feat: add proxy-all

docker-compose.yml

@@ -70,3 +70,12 @@ services:
         - IMAGE_TAG=${IMAGE_TAG:?}
     privileged: true
     image: ohmymndy/virter:${IMAGE_TAG:?}
+
+  proxy-all:
+    build:
+      context: ./proxy-all
+      args:
+        - IMAGE_TAG=${IMAGE_TAG:?}
+    ports:
+      - 9901:5000
+    image: ohmymndy/proxy-all:${IMAGE_TAG:?}

proxy-all/Dockerfile

@@ -0,0 +1,9 @@
+FROM python:3.13.2
+
+RUN pip install Flask requests beautifulsoup4 Pygments
+
+RUN mkdir -p /app
+COPY ./src /app 
+
+USER nobody
+ENTRYPOINT [ "/app/app.py" ]

proxy-all/src/app.py

@@ -0,0 +1,82 @@
+#!/usr/bin/env python
+
+from flask import Flask, request, Response
+import requests
+import urllib.parse
+import logging
+from bs4 import BeautifulSoup
+from pygments.lexers import guess_lexer
+
+app = Flask(__name__)
+
+logging.basicConfig(level=logging.INFO)
+
+ALLOWED_DOMAINS = []
+
+
+def is_allowed(url):
+    return True
+
+
+def prepend_base_url(base_url, target_base_url, url):
+    if url.startswith(("http://", "https://")):
+        return base_url + "?q=" + urllib.parse.quote_plus(url)
+    else:
+        return (
+            base_url
+            + "?q="
+            + urllib.parse.quote_plus(target_base_url + url.lstrip("/"))
+        )
+
+
+@app.route("/", methods=["GET"])
+def proxy():
+    target_url = request.args.get("q")
+    if not target_url:
+        return Response("Missing 'q' query parameter.", status=400)
+
+    if not is_allowed(target_url):
+        return Response("URL not allowed.", status=403)
+
+    try:
+        headers = {key: value for key, value in request.headers if key != "Host"}
+        headers = {}
+        resp = requests.get(
+            target_url, headers=headers, stream=True, timeout=10, allow_redirects=True
+        )
+
+        if resp.status_code != 200:
+            return resp
+
+        excluded_headers = ["content-encoding", "transfer-encoding", "connection"]
+        headers = [
+            (name, value)
+            for (name, value) in resp.raw.headers.items()
+            if name.lower() not in excluded_headers
+        ]
+
+        logging.info(f"Proxied URL: {target_url} | Status: {resp.status_code}")
+
+        soup = BeautifulSoup(resp.content, "html.parser")
+        language = guess_lexer(resp.content).name
+
+        content = resp.content
+        if bool(soup.find()) and language in ["XML", "HTML"]:
+            for tag in soup.find_all(href=True):
+                tag["href"] = prepend_base_url(
+                    request.base_url, target_url, tag["href"]
+                )
+
+            for tag in soup.find_all(src=True):
+                tag["src"] = prepend_base_url(request.base_url, target_url, tag["src"])
+
+            content = soup.encode("utf8")
+
+        return Response(content, resp.status_code, headers)
+    except requests.exceptions.RequestException as e:
+        logging.error(f"Error fetching {target_url}: {e}")
+        return Response(f"Error fetching the URL: {e}", status=500)
+
+
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=5000)

version.txt

@@ -1 +1 @@
-0.3
+0.4