[feature/ISSUE-66] Add proxy support

azihassan · Apr 6, 2024 · 537cbaf · 537cbaf
1 parent e943fd5
commit 537cbaf
Show file tree

Hide file tree

Showing 9 changed files with 123 additions and 25 deletions.
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -6,7 +6,7 @@ jobs:
   unit_test:
     strategy:
       matrix:
-        os: [ { name: ubuntu-20.04, extension: '' } , { name: windows-2019, extension: '.exe' }, { name: macos-latest, extension: '' } ]
+        os: [ { name: macos-latest, extension: '' } ]
 
     runs-on: ${{ matrix.os.name }}
 
@@ -31,7 +31,7 @@ jobs:
   console_test:
     strategy:
       matrix:
-        os: [ { name: ubuntu-20.04, command: sh tests/tests.sh, extension: '' } , { name: windows-2019, command: powershell.exe -file tests\tests.ps1, extension: '.exe' }, { name: macos-latest, command: sh tests/tests-macos.sh, extension: '' }]
+        os: [{ name: macos-latest, command: sh tests/tests-macos.sh, extension: '' }]
 
     runs-on: ${{ matrix.os.name }}
     needs: unit_test
@@ -52,6 +52,17 @@ jobs:
         tar -xf libcurl-7.68.0-WinSSL-zlib-x86-x64.zip
         xcopy dmd2\windows\bin64\libcurl.dll .
 
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.x'
+
+    - name: Install mitmproxy
+      run: pip install mitmproxy
+
+    - name: Setup tmate session
+      uses: mxschmitt/action-tmate@v3
+
     - name: e2e tests
       run: ${{ matrix.os.command }}
 
diff --git a/script.py b/script.py
@@ -0,0 +1,5 @@
+from mitmproxy import ctx
+
+def request(flow):
+    print(flow.request.url)
+    ctx.master.shutdown()
diff --git a/source/app.d b/source/app.d
@@ -38,6 +38,7 @@ void main(string[] args)
     bool noProgress;
     bool noCache;
     bool dethrottle = true;
+    string proxy;
 
     version(linux)
     {
@@ -56,6 +57,7 @@ void main(string[] args)
         "no-cache", "Skip caching of HTML and base.js", &noCache,
         "d|dethrottle", "Attempt to dethrottle download speed by solving the N challenge (defaults to true)", &dethrottle,
         "no-dethrottle", "Skip N-challenge dethrottling attempt", () { dethrottle = false; },
+        "proxy", "Specifies a proxy in the type://host:port format", &proxy
     );
 
     if(help.helpWanted || args.length == 1)
@@ -85,7 +87,8 @@ void main(string[] args)
                     parallel,
                     noProgress,
                     retry > 0 ? true : noCache, //force cache refresh on failure,
-                    dethrottle
+                    dethrottle,
+                    proxy
                 );
                 break;
             }
@@ -105,10 +108,12 @@ void main(string[] args)
     }
 }
 
-void handleURL(string url, int itag, StdoutLogger logger, bool displayFormats, bool outputURL, bool parallel, bool noProgress, bool noCache, bool dethrottle)
+void handleURL(string url, int itag, StdoutLogger logger, bool displayFormats, bool outputURL, bool parallel, bool noProgress, bool noCache, bool dethrottle, string proxy)
 {
     logger.display(formatTitle("Handling " ~ url));
-    YoutubeVideoURLExtractor parser = Cache(logger, noCache ? Yes.forceRefresh : No.forceRefresh).makeParser(url, itag);
+    Cache cache = Cache(logger, noCache ? Yes.forceRefresh : No.forceRefresh);
+    cache.proxy = proxy;
+    YoutubeVideoURLExtractor parser = cache.makeParser(url, itag);
     logger.displayVerbose("Downloaded video HTML");
     logger.displayVerbose("Attempt to dethrottle : " ~ (dethrottle ? "Yes" : "No"));
 
@@ -171,5 +176,6 @@ void handleURL(string url, int itag, StdoutLogger logger, bool displayFormats, b
             return 0;
         }, !noProgress);
     }
+    downloader.setProxy(proxy);
     downloader.download(destination, link, url);
 }
diff --git a/source/cache.d b/source/cache.d
@@ -11,7 +11,7 @@ import std.string : indexOf;
 import std.regex : ctRegex, matchFirst;
 import std.algorithm : map;
 
