Optimize downloadMedia

[brunoferreiras]

I had the same problem as described on #384

This pull request introduces a significant optimization in a specific function of the code. I replaced the usage of io.ReadAll with io.Copy to improve memory efficiency.

In my tests, I used a document of 10Mb using the go tool pprof to get this.

Before this change:

After this change:

This change aims to optimize memory consumption when processing large amounts of data. Additionally, I included examples of using go tool pprof to demonstrate performance improvements.

[tulir]

That doesn't make any sense. You need 10mb of memory to download a 10mb file into memory, so my first guess would be that the second pprof is hiding the actual memory usage somewhere else. bytes.Buffer also shouldn't be any different from io.ReadAll, both of them have a byte slice as a buffer and expand it dynamically as they read data.

A potential small optimization would be pre-allocating the correct size of byte array and using io.ReadFull, rather than having bytes.Buffer or io.ReadAll dynamically grow a byte slice.