicy Bypass -File "${scriptPathWin()}" ${quoted}`;
+    try {
+        return execSync(command, { stdio: ["ignore", "pipe", "pipe"] })
+            .toString()
+            .trim()
+            .split("\n")
+            .map((line) => line.trim().replace(/\r/, ""))
+            .filter((line) => line.length > 0);
+    }
+    catch (error) {
+        const stderr = error.stderr?.toString().trim();
+        throw new Error(stderr && stderr.length > 0 ? stderr : String(error), {
+            cause: error,
+        });
+    }
+}
+export function parseSize(value, label) {
+    const match = /^(\d+)x(\d+)$/.exec(value);
+    if (!match) {
+        throw new Error(`Invalid ${label}: "${value}". Expected <width>x<height>.`);
+    }
+    return { width: Number(match[1]), height: Number(match[2]) };
+}

package/dist/chrome.ps1

@@ -0,0 +1,212 @@
+param(
+    [Parameter(Position = 0)][ValidateSet("info", "resize")][string]$Command = "info",
+    [string]$Title = "",
+    [int]$ViewportWidth = 0,
+    [int]$ViewportHeight = 0,
+    [int]$PaddingWidth = 0,
+    [int]$PaddingHeight = 0
+)
+
+$ErrorActionPreference = "Stop"
+[Console]::OutputEncoding = [System.Text.Encoding]::UTF8
+Add-Type -AssemblyName System.Windows.Forms
+
+Add-Type -TypeDefinition @"
+using System;
+using System.Collections.Generic;
+using System.Runtime.InteropServices;
+using System.Text;
+public class Win32 {
+    public delegate bool EnumWindowsProc(IntPtr hwnd, IntPtr lParam);
+    [DllImport("user32.dll")] public static extern bool EnumWindows(EnumWindowsProc cb, IntPtr lParam);
+    [DllImport("user32.dll")] public static extern bool EnumChildWindows(IntPtr parent, EnumWindowsProc cb, IntPtr lParam);
+    [DllImport("user32.dll")] public static extern bool GetWindowRect(IntPtr h, out RECT r);
+    [DllImport("user32.dll")] public static extern int GetClassName(IntPtr hwnd, StringBuilder buf, int maxCount);
+    [DllImport("user32.dll", CharSet = CharSet.Unicode)] public static extern int GetWindowTextW(IntPtr hwnd, StringBuilder buf, int maxCount);
+    [DllImport("user32.dll")] public static extern bool IsWindowVisible(IntPtr h);
+    [DllImport("user32.dll")] public static extern uint GetWindowThreadProcessId(IntPtr h, out uint pid);
+    public struct RECT { public int Left, Top, Right, Bottom; }
+
+    public const string TitleSuffix = " - Google Chrome";
+
+    public static List<IntPtr> FindBrowserWindows(uint[] pids, string query) {
+        var set = new HashSet<uint>(pids);
+        var results = new List<IntPtr>();
+        EnumWindows((h, _) => {
+            uint pid;
+            GetWindowThreadProcessId(h, out pid);
+            if (!set.Contains(pid)) return true;
+            var cls = new StringBuilder(256);
+            GetClassName(h, cls, 256);
+            if (cls.ToString() != "Chrome_WidgetWin_1") return true;
+            if (!IsWindowVisible(h)) return true;
+            var buf = new StringBuilder(512);
+            GetWindowTextW(h, buf, 512);
+            var title = buf.ToString();
+            if (!title.EndsWith(TitleSuffix)) return true;
+            if (query.Length > 0) {
