Apache Tika has worked well for me in the past, ended up running it on an AWS Lambda https://tika.apache.org/. - Source: Hacker News / 10 months ago

If you accept running Java, the Apache Tika is extremely good at parsing content (https://tika.apache.org/). - Source: Hacker News / 11 months ago

Apache Tika can spit out text from lots of formats. I've used it with grep (or rg) to make a small scale searching of local folders. Tika does a really good job at OCR for finding if text is in a file. Source: about 1 year ago

Https://tika.apache.org Meta data from things. Source: over 1 year ago

At my previous job we had the same problem which we solved by using Tika. We called it on the server along with other stuff, but there is also a Python binding. Source: almost 2 years ago

I'd suggest using both FakeSpot and ReviewMeta together. In the end though, neither should be treated as perfect. They both can give skewed ratings on legitimately good products. It just helps give an idea of if you should be cautious and look into things more before purchasing. Source: 11 months ago

I recommend https://reviewmeta.com if you want to see if the reviews are authentic or not. Source: 12 months ago

This does a pretty good job of filtering out the crap - https://reviewmeta.com/. Source: about 1 year ago

This is only for amazon, but https://reviewmeta.com/ checks if reviews under a product are likely to be real or not. Source: about 1 year ago

I'll also double-check those Amazon reviews with a tool like Reviewmeta to see if they have a high number of suspicious or fake-sounding ratings or comments. Source: about 1 year ago