CR changes and updated DROID XML

richardlehane · Nov 5, 2023 · a8c5b49 · a8c5b49
1 parent f2f643d
commit a8c5b49
Show file tree

Hide file tree

Showing 5 changed files with 22 additions and 24 deletions.
diff --git a/cmd/roy/data/DROID_SignatureFile_V114.xml b/cmd/roy/data/DROID_SignatureFile_V114.xml
@@ -1,4 +1,5 @@
 <?xml version="1.0" encoding="UTF-8"?>
+<!-- THIS FILE WILL BE REPLACED BEFORE MERGE -->
 <FFSignatureFile DateCreated="2023-08-18T15:27:49" Version="114" xmlns="http://www.nationalarchives.gov.uk/pronom/SignatureFile">
     <InternalSignatureCollection>
         <InternalSignature ID="9" Specificity="Specific">
@@ -52744,7 +52745,7 @@
             PUID="fmt/9" Version="5"/>
         <FileFormat ID="612" Name="Tagged Image File Format"
             PUID="fmt/10" Version="6"/>
-        <FileFormat ID="613" MIMEType="application/pdf"
+        <FileFormat  FormatType="Page Description" ID="613" MIMEType="application/pdf"
             Name="Acrobat PDF 1.0 - Portable Document Format"
             PUID="fmt/14" Version="1.0">
             <InternalSignatureID>123</InternalSignatureID>
@@ -52792,7 +52793,7 @@
             <HasPriorityOverFileFormatID>687</HasPriorityOverFileFormatID>
             <HasPriorityOverFileFormatID>869</HasPriorityOverFileFormatID>
         </FileFormat>
-        <FileFormat ID="619" MIMEType="image/gif"
+        <FileFormat FormatType="Image (Raster)" ID="619" MIMEType="image/gif"
             Name="Graphics Interchange Format" PUID="fmt/3" Version="87a">
             <InternalSignatureID>18</InternalSignatureID>
             <Extension>gif</Extension>
@@ -53005,7 +53006,7 @@
             <Extension>wav</Extension>
             <HasPriorityOverFileFormatID>2741</HasPriorityOverFileFormatID>
         </FileFormat>
-        <FileFormat ID="655" MIMEType="video/x-msvideo"
+        <FileFormat FormatType="Audio, Video" ID="655" MIMEType="video/x-msvideo"
             Name="Audio/Video Interleaved Format" PUID="fmt/5">
             <InternalSignatureID>51</InternalSignatureID>
             <Extension>avi</Extension>
@@ -53060,7 +53061,7 @@
             <InternalSignatureID>54</InternalSignatureID>
             <Extension>wrl</Extension>
         </FileFormat>
-        <FileFormat ID="664" MIMEType="image/png"
+        <FileFormat FormatType="Image (Raster)" ID="664" MIMEType="image/png"
             Name="Portable Network Graphics" PUID="fmt/11" Version="1.0">
             <InternalSignatureID>58</InternalSignatureID>
             <Extension>png</Extension>
@@ -53523,7 +53524,7 @@
             <InternalSignatureID>127</InternalSignatureID>
             <Extension>doc</Extension>
         </FileFormat>
-        <FileFormat ID="735" MIMEType="audio/x-wav"
+        <FileFormat FormatType="Audio" ID="735" MIMEType="audio/x-wav"
             Name="Broadcast WAVE" PUID="fmt/1" Version="0 Generic">
             <InternalSignatureID>1032</InternalSignatureID>
             <Extension>wav</Extension>
@@ -58786,7 +58787,7 @@
             <InternalSignatureID>1364</InternalSignatureID>
             <Extension>exr</Extension>
         </FileFormat>
-        <FileFormat ID="1807" Name="Nearly Raw Raster Data"
+        <FileFormat FormatType="Image (Raster), Dataset" ID="1807" Name="Nearly Raw Raster Data"
             PUID="fmt/1002" Version="1">
             <InternalSignatureID>1357</InternalSignatureID>
             <Extension>nrrd</Extension>

diff --git a/cmd/sf/pronom_test.go b/cmd/sf/pronom_test.go
@@ -29,7 +29,7 @@ type pronomIdentificationTests struct {
 
 var skeletons = make(map[string]*fstest.MapFile)
 
-var minimalPronom = []string{"fmt/1", "fmt/3", "fmt/5", "fmt/11", "fmt/14"}
+var minimalPronom = []string{"fmt/1", "fmt/3", "fmt/5", "fmt/11", "fmt/14", "fmt/1002"}
 