-import helpers : StdoutLogger, parseID, parseQueryString, parseBaseJSKey;
+import helpers : StdoutLogger, parseID, parseQueryString, parseBaseJSKey, USER_AGENT;
 import parsers : parseBaseJSURL, YoutubeVideoURLExtractor, SimpleYoutubeVideoURLExtractor, AdvancedYoutubeVideoURLExtractor;
 
 struct Cache
@@ -20,6 +20,7 @@ struct Cache
     private string delegate(string url) downloadAsString;
     private Flag!"forceRefresh" forceRefresh;
     string cacheDirectory;
+    string proxy;
 
     this(StdoutLogger logger, Flag!"forceRefresh" forceRefresh = No.forceRefresh)
     {
@@ -32,8 +33,14 @@ struct Cache
             Curl curl;
             curl.initialize();
             curl.set(CurlOption.url, url);
+            curl.set(CurlOption.useragent, USER_AGENT);
             curl.set(CurlOption.encoding, "deflate, gzip");
             curl.set(CurlOption.followlocation, true);
+            curl.set(CurlOption.followlocation, true);
+            if(proxy)
+            {
+                curl.set(CurlOption.proxy, proxy);
+            }
 
             curl.onReceive = (ubyte[] chunk) {
                 result ~= chunk.map!(to!(const(char))).to!string;

diff --git a/source/downloaders.d b/source/downloaders.d
@@ -6,20 +6,22 @@ import std.string : startsWith, indexOf, format, split;
 import std.file : append, exists, read, remove, getSize;
 import std.range : iota;
 import std.net.curl : Curl, CurlOption;
-import helpers : getContentLength, sanitizePath, StdoutLogger, formatSuccess;
+import helpers : getContentLength, sanitizePath, StdoutLogger, formatSuccess, USER_AGENT;
 
 import parsers : YoutubeFormat;
 
 interface Downloader
 {
     void download(string destination, string url, string referer);
+    Downloader setProxy(string proxy);
 }
 
 class RegularDownloader : Downloader
 {
     private StdoutLogger logger;
     private int delegate(ulong length, ulong currentLength) onProgress;
     private bool progress;
+    private string proxy;
 
     this(StdoutLogger logger, int delegate(ulong length, ulong currentLength) onProgress, bool progress = true)
     {
@@ -28,6 +30,12 @@ class RegularDownloader : Downloader
         this.progress = progress;
     }
 
+    public Downloader setProxy(string proxy)
+    {
+        this.proxy = proxy;
+        return this;
+    }
+
     public void download(string destination, string url, string referer)
     {
         auto http = Curl();
@@ -49,13 +57,17 @@ class RegularDownloader : Downloader
 
         auto file = File(destination, "ab");
         http.set(CurlOption.url, url);
-        http.set(CurlOption.useragent, "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:95.0) Gecko/20100101 Firefox/95.0");
+        http.set(CurlOption.useragent, USER_AGENT);
         http.set(CurlOption.referer, referer);
         http.set(CurlOption.followlocation, true);
         http.set(CurlOption.failonerror, true);
         http.set(CurlOption.timeout, 60 * 3);
         http.set(CurlOption.connecttimeout, 60 * 3);
         http.set(CurlOption.nosignal, true);
+        if(proxy != "")
+        {
+            http.set(CurlOption.proxy, proxy);
+        }
 
         http.onReceiveHeader = (in char[]  header) {
             logger.displayVerbose(header);
@@ -83,6 +95,7 @@ class ParallelDownloader : Downloader
     private string title;
     private YoutubeFormat youtubeFormat;
     private bool progress;
+    private string proxy;
 
     this(StdoutLogger logger, string id, string title, YoutubeFormat youtubeFormat, bool progress = true)
     {
@@ -93,6 +106,13 @@ class ParallelDownloader : Downloader
         this.progress = progress;
     }
 
+    public Downloader setProxy(string proxy)
+    {
+        this.proxy = proxy;
+        return this;
+    }
+
+
     public void download(string destination, string url, string referer)
     {
         ulong length = url.getContentLength();
@@ -119,7 +139,7 @@ class ParallelDownloader : Downloader
                 logger.displayVerbose(partialDestination, " already has ", partialDestination.getSize(), " bytes, skipping");
                 continue;
             }
-            new RegularDownloader(logger, (ulong _, ulong __) {
+            auto downloader = new RegularDownloader(logger, (ulong _, ulong __) {
                 if(length == 0)
                 {
                     return 0;
@@ -128,7 +148,9 @@ class ParallelDownloader : Downloader
                 auto percentage = 100.0 * (cast(float)(current) / length);
                 writef!"\r[%.2f %%] %.2f / %.2f MB"(percentage, current / 1024.0 / 1024.0, length / 1024.0 / 1024.0);
                 return 0;
-            }, progress).download(partialDestination, partialLink, url);
+            }, progress);
+            downloader.proxy = proxy;
+            downloader.download(partialDestination, partialLink, url);
         }
 
         writeln();

diff --git a/source/helpers.d b/source/helpers.d
@@ -7,11 +7,13 @@ import std.net.curl : HTTP;
 import std.string : split, indexOf, startsWith;
 import std.format : formattedRead;
 