+                var pageTitle = title.Substring(0, title.Length - TitleSuffix.Length);
+                if (pageTitle.IndexOf(query, StringComparison.OrdinalIgnoreCase) < 0) return true;
+            }
+            results.Add(h);
+            return true;
+        }, IntPtr.Zero);
+        return results;
+    }
+
+    public static List<RECT> FindContentRects(IntPtr chromeHwnd) {
+        RECT mainRect;
+        GetWindowRect(chromeHwnd, out mainRect);
+        var results = new List<RECT>();
+        EnumChildWindows(chromeHwnd, (h, _) => {
+            var buf = new StringBuilder(256);
+            GetClassName(h, buf, 256);
+            if (buf.ToString() == "Chrome_RenderWidgetHostHWND" && IsWindowVisible(h)) {
+                RECT r;
+                GetWindowRect(h, out r);
+                if (r.Top >= mainRect.Top && r.Top < mainRect.Bottom) {
+                    results.Add(r);
+                }
+            }
+            return true;
+        }, IntPtr.Zero);
+        return results;
+    }
+}
+"@
+
+$chromePids = @(Get-Process chrome -ErrorAction SilentlyContinue | ForEach-Object { [uint32]$_.Id })
+if ($chromePids.Count -eq 0) {
+    [Console]::Error.WriteLine("Chrome is not running.")
+    exit 1
+}
+
+function Get-TargetWindow {
+    $windows = [Win32]::FindBrowserWindows($chromePids, $Title)
+    if ($windows.Count -eq 0) {
+        if ($Title.Length -gt 0) {
+            [Console]::Error.WriteLine("No Chrome window matching '$Title' was found.")
+        }
+        else {
+            [Console]::Error.WriteLine("Chrome is not running or has no visible window.")
+        }
+        exit 1
+    }
+    return $windows[0]
+}
+
+if ($Command -eq "info") {
+    $hwnd = Get-TargetWindow
+    $rects = [Win32]::FindContentRects($hwnd)
+    $best = $rects | Sort-Object { $_.Left }, { $_.Top } | Select-Object -First 1
+    if (-not $best) {
+        [Console]::Error.WriteLine("Chrome content area (Chrome_RenderWidgetHostHWND) not found.")
+        exit 1
+    }
+
+    $x = $best.Left
+    $y = $best.Top
+    $w = $best.Right - $best.Left
+    $h = $best.Bottom - $best.Top
+    Write-Output "content $x $y $w $h"
+
+    $i = 0
+    foreach ($s in [System.Windows.Forms.Screen]::AllScreens) {
+        $b = $s.Bounds
+        Write-Output "monitor $i $($b.Left) $($b.Top) $($b.Right) $($b.Bottom)"
+        $i++
+    }
+    exit 0
+}
+
+if ([Win32]::FindBrowserWindows($chromePids, "").Count -eq 0) {
+    [Console]::Error.WriteLine("Chrome is not running or has no visible window.")
+    exit 1
+}
+
+$portFile = Join-Path $env:LOCALAPPDATA "Google\Chrome\User Data\DevToolsActivePort"
+if (-not (Test-Path $portFile)) {
+    [Console]::Error.WriteLine("DevToolsActivePort not found. Enable remote debugging at chrome://inspect/#remote-debugging.")
+    exit 1
+}
+$portLines = Get-Content $portFile
+$endpoint = "ws://127.0.0.1:$($portLines[0].Trim())$($portLines[1].Trim())"
+
+$ws = New-Object System.Net.WebSockets.ClientWebSocket
+$token = [System.Threading.CancellationToken]::None
+$connected = $false
+try {
+    $connected = $ws.ConnectAsync([Uri]$endpoint, $token).Wait(60000)
