cleanup

dataFormat.go

@@ -6,6 +6,8 @@ import (
 	"fmt"
 )
 
+var ErrUnknownDataFormat = fmt.Errorf("Unknown Data Format")
+
 // DataFormat represents different types of data
 type DataFormat int
 
@@ -63,7 +65,7 @@ func ParseDataFormatString(s string) (df DataFormat, err error) {
 // MarshalJSON satisfies the json.Marshaler interface
 func (f DataFormat) MarshalJSON() ([]byte, error) {
 	if f == UnknownDataFormat {
-		return nil, nil
+		return nil, ErrUnknownDataFormat
 	}
 	return []byte(fmt.Sprintf(`"%s"`, f.String())), nil
 }

hash.go

@@ -7,16 +7,20 @@ import (
 	"github.com/multiformats/go-multihash"
 )
 
-// CalcHash calculates the hash of a json.Marshaler
-// TODO - this will have to place nice with IPFS block hashing strategies
-func CalcHash(m json.Marshaler) (hash string, err error) {
-	h := sha256.New()
-
+// JSONHash calculates the hash of a json.Marshaler
+func JSONHash(m json.Marshaler) (hash string, err error) {
 	// marshal to cannoncical JSON representation
 	data, err := m.MarshalJSON()
 	if err != nil {
 		return
 	}
+	return HashBytes(data)
+}
+
+// TODO - this will have to place nice with IPFS block hashing strategies
+func HashBytes(data []byte) (hash string, err error) {
+	h := sha256.New()
+
 	if _, err = h.Write(data); err != nil {
 		return
 	}

hash_test.go

@@ -0,0 +1,28 @@
+package dataset
+
+import (
+	"testing"
+)
+
+func TestHashBytes(t *testing.T) {
+	cases := []struct {
+		in  []byte
+		out string
+		err error
+	}{
+		{[]byte(""), "1220e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", nil},
+	}
+
+	for i, c := range cases {
+		got, err := HashBytes(c.in)
+		if err != c.err {
+			t.Errorf("case %d error mismatch. expected: %s got: %s", i, c.err, err)
+			continue
+		}
+
+		if got != c.out {
+			t.Errorf("case %d result mismatch. expected: %s got: %s", i, c.out, got)
+			continue
+		}
+	}
+}

resource.go

@@ -2,7 +2,6 @@ package dataset
 
 import (
 	"encoding/json"
-	"errors"
 	"github.com/ipfs/go-datastore"
 )
 
@@ -40,7 +39,7 @@ type Resource struct {
 
 // Hash gives the hash of this resource
 func (r *Resource) Hash() (string, error) {
-	return CalcHash(r)
+	return JSONHash(r)
 }
 
 // truthCount returns the number of arguments that are true

resource_test.go

@@ -8,6 +8,29 @@ import (
 	"testing"
 )
 
+func TestResouceHash(t *testing.T) {
+	cases := []struct {
+		r    *Resource
+		hash string
+		err  error
+	}{
+		{&Resource{Format: CsvDataFormat}, "1220c2f881931bffda4b33de1fcc9c6085b4d4b9dcc5d18083d97c6415c1a3590b66", nil},
+	}
+
+	for i, c := range cases {
+		hash, err := c.r.Hash()
+		if err != c.err {
+			t.Errorf("case %d error mismatch. expected %s, got %s", i, c.err, err)
+			continue
+		}
+
+		if hash != c.hash {
+			t.Errorf("case %d hash mismatch. expected %s, got %s", i, c.hash, hash)
+			continue
+		}
+	}
+}
+
 func TestResourceUnmarshalJSON(t *testing.T) {
 	cases := []struct {
 		FileName string
@@ -30,7 +53,7 @@ func TestResourceUnmarshalJSON(t *testing.T) {
 			continue
 		}
 
-		if err = ResourceEqual(ds, c.result); err != nil {
+		if err = CompareResources(ds, c.result); err != nil {
 			t.Errorf("case %d resource comparison error: %s", i, err)
 			continue
 		}
@@ -42,7 +65,7 @@ func TestResourceMarshalJSON(t *testing.T) {
 
 }
 
-func ResourceEqual(a, b *Resource) error {
+func CompareResources(a, b *Resource) error {
 	if a == nil && b == nil {
 		return nil
 	} else if a == nil && b != nil || a != nil && b == nil {