From ca7f5d2f9598c81b1fe4cf5ea50d435ff7d929e4 Mon Sep 17 00:00:00 2001
From: Stefan Kamphausen <stefan.kamphausen@acrolinx.com>
Date: Sun, 20 May 2018 22:35:17 +0200
Subject: [PATCH] Working prototype for license normalization

---
 build.boot                               |  2 +-
 resources/license-normalizations.edn     | 28 +++++++++++++++++
 src/de/clojure_buch/boot_attribution.clj | 39 +++++++++++++++++++-----
 3 files changed, 60 insertions(+), 9 deletions(-)
 create mode 100644 resources/license-normalizations.edn

diff --git a/build.boot b/build.boot
index 26a1f0f..9c2e5e9 100644
--- a/build.boot
+++ b/build.boot
@@ -16,7 +16,7 @@
 ;; Now, boot
 (set-env!
  :source-paths #{"src/"}
- ;; resource-paths
+ :resource-paths #{"resources/"}
  :dependencies (into runtime-deps test-deps)
  )
 
diff --git a/resources/license-normalizations.edn b/resources/license-normalizations.edn
new file mode 100644
index 0000000..00756b2
--- /dev/null
+++ b/resources/license-normalizations.edn
@@ -0,0 +1,28 @@
+;; Keys are taken from SPDX. Values are sets or vactors of variations
+;; found in real-world examples.
+
+;; All comparisons will be done case insensitively at runtime. 
+{"Apache-2.0"
+ ["Apache 2"
+  "Apache 2.0 License"
+  "Apache License"
+  "Apache License 2.0"
+  "Apache License Version 2.0"
+  "Apache License, Version 2.0"
+  "Apache Software License - Version 2.0"
+  "ASF 2.0"]
+
+ "EPL-1.0"
+ ["Eclipse Public License"
+  "Eclipse Public License 1.0"
+  "Eclipse Public License v1.0"
+  "Eclipse Public License - v 1.0"
+  "EPL"]
+
+ "LGPL"
+ ["GNU Lesser General Public Licence"
+  "GNU Lesser General Public License"
+  "GNU Lesser General Public License, Version 2.1"
+  "GNU Lesser Public License"
+  "GNU Library or Lesser General Public License"]
+ }
diff --git a/src/de/clojure_buch/boot_attribution.clj b/src/de/clojure_buch/boot_attribution.clj
index cc66640..31df219 100644
--- a/src/de/clojure_buch/boot_attribution.clj
+++ b/src/de/clojure_buch/boot_attribution.clj
@@ -1,6 +1,8 @@
 (ns de.clojure-buch.boot-attribution
   {:boot/export-tasks true}
   (:require [clojure.java.io :as io]
+            [clojure.edn     :as edn]
+            [clojure.string  :as string]
             [clojure.xml     :as xml]
             [boot.core       :as boot :refer [deftask]]
             [boot.pod        :as pod]
@@ -52,19 +54,21 @@
        (catch Throwable _ nil)))
 
 
-(defn- library-attributions [the-dep strategies]
+(defn- library-attributions
+  [the-dep strategies lnormalizer]
   (let [[lib vers jar] (unfold-dep the-dep)]
     (util/info "Resolving POM for %s %s %s\n" lib vers jar)
     (some (fn resolve-license-for [strategy]
-            (resolve-licenses strategy jar))
+            (->> jar
+                 (resolve-licenses strategy)
+                 (map lnormalizer)))
           strategies)))
 
 ;; FIXME: want more strategies:
 ;; - find a LICENSE(.txt) or similar
 ;; - Grep in a README(.md,txt,org)
 (def ^:private license-strategies
-  [::pom
-   ::unknown])
+  [::pom ::unknown])
 
 
 ;; FIXME
@@ -72,10 +76,10 @@
   (let [[lib vers jar] (unfold-dep the-dep)]
     "NYI"))
 
-(defn all-attributions [deps]
+(defn all-attributions [deps strategies lnormalizer]
   (for [dep deps]
     {:dep dep
-     :licenses (library-attributions dep license-strategies)
+     :licenses (library-attributions dep strategies lnormalizer)
      :copyrights (copyright-attributions dep)}))
 
   
@@ -83,10 +87,29 @@
   (pod/resolve-dependencies env))
 
 
+(defn- make-license-normalizer [norms]
+  (let [lookup
+        (into {}
+              (for [[lic alt-names] norms
+                    an alt-names]
+                {(string/lower-case an) lic}))]
+    (fn license-normalizer [license]
+      (if-let [normalized
+               (lookup (string/lower-case license))]
+        normalized
+        license))))
+
+(defn- license-normalizer []
+  (->
+   (io/resource "license-normalizations.edn")
+   slurp
+   edn/read-string
+   make-license-normalizer))
 
 (defn repl-test []
-  (-> (all-deps (boot/get-env))
-      all-attributions))
+  (all-attributions (all-deps (boot/get-env))
+                    license-strategies
+                    (license-normalizer)))