A commandline tool that counts the number of word occurences in an input.
This is just a placeholder repository for now. Please create issues for feature request and collaboration.
- Ignore words (regex pattern)
- Filter stopwords (similar to NLTK's stopwords)
- Different output formats (plaintext, JSON)
- Performance (SIMD support, async execution)
- Usage as library
- Recursion support
- Allow skipping files
- Allow specifying ignored words in a separate file
- Generate "heat bars" for words like shell-hist does
- Split report by file/folder (sort of like
slocdoes for code) - Choose language for stopwords
- Format output (e.g. justify counts a la
uniq -c) - Calculate TF-IDF score in a multi-file scenario
Idea contributors:
- @jamesmunns
- Want to see your name here? Create an issue!
uniq
A basic version would be
curl -L 'https://github.com/mre/freq/raw/main/README.md' | tr -cs '[:alnum:]' "\n" | grep -vEx 'and|or|for|a|of|to|an|in' | sort | uniq -c | sort
This works, but it's not very extensible by normal users. It would also lack most of the features listed above.
Lucene
Has all the bells and whistles, but there is no official CLI interface and requires a full Java installation.
wordcount
freqword <tab> freq
Nice and simple. Doesn't exclude stopwords and no regex support, though. https://github.com/juditacs/wordcount
word-frequency
Haskell-based approach: Includes features like min length for words, or min occurrences of words in a text. https://github.com/cbzehner/word-frequency
What else?
There must be more tools out there. Can you help me find them?