+immutable string USER_AGENT = "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:95.0) Gecko/20100101 Firefox/95.0";
+
 ulong getContentLength(string url)
 {
     auto http = HTTP(url);
     http.method = HTTP.Method.head;
-    http.addRequestHeader("User-Agent", "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:95.0) Gecko/20100101 Firefox/95.0");
+    http.addRequestHeader("User-Agent", USER_AGENT);
     http.perform();
     if(http.statusLine.code >= 400)
     {

diff --git a/source/parsers.d b/source/parsers.d
@@ -1,5 +1,4 @@
 import std.json;
-import std.net.curl : get;
 import std.uri : decodeComponent, encodeComponent;
 import std.stdio;
 import std.typecons : tuple, Tuple;
@@ -356,11 +355,6 @@ unittest
     assert(extractor.getURL(396) == "https://rr2---sn-f5o5-jhod.googlevideo.com/videoplayback?expire=1677997809&ei=keIDZIHQKMWC1ga62YWIDQ&ip=105.66.0.249&id=o-ADmt4SY6m6445pG7f4G5f72y1NE48ZiWiqWDA9pi6iQo&itag=396&aitags=133%2C134%2C135%2C136%2C137%2C160%2C242%2C243%2C244%2C247%2C248%2C278%2C394%2C395%2C396%2C397%2C398%2C399&source=youtube&requiressl=yes&mh=7c&mm=31%2C29&mn=sn-f5o5-jhod%2Csn-h5q7knes&ms=au%2Crdu&mv=m&mvi=2&pl=24&initcwndbps=275000&vprv=1&mime=video%2Fmp4&ns=V1YGXTHGUU0a4PsRJqmYKX0L&gir=yes&clen=5953258&dur=212.040&lmt=1674230525337110&mt=1677975897&fvip=4&keepalive=yes&fexp=24007246&c=WEB&txp=4537434&n=iRrA3X-4scFA5la&sparams=expire%2Cei%2Cip%2Cid%2Caitags%2Csource%2Crequiressl%2Cvprv%2Cmime%2Cns%2Cgir%2Cclen%2Cdur%2Clmt&lsparams=mh%2Cmm%2Cmn%2Cms%2Cmv%2Cmvi%2Cpl%2Cinitcwndbps&lsig=AG3C_xAwRQIgE-grPIIwKVqUa_siK-FtbLtMME0LPjp9rNlzuvLN7XQCIQCfVt03aw8T9cNgG3u_pFuQafSG4AQeKpgLEHcvodbUjA%3D%3D&sig=AOq0QJ8wRQIgadIMr0vpR2qXdJuUXwsemVtnHk62MbU6kF5SrAfOGlwCIQDYj3buw7XBrdJDtAAUL42iVe5Bfi8PRLVUK3aq-Zc2iA%3D%3D");
 }
 
-YoutubeVideoURLExtractor makeParser(string html, StdoutLogger logger)
-{
-    return makeParser(html, baseJSURL => baseJSURL.get().idup, logger);
-}
-
 YoutubeVideoURLExtractor makeParser(string html, string delegate(string) performGETRequest, StdoutLogger logger)
 {
     if(html.canFind("signatureCipher"))

diff --git a/tests/tests-macos.sh b/tests/tests-macos.sh
@@ -1,15 +1,24 @@
 set -e
 
 chmod +x youtube-d
+mitmdump -q -w proxydump & 1>/dev/null 2>/dev/null
+trap "pkill -INT mitmdump" EXIT
+echo Launching proxy...
+sleep 2
+echo Proxy running in pid $(pgrep mitmdump)
 
