smart_open VS OkHttp

We’re using smart_open, which is an amazing library that lets you open objects in S3 (and other cloud object stores) as if they’re files on your filesystem. It’s obviously critical that we’re able to seek to an arbitrary position in an S3 file without first downloading the whole thing. We’ll assume you’re using Poetry, but you should be able to follow along with any other package manager:

If you didn’t follow along with the first article in this series, you should be able to follow this article with your own dataset as long as you install smart_open and Meadowrun. smart_open is an amazing library that lets you open objects in S3 (and other cloud object stores) as if they’re files on your filesystem, and Meadowrun makes it easy to run your Python code on the cloud.

I see that you're primarily looking into Python work, so I'd recommend `smart_open` as a nice, compact way to get started.

https://github.com/RaRe-Technologies/smart_open

You want smart_open. It gives you a (more complete) file-like interface to many different storage systems, including s3. You can read and seek as needed.

See also smart_open: https://github.com/RaRe-Technologies/smart_open which might be more user-friendly? Never used it myself but it was on HN before. Discussion on their bugtracker: https://github.com/RaRe-Technologies/smart_open/issues/579

We will use Square’s Mock Webserver to spin up a mock server which we can use to simulate real api's request to the get coffee endpoint.

OkHttp Documentation

mockwebserver -> https://github.com/square/okhttp/tree/master/mockwebserver

I searched in the OkHttp GitHub project for an advice on which values may be suitable for Android apps nowadays but found no answers (only this old issue which does not help). Since we share a single OkHttp client Singleton for all our retrofit APIs and even Coil, I wonder if the default 5 maxRequestsPerHost is really enough.

You can technically add a Content-Length header for each part. It's not forbidden by the RFC, but nor is it common. It caused [problems](https://github.com/square/okhttp/issues/2138) for OkHttp, and they eventually removed it. Might be fine for internal-only use, though.

Boundaries are a lot like UUIDs, and rely on the same logic. When generating random data, once you have enough bits, the odds are against that sequence of bits ever having been generated before in the universe.

Lets be specific. This is the gradle build file for Squares okhttp client library. How exactly would your bld tool "predict" or "help" with all the parameters needed? There is no need to be defensive. Replace those large build files with your own, show where your approach is better and then understanding will lead to better solutions.

And adding to that: The asynchttpclient library is just a thin wrapper around OkHttp3, so it might be easier to just go with that instead: https://square.github.io/okhttp/

Retrofit in combination with OkHttp for fetching data from server (which hopefully already exists)

OkHttp