From 5df84816f5473d2f2662a8b87560dc584ca01a96 Mon Sep 17 00:00:00 2001
From: Peter Monks <pmonks@gmail.com>
Date: Thu, 2 Nov 2023 19:11:26 -0700
Subject: [PATCH] :new: Add normalisation of concluded SPDX expressions as well
 as declared ones

---
 src/lice_comb/impl/matching.clj  | 6 +++++-
 src/lice_comb/impl/utils.clj     | 8 ++++++++
 test/lice_comb/matching_test.clj | 2 +-
 3 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/src/lice_comb/impl/matching.clj b/src/lice_comb/impl/matching.clj
index 0121131..d39b762 100644
--- a/src/lice_comb/impl/matching.clj
+++ b/src/lice_comb/impl/matching.clj
@@ -25,6 +25,7 @@
             [spdx.licenses                   :as sl]
             [spdx.exceptions                 :as se]
             [spdx.matching                   :as sm]
+            [spdx.expressions                :as sexp]
             [lice-comb.impl.spdx             :as lcis]
             [lice-comb.impl.regex-matching   :as lcirm]
             [lice-comb.impl.expressions-info :as lciei]
@@ -132,6 +133,7 @@
         lic (sm/licenses-within-text s @lcis/license-ids-d)
         ids (set/union lic @exc)]
     (when ids
+      ; We don't need to sexp/normalise the keys here, as we never detect an expression from a text
       (manual-fixes (into {} (map #(hash-map % (list {:id % :type :concluded :confidence :high :strategy :spdx-text-matching})) ids))))))
 
 (defmethod text->expressions java.io.Reader
@@ -155,6 +157,7 @@
   are found."
   [uri]
   (when-not (s/blank? uri)
+      ; We don't need to sexp/normalise the keys here, as we never detect an expression from a URI
     (lciei/prepend-source uri
                           (manual-fixes
                             (let [suri (lciu/simplify-uri uri)]
@@ -304,7 +307,8 @@
                                          (map #(if (keyword? %) % (string->ids-info %)))
                                          flatten
                                          seq
-                                         build-expressions-info-map))))))
+                                         build-expressions-info-map
+                                         (lciu/mapfonk sexp/normalise)))))))
 
 (defn init!
   "Initialises this namespace upon first call (and does nothing on subsequent
diff --git a/src/lice_comb/impl/utils.clj b/src/lice_comb/impl/utils.clj
index 4266011..2d7f5b1 100644
--- a/src/lice_comb/impl/utils.clj
+++ b/src/lice_comb/impl/utils.clj
@@ -24,6 +24,14 @@
             [clojure.java.io :as io]
             [clj-base62.core :as base62]))
 
+(defn mapfonk
+  "Returns a new map where f has been applied to all of the keys of m."
+  [f m]
+  (when m
+    (into {}
+          (for [[k v] m]
+            [(f k) v]))))
+
 (defn mapfonv
   "Returns a new map where f has been applied to all of the values of m."
   [f m]
diff --git a/test/lice_comb/matching_test.clj b/test/lice_comb/matching_test.clj
index 33155c0..5ca8682 100644
--- a/test/lice_comb/matching_test.clj
+++ b/test/lice_comb/matching_test.clj
@@ -115,7 +115,7 @@
     (is (valid= #{"MIT" "BSD-4-Clause"}                 (name->expressions "MIT / BSD")))
     (is (valid= #{"Apache-2.0" "GPL-3.0-only"}          (name->expressions "Apache License version 2.0 / GNU General Public License version 3")))
     (is (valid= #{"Apache-2.0" "GPL-3.0-only WITH Classpath-exception-2.0"} (name->expressions "Apache License version 2.0 / GNU General Public License version 3 with classpath exception")))
-    (is (valid= #{"EPL-2.0 OR GPL-2.0-or-later WITH Classpath-exception-2.0 OR MIT OR BSD-3-Clause AND Apache-2.0"} (name->expressions "Eclipse Public License or General Public License 2.0 or (at your discretion) later w/ classpath exception or MIT Licence or three clause bsd and Apache Licence"))))
+    (is (valid= #{"EPL-2.0 OR (GPL-2.0-or-later WITH Classpath-exception-2.0 AND MIT) OR (BSD-3-Clause AND Apache-2.0)"} (name->expressions "Eclipse Public License or General Public License 2.0 or (at your discretion) later w/ classpath exception aNd MIT Licence or three clause bsd and Apache Licence"))))
   (testing "Messed up license expressions"
     (is (valid= #{"Apache-2.0" "MIT"}                   (name->expressions "Apache with MIT"))))
   (testing "Names seen in handpicked POMs on Maven Central"