+}
+catch {
+    $connected = $false
+}
+if (-not $connected) {
+    [Console]::Error.WriteLine("Could not connect to Chrome. Approve the connection prompt in the browser and try again.")
+    exit 1
+}
+
+$script:cdpId = 0
+function Invoke-Cdp([string]$method, $params, [string]$sessionId) {
+    $script:cdpId++
+    $myId = $script:cdpId
+    $payload = [ordered]@{ id = $myId; method = $method; params = if ($params) { $params } else { @{} } }
+    if ($sessionId) { $payload.sessionId = $sessionId }
+    $json = (New-Object psobject -Property $payload) | ConvertTo-Json -Depth 10 -Compress
+    $bytes = [System.Text.Encoding]::UTF8.GetBytes($json)
+    $segment = New-Object System.ArraySegment[byte] -ArgumentList @(, $bytes)
+    $ws.SendAsync($segment, [System.Net.WebSockets.WebSocketMessageType]::Text, $true, $token).Wait(10000) | Out-Null
+
+    for ($i = 0; $i -lt 500; $i++) {
+        $sb = New-Object System.Text.StringBuilder
+        do {
+            $buf = New-Object byte[] 65536
+            $recvSegment = New-Object System.ArraySegment[byte] -ArgumentList @(, $buf)
+            $recv = $ws.ReceiveAsync($recvSegment, $token)
+            if (-not $recv.Wait(10000)) { throw "Timed out waiting for a CDP response." }
+            [void]$sb.Append([System.Text.Encoding]::UTF8.GetString($buf, 0, $recv.Result.Count))
+        } while (-not $recv.Result.EndOfMessage)
+        $message = $sb.ToString() | ConvertFrom-Json
+        if ($message.id -eq $myId) {
+            if ($message.error) { throw "CDP error on $method : $($message.error.message)" }
+            return $message
+        }
+    }
+    throw "No CDP response for $method."
+}
+
+try {
+    $targets = (Invoke-Cdp "Target.getTargets" $null $null).result.targetInfos
+    $pages = @($targets | Where-Object { $_.type -eq "page" })
+    if ($Title.Length -gt 0) {
+        $pages = @($pages | Where-Object { $_.title -like "*$Title*" })
+    }
+    if ($pages.Count -eq 0) {
+        if ($Title.Length -gt 0) { throw "No page matching '$Title' was found." }
+        throw "No page was found."
+    }
+    $target = $pages[0]
+
+    $windowId = (Invoke-Cdp "Browser.getWindowForTarget" @{ targetId = $target.targetId } $null).result.windowId
+    $contentWidth = $ViewportWidth + $PaddingWidth
+    $contentHeight = $ViewportHeight + $PaddingHeight
+    Invoke-Cdp "Browser.setContentsSize" @{ windowId = $windowId; width = $contentWidth; height = $contentHeight } $null | Out-Null
+
+    $sessionId = (Invoke-Cdp "Target.attachToTarget" @{ targetId = $target.targetId; flatten = $true } $null).result.sessionId
+    Invoke-Cdp "Emulation.setDeviceMetricsOverride" @{
+        width             = $ViewportWidth
+        height            = $ViewportHeight
+        deviceScaleFactor = 1
+        mobile            = $false
+    } $sessionId | Out-Null
+
+    Write-Output "viewport $ViewportWidth $ViewportHeight"
+    Write-Output "content $contentWidth $contentHeight"
+    Write-Output "title $($target.title)"
+}
+catch {
+    [Console]::Error.WriteLine($_.Exception.Message)
+    $ws.Dispose()
+    exit 1
+}
+$ws.Dispose()

