From 5e4f57a8b360ed0bbd1ffdd3d4dde8e161b0fd6b Mon Sep 17 00:00:00 2001
From: Thomas Sileo <t@a4.io>
Date: Sun, 18 Aug 2019 11:40:57 +0200
Subject: [PATCH] Proxy media that aren't cached

---
 app.py                    | 20 ++++++++++++++++++++
 utils/template_filters.py |  2 +-
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/app.py b/app.py
index 4dd1e1b..b50d6c3 100644
--- a/app.py
+++ b/app.py
@@ -3,8 +3,10 @@ import logging
 import os
 import traceback
 from datetime import datetime
+from typing import Any
 from uuid import uuid4
 
+import requests
 from bson.errors import InvalidId
 from bson.objectid import ObjectId
 from flask import Flask
@@ -226,12 +228,30 @@ def robots_txt():
 
 @app.route("/microblogpub-0.0.jsonld")
 def microblogpub_jsonld():
+    """Returns our AP context (embedded in activities @context)."""
     return Response(
         response=json.dumps(jsonld.MICROBLOGPUB),
         headers={"Content-Type": "application/ld+json"},
     )
 
 
+@app.route("/p/<path:url>")
+def proxy(url: str) -> Any:
+    req_headers = {
+        k: v
+        for k, v in dict(request.headers).items()
+        if k.lower() not in ["host", "cookie"]
+    }
+    resp = requests.get(url, stream=True, headers=req_headers)
+    app.logger.info(f"proxied req {url}: {resp!r}")
+
+    def data():
+        for chunk in resp.raw.stream(decode_content=False):
+            yield chunk
+
+    return Response(data(), headers=dict(resp.raw.headers))
+
+
 @app.route("/media/<media_id>")
 @noindex
 def serve_media(media_id):
diff --git a/utils/template_filters.py b/utils/template_filters.py
index 7eb3701..c31b585 100644
--- a/utils/template_filters.py
+++ b/utils/template_filters.py
@@ -249,7 +249,7 @@ def _get_file_url(url, size, kind) -> str:
 
     # MEDIA_CACHE.cache(url, kind)
     _logger.error(f"cache not available for {url}/{size}/{kind}")
-    return url
+    return f"/p/{url}"
 
 
 @filters.app_template_filter()