Hmm, that’s weird. Do you have a gpu attached?
yeah, 1050 ti
I think it might be because you may not have CUDA installed. It should be defaulting to cpu backend but I guess it didn’t in your case. Can you install CUDA toolkit (11.7) from nvidia’s website onto your system?
i got the conclusion it isnt only showing, because the terminal is raising cpu usage
what if run the thing on a website like this?
You’ll need to get a python runtime on it as well. Additionally, replit is very slow (not an exaggeration). I’d try using a VM to get it working instead of bare metal
i just gave up, hoping someday it wil work on my pc, thanks for your help
Another update: Recovering twitch vods will now return the results in sorted descending order (before it was completely random). Additionally, the output will display the date and time the vod was started
could you give a typed out example of how you’d use the filtering on twitch?
chat_reader --filter "(?i)(burp|belch)" twitch channel somechannel --clips --vods
Goes through every vod on the channel and only prints out messages/users from chat containing ‘burp’ or ‘belch’ (case insensitive)
The replit instance of chat_reader uses old code and has not been updated in a long time. It is recommended to use the native version. I’ll be updating it shortly so that it hooks up to the latest version of chat_reader.
The replit has been updated and now uses the latest version of chat_reader, including the speed improvements and different search functions
That is what the filter is for
Hello, I am also hitting the same issue when running the twitch channel search:
error decoding response body: EOF while parsing a value at line 1 column 0
I am using the native version, though also notice the same issue on the replit. I tried with a few different usernames and it’s a consistent response. Any clues or suggestions?
Are you using the most recently updated version for the native version? If so, pm me some details about what you’re typing to produce this message as well as the replit
That was it, I was using an old version. All good now, thanks a lot!
where is the current version?
Top of the thread:
https://drive.google.com/drive/folders/1ZCUWVsjN-Y2kKFQOJ1L-5JS87wcZz4zF
The tool will only find whatever strings match with the regular expression/filter provided, nothing more, nothing less. If you have questions related to matching- it would be best to consult with a regex reference
Yes there is a problem, currently looking into it
A patch has been made to stop the thread crashes and the builds have been updated for the cli client (replit is still using the old version) releases – Google Drive.
The problem was due to a removal on twitch’s part for m3u8 extraction on highlights. As noted, this is a patch, but not necessarily a fix- at the moment, highlights will not have an m3u8 url generated for them, as such you may need to get it another way. Vod archives should continue to display the urls as usual. There are no changes to any other systems, such as those handling comment reading for twitch or other services, like afreecatv.