package/dist/cli.js

@@ -2,10 +2,18 @@
 import { Command } from "commander";
 import { registerConcatCommand } from "./concat.js";
 import { registerRecordCommand } from "./record.js";
+import { registerResizeCommand } from "./resize.js";
+function reportAndExit(error) {
+    console.error(error instanceof Error ? error.message : String(error));
+    process.exit(1);
+}
+process.on("uncaughtException", reportAndExit);
+process.on("unhandledRejection", reportAndExit);
 const program = new Command();
 program
     .name("chrome-recording")
     .description("Record Chrome browser content in WSL");
 registerRecordCommand(program);
+registerResizeCommand(program);
 registerConcatCommand(program);
 program.parse();

package/dist/postinstall.js

@@ -6,7 +6,7 @@ import { execSync } from "child_process";
 import pkg from "../package.json" with { type: "json" };
 const [scope, pkgName] = pkg.name.replace(/^@/, "").split("/");
 const scriptDir = dirname(fileURLToPath(import.meta.url));
-const ps1Src = resolve(scriptDir, "get_chrome_info.ps1");
+const ps1Src = resolve(scriptDir, "chrome.ps1");
 const localAppDataWin = execSync("powershell.exe -Command '$env:LOCALAPPDATA'")
     .toString()
     .trim();
@@ -15,5 +15,5 @@ const localAppDataLinux = execSync(`wslpath -u "${localAppDataWin}"`)
     .trim();
 const destDir = resolve(localAppDataLinux, scope ?? pkg.name, pkgName ?? "");
 mkdirSync(destDir, { recursive: true });
-copyFileSync(ps1Src, resolve(destDir, "get_chrome_info.ps1"));
-console.log(`Installed get_chrome_info.ps1 to ${destDir}`);
+copyFileSync(ps1Src, resolve(destDir, "chrome.ps1"));
+console.log(`Installed chrome.ps1 to ${destDir}`);

package/dist/record.js

@@ -2,6 +2,7 @@ import { execSync, spawn } from "child_process";
 import { mkdirSync } from "fs";
 import { dirname, resolve } from "path";
 import { createInterface } from "readline";
+import { runChromeScript } from "./chrome.js";
 export function registerRecordCommand(program) {
     program
         .command("record")
@@ -13,6 +14,8 @@ export function registerRecordCommand(program) {
         .option("-d, --duration <seconds>", "Recording duration in seconds")
         .option("-q, --quality <value>", "Quality: number for CRF, number+unit for bitrate (e.g. 23, 4000k, 4M)")
         .option("-f, --fps <fps>", "Frame rate")
+        .option("-n, --no-pointer", "Hide the mouse pointer")
+        .option("-t, --title <text>", "Select a window whose page title contains this text")
         .option("-v, --verbose", "Show ffmpeg output")
         .action(async (options) => {
         const missing = ["powershell.exe", "ffmpeg.exe", "wslpath"].filter((cmd) => {
@@ -27,15 +30,10 @@ export function registerRecordCommand(program) {
         if (missing.length > 0) {
             throw new Error(`Command(s) not found: ${missing.join(", ")}\nWSL + Windows environment required.`);
         }
-        const localAppDataWin = execSync("powershell.exe -Command '$env:LOCALAPPDATA'")
-            .toString()
-            .trim();
-        const ps1Win = `${localAppDataWin}\\y11i-3d\\chrome-recording\\get_chrome_info.ps1`;
-        const info = execSync(`powershell.exe -ExecutionPolicy Bypass -File "${ps1Win}"`)
-            .toString()
-            .trim()
-            .split("\n")
-            .map((l) => l.trim().replace(/\r/, ""));
+        const infoArgs = ["info"];
+        if (options.title)
+            infoArgs.push("-Title", options.title);
+        const info = runChromeScript(infoArgs);
         const contentLine = info.find((l) => l.startsWith("content"));
         if (!contentLine) {
             throw new Error("Failed to get Chrome content area info");
@@ -102,6 +100,7 @@ export function registerRecordCommand(program) {
             `offset_y=${offsetY}`,
             `video_size=${contentW}x${contentH}`,
             ...(options.fps !== undefined ? [`framerate=${options.fps}`] : []),
+            ...(options.pointer ? [] : ["draw_mouse=0"]),
         ];
         const ddagrab = ddagrabParams.join(":");
         const ffmpegArgs = screenshot
@@ -119,6 +118,8 @@ export function registerRecordCommand(program) {
                 ...(duration !== null ? ["-t", String(duration)] : []),
                 "-c:v",
                 "libx264",
+                "-pix_fmt",
+                "yuv420p",
                 ...qualityArgs,
                 "-y",
                 outputWin,

package/dist/resize.d.ts

@@ -0,0 +1,2 @@
+import type { Command } from "commander";
+export declare function registerResizeCommand(program: Command): void;

