Download Through a Proxy with Requests and iter_content

with session.get(url, stream=True, proxies=routes,
                 timeout=(5, 30)) as response:
    response.raise_for_status()
    with open(part_path, 'wb') as output:
        for chunk in response.iter_content(chunk_size=64 * 1024):
            if chunk:
                output.write(chunk)

Use this pattern with the application's existing Requests session, proxy routes, URL and temporary output path. For a public HTTP route, the HTTP proxy list supplies candidate addresses for your proxy mapping. The streaming pattern consumes the body incrementally instead of obtaining response.content and then writing a second full-sized object.

Keep that property throughout the download: stream=True defers the body, but accessing response.content afterward can still gather it all into memory. iter_content's chunk size is a read preference, not a request for HTTP chunked transfer coding at the website or proxy.

The response context manager closes the response even if a write fails or the application stops early. The illustrative timeout pair is not a wall-clock deadline for the whole transfer; set the download's lifetime policy separately as needed.

After successful consumption, check the artifact expected by the application before publishing its final name. On interruption, keep the partial file only under a deliberate resume policy, with metadata identifying the remote representation. A nonempty temporary file is not a successful download.