1 files changed, 66 insertions, 0 deletions
diff --git a/skate/unstructured.go b/skate/unstructured.go
new file mode 100644
index 0000000..082c685
--- /dev/null
+++ b/skate/unstructured.go
@@ -0,0 +1,66 @@
+package skate
+
+import (
+	"regexp"
+	"strings"
+)
+
+var (
+	PatDOI         = regexp.MustCompile(`10[.][0-9]{1,8}/[^ ]*[\w]`)
+	PatDOINoHyphen = regexp.MustCompile(`10[.][0-9]{1,8}/[^ -]*[\w]`)
+	PatArxivPDF    = regexp.MustCompile(`https?://arxiv.org/pdf/([0-9]{4,4}[.][0-9]{1,8})(v[0-9]{1,2})?(.pdf)?`)
+	PatArxivAbs    = regexp.MustCompile(`https?://arxiv.org/abs/([0-9]{4,4}[.][0-9]{1,8})(v[0-9]{1,2})?(.pdf)?`)
+
+	urlPrefixes = []string{
+		"http://doi.org/",
+		"https://doi.org/",
+		"http://dx.doi.org/",
+		"https://dx.doi.org/",
+	}
+)
+
+// ParseUnstructured will in-place augment missing DOI, arxiv id and so on.
+func ParseUnstructured(ref *Ref) error {
+	var (
+		uns = ref.Biblio.Unstructured
+		v   string
+		vs  []string
+	)
+	// Handle things like: 10.1111/j.1550-7408.1968.tb02138.x-BIB5|cit5,
+	// 10.1111/j.1558-5646.1997.tb02431.x-BIB0008|evo02431-cit-0008, ...
+	if strings.Contains(strings.ToLower(ref.Key), "-bib") && ref.Biblio.DOI == "" {
+		parts := strings.Split(strings.ToLower(ref.Key), "-bib")
+		ref.Biblio.DOI = parts[0]
+	}
+	// DOI
+	v = PatDOI.FindString(uns)
+	if v != "" && ref.Biblio.DOI == "" {
+		ref.Biblio.DOI = v
+	}
+	// DOI in Key
+	v = PatDOINoHyphen.FindString(ref.Key)
+	if v != "" && ref.Biblio.DOI == "" {
+		ref.Biblio.DOI = v
+	}
+	// DOI in URL
+	for _, prefix := range urlPrefixes {
+		if ref.Biblio.DOI != "" && strings.HasPrefix(ref.Biblio.Url, prefix) {
+			ref.Biblio.DOI = strings.Replace(ref.Biblio.Url, prefix, "", -1)
+		}
+	}
+	v = PatDOINoHyphen.FindString(ref.Key)
+	if v != "" && ref.Biblio.DOI == "" {
+		ref.Biblio.DOI = v
+	}
+	// Arxiv
+	vs = PatArxivPDF.FindStringSubmatch(uns)
+	if len(vs) != 0 && ref.Biblio.ArxivId == "" {
+		ref.Biblio.ArxivId = vs[1]
+	} else {
+		vs = PatArxivAbs.FindStringSubmatch(uns)
+		if len(vs) != 0 && ref.Biblio.ArxivId == "" {
+			ref.Biblio.ArxivId = vs[1]
+		}
+	}
+	return nil
+}