Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
rasguanabana
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
12 ms
·
31.
▲
by
rasguanabana
7y ago
It’s a real pain to handle subprocesses in Python. If you need to automate certain command based workflows, Bash scripts are much easier, both to write and read, until they reach certain size. At my previous job it was a daily task and scri
32.
▲
What Is Minimum Viable Product? [video]
(youtube.com)
2 points
by
rasguanabana
7y ago
|
0 comments
33.
▲
by
rasguanabana
8y ago
I have no data supporting this (ironically, just intuition), but examples you give might only be a small chunk of attempts that brought a successful outcome. In other words, when one tries to achieve something, they'll initially fail a
34.
▲
Show HN: Xml_browser.py – Edit XML documents as a directory structure
(github.com)
2 points
by
rasguanabana
10y ago
|
0 comments
35.
▲
by
rasguanabana
10y ago
Thank you for suggestion, I have added a simple example.
36.
▲
Show HN: Pattern matching and guards for Python functions
(github.com)
6 points
by
rasguanabana
10y ago
|
4 comments
37.
▲
by
rasguanabana
11y ago
I have to think that through. Search pages are tricky as you can't simply go to a certain page. They're identified by tokens, which aren't known a priori - you just get adjacent ones with search results. A reasonable solution
38.
▲
by
rasguanabana
11y ago
There's no authorisation at the moment. Plain requests.get is used.
39.
▲
by
rasguanabana
11y ago
That is going to be the first thing to implement now.
40.
▲
by
rasguanabana
11y ago
Maybe using Fourier/wavelet/whatsoever transform would be the way to go, just like in digital watermarking techniques. Both high capacity and robustness would seem easier to achieve.
41.
▲
by
rasguanabana
11y ago
Oh, I don't want to clutter it too much, it's better to keep it simple. It definitely needs a mount option to set the default quality. Changing it later might be tricky, maybe passing options in directory names would be a good ide
42.
▲
by
rasguanabana
11y ago
Some time ago Youtube started to provide audio and video separately. Due to that, for full videos whole data needs to be downloaded before reading (it's hard to merge them on-the-fly). For just audio or just video data streaming is int
43.
▲
by
rasguanabana
11y ago
If there was a way to do massive changes in existing videos it would be feasible to use them as storage - that would be a definitely interesting approach. Anyway, advanced data encoding would be needed to prevent it from damage after format
44.
▲
by
rasguanabana
11y ago
It's read-only - you can play/copy youtube movies as a file
45.
▲
Show HN: YouTube Filesystem – YTFS
(github.com)
185 points
by
rasguanabana
11y ago
|
61 comments