 // Populate the global skeletons map from string-based byte-sequences to
 // save having to store skeletons on disk and read from them.
@@ -59,6 +59,7 @@ func makeSkeletons() {
 		"")
 	files["fmt-3-signature-id-18.gif"] = "4749463837613b"
 	files["badf00d.unknown"] = "badf00d"
+	files["fmt-1002-signature-id-1357.nrrd"] = "4e52524430302e3031"
 	for key, val := range files {
 		data, _ := hex.DecodeString(val)
 		skeletons[key] = &fstest.MapFile{Data: []byte(data)}
@@ -125,6 +126,15 @@ var pronomIDs = []pronomIdentificationTests{
 		"Audio, Video",
 		"extension match avi; byte match at 0, 12",
 		"",
+	}, {
+		"pronom",
+		"fmt/1002",
+		"Nearly Raw Raster Data",
+		"1",
+		"",
+		"Image (Raster), Dataset",
+		"extension match nrrd; byte match at 0, 9",
+		"",
 	},
 }
 

diff --git a/pkg/config/identifier.go b/pkg/config/identifier.go
@@ -324,6 +324,7 @@ func Clear() func() private {
 		identifier.multi = Conclusive
 		loc.fdd = ""
 		mimeinfo.mi = ""
+		pronom.reports = "pronom"
 		return private{}
 	}
 }

diff --git a/pkg/pronom/internal/mappings/droid.go b/pkg/pronom/internal/mappings/droid.go
@@ -12,40 +12,30 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-// Package mappings contains struct mappings to unmarshall three
+// Package mappings contains struct mappings to unmarshal three
 // different PRONOM XML formats: the DROID signature file format, the report
 // format, and the container format.
 package mappings
 
 import "encoding/xml"
 
-// Droid Signature File
-
-// Droid describes the basic top-level structure of a DROID signature
-// file.
 type Droid struct {
 	XMLName     xml.Name            `xml:"FFSignatureFile"`
 	Version     int                 `xml:",attr"`
 	Signatures  []InternalSignature `xml:"InternalSignatureCollection>InternalSignature"`
 	FileFormats []FileFormat        `xml:"FileFormatCollection>FileFormat"`
 }
 
-// InternalSignature describes the structure of the InternalSignature
-// section of a DROID signature file.
 type InternalSignature struct {
 	ID            int       `xml:"ID,attr"`
 	ByteSequences []ByteSeq `xml:"ByteSequence"`
 }
 
-// ByteSeq describes the structure of the ByteSequence sections of a
-// DROID signature file.
 type ByteSeq struct {
 	Reference    string        `xml:"Reference,attr"`
 	SubSequences []SubSequence `xml:"SubSequence"`
 }
 
-// SubSequence describes the structure of the SubSequence sections of a
-// DROID signature file.
 type SubSequence struct {
 	Position        int    `xml:",attr"`
 	SubSeqMinOffset string `xml:",attr"` // and empty int values are unmarshalled to 0
@@ -55,25 +45,21 @@ type SubSequence struct {
 	RightFragments  []Fragment `xml:"RightFragment"`
 }
 
-// Fragment describes the structure of the fragment sections of a DROID
-// signature file.
 type Fragment struct {
 	Value     string `xml:",chardata"`
 	MinOffset string `xml:",attr"`
 	MaxOffset string `xml:",attr"`
 	Position  int    `xml:",attr"`
 }
 
-// FileFormat describes the structure of the FileFormat section of a
-// DROID signature file.
 type FileFormat struct {
 	XMLName    xml.Name `xml:"FileFormat"`
 	ID         int      `xml:"ID,attr"`
 	Puid       string   `xml:"PUID,attr"`
 	Name       string   `xml:",attr"`
 	Version    string   `xml:",attr"`
 	MIMEType   string   `xml:",attr"`
-	Types      string   `xml:"FormatTypes"`
+	FormatType string   `xml:",attr"`
 	Extensions []string `xml:"Extension"`
 	Signatures []int    `xml:"InternalSignatureID"`
 	Priorities []int    `xml:"HasPriorityOverFileFormatID"`

diff --git a/pkg/pronom/parseable.go b/pkg/pronom/parseable.go
@@ -233,7 +233,7 @@ func (d *droid) Infos() map[string]identifier.FormatInfo {
 			name:     strings.TrimSpace(v.Name),
 			version:  strings.TrimSpace(v.Version),
 			mimeType: strings.TrimSpace(v.MIMEType),
-			class:    strings.TrimSpace(v.Types),
+			class:    strings.TrimSpace(v.FormatType),
 		}
 	}
 	return infos