package/dist/resize.js

@@ -0,0 +1,29 @@
+import { parseSize, runChromeScript } from "./chrome.js";
+export function registerResizeCommand(program) {
+    program
+        .command("resize")
+        .description("Resize the Chrome viewport")
+        .argument("<size>", "Viewport size as <width>x<height>")
+        .option("-p, --padding <width>x<height>", "Padding kept outside the viewport (right and bottom)", "0x0")
+        .option("-t, --title <text>", "Select a page whose title contains this text")
+        .action((size, options) => {
+        const viewport = parseSize(size, "size");
+        const padding = parseSize(options.padding, "padding");
+        const args = [
+            "resize",
+            "-ViewportWidth",
+            String(viewport.width),
+            "-ViewportHeight",
+            String(viewport.height),
+            "-PaddingWidth",
+            String(padding.width),
+            "-PaddingHeight",
+            String(padding.height),
+        ];
+        if (options.title)
+            args.push("-Title", options.title);
+        for (const line of runChromeScript(args)) {
+            console.log(line);
+        }
+    });
+}

package/package.json

@@ -8,7 +8,7 @@
     "windows",
     "wsl"
   ],
-  "version": "1.0.0",
+  "version": "1.1.0",
   "type": "module",
   "files": [
     "dist"
@@ -17,7 +17,7 @@
     "chrome-recording": "dist/cli.js"
   },
   "scripts": {
-    "build": "rm -rf dist && tsc && cp src/get_chrome_info.ps1 dist/",
+    "build": "rm -rf dist && tsc && cp src/chrome.ps1 dist/",
     "postinstall": "node dist/postinstall.js"
   },
   "devDependencies": {

package/dist/get_chrome_info.ps1

@@ -1,62 +0,0 @@
-Add-Type -AssemblyName System.Windows.Forms
-
-Add-Type -TypeDefinition @"
-using System;
-using System.Collections.Generic;
-using System.Runtime.InteropServices;
-using System.Text;
-public class Win32 {
-    public delegate bool EnumWindowsProc(IntPtr hwnd, IntPtr lParam);
-    [DllImport("user32.dll")] public static extern bool GetWindowRect(IntPtr h, out RECT r);
-    [DllImport("user32.dll")] public static extern int GetClassName(IntPtr hwnd, StringBuilder buf, int maxCount);
-    [DllImport("user32.dll")] public static extern bool EnumChildWindows(IntPtr parent, EnumWindowsProc cb, IntPtr lParam);
-    public struct RECT { public int Left, Top, Right, Bottom; }
-
-    public static List<RECT> FindContentRects(IntPtr chromeHwnd) {
-        RECT mainRect;
-        GetWindowRect(chromeHwnd, out mainRect);
-        var results = new List<RECT>();
-        EnumChildWindows(chromeHwnd, (h, _) => {
-            var buf = new StringBuilder(256);
-            GetClassName(h, buf, 256);
-            if (buf.ToString() == "Chrome_RenderWidgetHostHWND") {
-                RECT r;
-                GetWindowRect(h, out r);
-                if (r.Top >= mainRect.Top && r.Top < mainRect.Bottom) {
-                    results.Add(r);
-                }
-            }
-            return true;
-        }, IntPtr.Zero);
-        return results;
-    }
-}
-"@
-
-$proc = Get-Process chrome -ErrorAction SilentlyContinue | Where-Object { $_.MainWindowTitle -ne "" } | Select-Object -First 1
-if (-not $proc) {
-    Write-Error "Chrome is not running or has no visible window."
-    exit 1
-}
-
-$rects = [Win32]::FindContentRects($proc.MainWindowHandle)
-$best = $rects | Sort-Object { $_.Left }, { $_.Top } | Select-Object -First 1
-
-if (-not $best) {
-    Write-Error "Chrome content area (Chrome_RenderWidgetHostHWND) not found."
-    exit 1
-}
-
-$x = $best.Left
-$y = $best.Top
-$w = $best.Right - $best.Left
-$h = $best.Bottom - $best.Top
-
-Write-Output "content $x $y $w $h"
-
-$i = 0
-foreach ($s in [System.Windows.Forms.Screen]::AllScreens) {
-    $b = $s.Bounds
-    Write-Output "monitor $i $($b.Left) $($b.Top) $($b.Right) $($b.Bottom)"
-    $i++
-}