add benchmarks

2 files changed, 48 insertions, 4 deletions

diff --git a/skate/schema.go b/skate/schema.go
index 52aa91a..f97b55a 100644
--- a/skate/schema.go
+++ b/skate/schema.go
@@ -59,13 +59,13 @@ func RefToRelease(ref *Ref) (*Release, error) {
 	}
 	release.Contribs = contribs
 	if strings.Contains(strings.ToLower(ref.Biblio.Unstructured), "isbn") {
-		release.ExtIDs.ISBN = parseIsbn(ref.Biblio.Unstructured)
+		release.ExtIDs.ISBN = ParseIsbn(ref.Biblio.Unstructured)
 	}
 	return &release, nil
 }
 
-// parseIsbn tries to find and validate ISBN from unstructured data.
-func parseIsbn(s string) []string {
+// ParseIsbn tries to find and validate ISBN from unstructured data.
+func ParseIsbn(s string) []string {
 	// ISBN: 10: 0137822693, pp: 373
 	// Robotec, E. (1996). Scorbot ER VII, User's Manual, Eshed Robotec,
 	// ISBN9652910333. Shannon, C. (1948). A Mathematical Theory of
@@ -126,6 +126,10 @@ type Biblio struct {
 	Url             string   `json:"url,omitempty"`
 	Volume          string   `json:"volume,omitempty"`
 	Year            int64    `json:"year,omitempty"`
+	// More non-standard fields go into extra.
+	Extra struct {
+		ISBN []string `json:"isbn"`
+	} `json:"extra"`
 }
 
 // Ref is a reference document, can be very partial.
diff --git a/skate/schema_test.go b/skate/schema_test.go
index 0f84703..56d2d5b 100644
--- a/skate/schema_test.go
+++ b/skate/schema_test.go
@@ -198,9 +198,49 @@ func TestParseIsbn(t *testing.T) {
 		},
 	}
 	for _, c := range cases {
-		r := parseIsbn(c.s)
+		r := ParseIsbn(c.s)
 		if !reflect.DeepEqual(r, c.result) {
 			t.Fatalf("got %v, want %v", r, c.result)
 		}
 	}
 }
+
+func BenchmarkParseIsbn(b *testing.B) {
+	for n := 0; n < b.N; n++ {
+		ParseIsbn("House Pvt. Limited., (2006), ISBN 9788183561426. Date accessed: August 2015.")
+	}
+}
+
+func BenchmarkRefToRelease(b *testing.B) {
+	var ref Ref
+	_ = json.Unmarshal([]byte(`
+    {
+      "biblio": {
+        "arxiv_id": "123",
+        "container_name": "IEEE Trans. Pattern Anal. Mach. Intell",
+        "contrib_raw_names": [
+          "M Ben-Ezra",
+          "S K Nayar"
+        ],
+        "doi": "123",
+        "issue": "6",
+        "pages": "689-698",
+        "pmcid": "123",
+        "publisher": "ABC",
+        "title": "Motion-based motion deblurring",
+        "unstructured": "M. Ben-Ezra and S. K. Nayar. Motion-based motion deblurring. IEEE Trans. Pattern Anal. Mach. Intell., 26(6):689-698, 2004. 2",
+        "url": "https://abc.com",
+        "volume": "26",
+        "year": 2004
+      },
+      "index": 0,
+      "key": "b0",
+      "ref_source": "grobid",
+      "release_ident": "26qgat7mzrerjacrlsz3gdmcgy",
+      "release_year": 2014,
+      "work_ident": "aaaoe2wcbvdjthnv36dlqgkray"
+    }`), &ref)
+	for n := 0; n < b.N; n++ {
+		_, _ = RefToRelease(&ref)
+	}
+}