If the title of a large number of subscriptions can be translated literally, it can greatly increase the information filtering ability, which is a great improvement to the reading experience. If the title is found to be of interest, it can be opened in the original browser and translated in full
I just spent 20 minutes adding a rewriting rule that simply uses the ChatGPT-4 model to translate short titles. The token consumption and speed are acceptable. The results look good, allowing users with a lot of English RSS feeds to quickly filter content like a native language. I’m wondering if using a rewriting rule is a good idea and if I should submit a PR for this feature. I also spent an hour trying to add a checkbox on the edit page like a scraping proxy, but that would require a lot of code changes, which may not be a good idea. On the other hand, will the rewriting rules only apply to new content, or will they also trigger for old content?
If the title of a large number of subscriptions can be translated literally, it can greatly increase the information filtering ability, which is a great improvement to the reading experience. If the title is found to be of interest, it can be opened in the original browser and translated in full
I just spent 20 minutes adding a rewriting rule that simply uses the ChatGPT-4 model to translate short titles. The token consumption and speed are acceptable. The results look good, allowing users with a lot of English RSS feeds to quickly filter content like a native language. I’m wondering if using a rewriting rule is a good idea and if I should submit a PR for this feature. I also spent an hour trying to add a checkbox on the edit page like a scraping proxy, but that would require a lot of code changes, which may not be a good idea. On the other hand, will the rewriting rules only apply to new content, or will they also trigger for old content?