What to do about identification via the Siegfried module that both has an error and identification results

[ross-spencer]

Given the skeleton file example below by way of illustration:

sf fmt-128-signature-id-139.sxw 
[FILE] /home/ross-spencer/git/ocfl/corpus/false-positives/fmt-128-signature-id-139.sxw
[ERROR] zip: not a valid zip file
---
siegfried   : 1.11.1
scandate    : 2024-11-28T15:16:12+01:00
signature   : default.sig
created     : 2024-06-21T16:01:11+10:00
identifiers : 
  - name    : 'pronom'
    details : 'DROID_SignatureFile_V118.xml; container-signature-20240501.xml'
---
filename : 'fmt-128-signature-id-139.sxw'
filesize : 488
modified : 2024-11-28T14:30:01+01:00
errors   : 'zip: not a valid zip file'
matches  :
  - ns      : 'pronom'
    id      : 'fmt/128'
    format  : 'OpenOffice Writer'
    version : '1.0'
    mime    : 'application/vnd.sun.xml.writer'
    class   : 'Word Processor'
    basis   : 'extension match sxw; byte match at [[0 68] [88 400]]'
    warning : 

When identified via the Siegfried package, it will simultaneously return an error, but also an identification, mirroring the command line ID. However, when using Siegfried as a library the error returned usually takes on greater importance because it can often mean something more serious with the application.

Example code:

// Copy and pasta code demonstrating SF identification
// false-negative.
package main

import (
	"bufio"
	"bytes"
	"io"
	"log"
	"os"

	"github.com/richardlehane/siegfried"
	"github.com/richardlehane/siegfried/pkg/pronom"
)

const sigfile = "/home/ross-spencer/.local/share/siegfried/default.sig"
const skeleton = "fmt-128-signature-id-139.sxw"

func main() {
	sig, _ := os.Open(sigfile)
	stat, err := sig.Stat()
	if err != nil {
		log.Println(err)
		return
	}
	bs := make([]byte, stat.Size())
	_, err = bufio.NewReader(sig).Read(bs)
	if err != nil && err != io.EOF {
		log.Println(err)
		return
	}
	sf, err := siegfried.LoadReader(bytes.NewBuffer(bs))
	//var sf *siegfried.Siegfried
	file, err := os.Open(skeleton)
	if err != nil {
		log.Printf("error opening file: %s", err)
		return
	}
	ident, err := sf.Identify(bufio.NewReader(file), "testfile", "")
	if err != nil {
		log.Printf("error isn't nil: %s %+v", err, ident)
	}
	for _, id := range ident {
		if pid, ok := id.(pronom.Identification); ok {
			log.Printf("but we still get a positive format identification... '%s'", skeleton)
			log.Println("puid", pid.ID)
			log.Println("name", pid.Name)
			log.Println("version", pid.Version)
			log.Println("class", pid.Class)
			log.Println("warn", pid.Warning)
			log.Println("basis", pid.Basis)

		}
	}
}

Packaged with file and go.mod/go.sum: example_sf_reader_main.zip

Example out:

2024/11/28 15:09:10 error isn't nil: zip: not a valid zip file [fmt/128]
2024/11/28 15:09:10 but we still get a positive format identification... 'fmt-128-signature-id-139.sxw'
2024/11/28 15:09:10 puid fmt/128
2024/11/28 15:09:10 name OpenOffice Writer
2024/11/28 15:09:10 version 1.0
2024/11/28 15:09:10 class Word Processor
2024/11/28 15:09:10 warn extension mismatch
2024/11/28 15:09:10 basis [byte match at [[0 68] [88 400]]]

I'm torn about how to treat it. There's useful info here, so maybe keep the useful info? Or throw it away given the error?

Or, should sf's handling be different here? there's an intellectual "error" but is it an actual "software" error?

[richardlehane]

hey @ross-spencer this is interesting, perhaps this class of "errors" should go to the warning field instead?

