How to filter output of tail with babashka

Question

I want to filter and modify output of tail command. This is what I come up with:

#!/usr/bin/env bb

(ns script
  (:require
   [clojure.java.io :as io]
   [clojure.string :as str]
   ))

(->> (line-seq (io/reader *in*)
               (filter #(re-find #"^\[.*CONSOLE" %))
               (map #(str "carpenter " %)))

It works for normal tail. But I want to use it for "tail -f" command. Any ideas?

Thx

Can you describe the issue you are running into with `tail -f`? — Nathan Davis, Sep 20 '21 at 20:12
Above script works with "tail /tmp/data.txt | script.clj" but fail for "tail -f /tmp/data.txt | script.clj" — serezlan, Sep 21 '21 at 00:25
Ok. but there are many ways things can fail. Can you describe the way in which it is failing? — Nathan Davis, Sep 21 '21 at 22:17

score 2 · Accepted Answer · answered Sep 19 '21 at 09:18

This example starts writing to a file two kinds of messages: HELLO and BYE. Then it starts a tail -f process to watch the file and then reads from the output of that process and only captures the BYE lines and prints them with a custom string in front.

(ns tail-example
  (:require [babashka.process :as p]
            [clojure.java.io :as io]))

(future
  (loop []
    (spit "my-file.txt" "HELLO\n" :append true)
    (spit "my-file.txt" "BYE\n" :append true)
    (Thread/sleep 1)
    (recur)))

(def tail (p/process
           (p/tokenize "tail -f my-file.txt")
           ;; send stderr to stderr of bb, leave out stream unmodified
           {:err :inherit}))

(let [rdr (io/reader (:out tail))]
  (binding [*in* rdr]
    (loop []
      (when-let [l (read-line)]
        (when (re-matches #"BYE" l)
          (println (str "[log] " l)))
        (recur)))))

How to filter output of tail with babashka

1 Answers1