Fipe is a simple Clojure library for representing and running file pipelines. It is similar to tools like GNU Make and Factual's Drake, but allows defining targets with simple Clojure macros, allowing flexibility similar to that of Rake, but for Clojure, not Ruby.
Note that dependency resolution is not yet implemented, but this library can still be very useful by using fipe-glob, sometimes combined with doseq, to produce many targets at once.
This is an alpha release. The API is subject to change (especially for globbing and dependencies). Comments and contributions are much appreciated.
Fipe release are published on Clojars.
To use this version with Leiningen, add the following dependency to your project definition:
[mitchellkoch/fipe "0.1.0"]Use the deftarget macro to define how a target file should be produced. All targets are relative to fipe.core/data-dir, which by default is data/. The function fipe-rel gives files relative to this path.
(ns example
(:use fipe.core :only [deftarget dep fipe])
(:require [clojure.string :as str]))
(deftarget "hello.txt"
"hello")
(defn x4 [f]
(let [s (str/trim-newline (slurp f))]
(vec (repeat 4 s))))
(deftarget "hellos.edn"
(x4 (dep "hello.txt")))Here, running (fipe "hello.txt") will produce hello.txt (containing hello) or (fipe-glob "hello*") will produce both hello.txt and hellos.edn, (where hellos.edn contains ["hello" "hello" "hello" "hello"]).
Fipe uses the file's extension to determine the output format for writing the files (see fipe.util/write-to-file). It currently supports .txt, a naive .tsv and .tsv.gz, csv using clojure.data.csv, json using cheshire, edn with pretty printing, and .ser.gz for serialized Java objects.
Additionally, deftarget! can be used to just run for side effects, in which case Fipe does not write the value to the target file itself.
Copyright © 2014 Mitchell Koch. Distributed under the Eclipse Public License, the same as Clojure.