Errors should really be about reporting issues that prevent identification happening (e.g. permissions on the file, network failure) whereas warnings are there to suggest potential issues with files (empty, extension mismatch, etc.). This feels more like it belongs in that second category.

[ross-spencer]

Yeah, I agree with this at a library level. It would probably be an API change I guess and so require bumping a major version if the change were made?

In case it has any utility, these are the files (33) (probably non-exhaustive outside of the skeleton suite) that I see with different combinations of error and ID or none-ID:

{
  "file": "fmt-111-signature-id-170.nul",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}
{
  "file": "fmt-125-signature-id-171.ppt",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}
{
  "file": "fmt-126-signature-id-172.ppt",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}
{
  "file": "fmt-127-signature-id-142.sxd",
  "error": "zip: not a valid zip file",
  "id": "fmt/127",
  "basis": "fmt/127"
}
{
  "file": "fmt-128-signature-id-139.sxw",
  "error": "zip: not a valid zip file",
  "id": "fmt/128",
  "basis": "fmt/128"
}
{
  "file": "fmt-129-signature-id-141.sxc",
  "error": "zip: not a valid zip file",
  "id": "fmt/129",
  "basis": "fmt/129"
}
{
  "file": "fmt-130-signature-id-140.sxi",
  "error": "zip: not a valid zip file",
  "id": "fmt/130",
  "basis": "fmt/130"
}
{
  "file": "fmt-135-signature-id-202.nul",
  "error": "zip: not a valid zip file",
  "id": "fmt/135",
  "basis": "fmt/135"
}
{
  "file": "fmt-136-signature-id-203.odt",
  "error": "zip: not a valid zip file",
  "id": "fmt/135",
  "basis": "fmt/135"
}
{
  "file": "fmt-137-signature-id-204.ods",
  "error": "zip: not a valid zip file",
  "id": "fmt/135",
  "basis": "fmt/135"
}
{
  "file": "fmt-138-signature-id-205.odp",
  "error": "zip: not a valid zip file",
  "id": "fmt/135",
  "basis": "fmt/135"
}
{
  "file": "fmt-139-signature-id-206.odg",
  "error": "zip: not a valid zip file",
  "id": "fmt/135",
  "basis": "fmt/135"
}
{
  "file": "fmt-140-signature-id-207.odb",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-1568-signature-id-1910.isdocx",
  "error": "zip: not a valid zip file",
  "id": "fmt/1568",
  "basis": "fmt/1568"
}
{
  "file": "fmt-189-signature-id-258.nul",
  "error": "zip: not a valid zip file",
  "id": "fmt/189",
  "basis": "fmt/189"
}
{
  "file": "fmt-290-signature-id-306.odt",
  "error": "zip: not a valid zip file",
  "id": "x-fmt/3",
  "basis": "x-fmt/3"
}
{
  "file": "fmt-290-signature-id-307.odt",
  "error": "zip: not a valid zip file",
  "id": "x-fmt/3",
  "basis": "x-fmt/3"
}
{
  "file": "fmt-291-signature-id-308.odt",
  "error": "zip: not a valid zip file",
  "id": "x-fmt/3",
  "basis": "x-fmt/3"
}
{
  "file": "fmt-292-signature-id-309.odp",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-292-signature-id-310.odp",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-293-signature-id-311.odp",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-294-signature-id-312.ods",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-294-signature-id-313.ods",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-295-signature-id-314.ods",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-296-signature-id-315.odg",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-296-signature-id-316.odg",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-297-signature-id-317.odg",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-39-signature-id-128.doc",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}
{
  "file": "fmt-40-signature-id-182.doc",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}
{
  "file": "fmt-482-signature-id-906.ibooks",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "fmt-483-signature-id-905.epub",
  "error": "zip: not a valid zip file",
  "id": "UNKNOWN",
  "basis": "UNKNOWN"
}
{
  "file": "x-fmt-430-signature-id-208.msg",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}
{
  "file": "x-fmt-88-signature-id-181.ppt",
  "error": "mscfb: slicer read error (EOF); 0",
  "id": "fmt/111",
  "basis": "fmt/111"
}