-./youtube-d -p --no-progress https://www.youtube.com/watch?v=R85MK830mMo
+echo Installing certificate...
+curl -Lo mitmproxy-ca-cert.pem --proxy http://localhost:8080 http://mitm.it/cert/pem
+sudo security add-trusted-cert -d -p ssl -p basic -k /Library/Keychains/System.keychain mitmproxy-ca-cert.pem
+
+./youtube-d -p --no-progress --proxy http://localhost:8080 https://www.youtube.com/watch?v=R85MK830mMo
 
 filename="Debugging Github actions-R85MK830mMo-18.mp4"
 if [ ! -e "$filename" ]; then
     echo "$filename not found"
     exit 1
 else
-    echo "[1/3] OK, $filename exists"
+    echo "[1/4] OK, $filename exists"
 fi
 
 expected_size=7079820
@@ -18,7 +27,7 @@ if [ $expected_size -ne $actual_size ]; then
     echo "Wrong size. Expected $expected_size, found $actual_size"
     exit 1
 else
-    echo "[2/3] OK, size is correct"
+    echo "[2/4] OK, size is correct"
 fi
 
 expected_hash="e7160d310e79a5a65f382b8ca0b198dd"
@@ -27,5 +36,21 @@ if [ $expected_hash != $actual_hash ]; then
     echo "Wrong hash. Expected $expected_hash, found $actual_hash"
     exit 1
 else
-    echo "[3/3] OK, md5sum is correct"
+    echo "[3/4] OK, md5sum is correct"
 fi
+
+urls=$(mitmdump -nr proxydump -s script.py)
+
+#if [[ $urls =~ "https://www.youtube.com/watch?v=R85MK830mMo" -a $urls =~ "base.js$" -a $urls =~ "googlevideo.com" ]]; then
+for url in  'https://www.youtube.com/watch?v=R85MK830mMo' 'base.js' 'googlevideo.com';
+do
+    if echo $urls | grep -q $url
+    then
+        echo "\t[OK] $url"
+    else
+        echo "[4/4] Missing URL in proxy dump:"
+        echo $url
+        exit 1
+    fi
+done
+echo "[4/4] OK, proxying worked as expected"
diff --git a/tests/tests.sh b/tests/tests.sh
@@ -1,14 +1,24 @@
 set -e
 chmod +x youtube-d
+mitmdump -q -w proxydump & 1>/dev/null 2>/dev/null
+trap "pkill -INT mitmdump" EXIT
+echo Launching proxy...
+sleep 2
+echo Proxy running in pid $(pgrep mitmdump)
 
-./youtube-d -p --no-progress https://www.youtube.com/watch?v=R85MK830mMo
+echo Installing certificate...
+curl -Lo mitmproxy-ca-cert.pem --proxy http://localhost:8080 http://mitm.it/cert/pem
+sudo mv mitmproxy-ca-cert.pem /usr/local/share/ca-certificates/mitmproxy.crt
+sudo update-ca-certificates
+
+./youtube-d -p --no-progress --proxy http://localhost:8080 https://www.youtube.com/watch?v=R85MK830mMo
 
 filename="Debugging Github actions-R85MK830mMo-18.mp4"
 if [ ! -e "$filename" ]; then
     echo "$filename not found"
     exit 1
 else
-    echo "[1/3] OK, $filename exists"
+    echo "[1/4] OK, $filename exists"
 fi
 
 expected_size=7079820
@@ -17,7 +27,7 @@ if [ $expected_size -ne $actual_size ]; then
     echo "Wrong size. Expected $expected_size, found $actual_size"
     exit 1
 else
-    echo "[2/3] OK, size is correct"
+    echo "[2/4] OK, size is correct"
 fi
 
 expected_hash="e7160d310e79a5a65f382b8ca0b198dd"
@@ -26,5 +36,21 @@ if [ $expected_hash != $actual_hash ]; then
     echo "Wrong hash. Expected $expected_hash, found $actual_hash"
     exit 1
 else
-    echo "[3/3] OK, md5sum is correct"
+    echo "[3/4] OK, md5sum is correct"
 fi
+
+urls=$(mitmdump -nr proxydump -s script.py)
+
+#if [[ $urls =~ "https://www.youtube.com/watch?v=R85MK830mMo" -a $urls =~ "base.js$" -a $urls =~ "googlevideo.com" ]]; then
+for url in  'https://www.youtube.com/watch?v=R85MK830mMo' 'base.js' 'googlevideo.com';
+do
+    if echo $urls | grep -q $url
+    then
+        echo "\t[OK] $url"
+    else
+        echo "[4/4] Missing URL in proxy dump:"
+        echo $url
+        exit 1
+    fi
+done
+echo "[4/4] OK, proxying worked as expected"