Updated PRONOM signature files - v77

[mistydemeo]

I noticed the signature files have fallen a few releases behind; there's the V74 signature files.

If this is accepted, can you also tag a new release? Thanks, and thanks for all your work on FIDO. :D I'll submit updates for the upcoming signature files too.

[Kris-LIBIS]

I guess, you should be able to update the signature file yourself, not? The included script update_signatures.py should do the job.

However, when I tried it, I got an error message:

FIDO signature updater v1.2.2
Contacting PRONOM...
Querying latest signaturefile version...
Downloading signature file version 74...
Writing DROID_SignatureFile-v74.xml...
Extracting PRONOM PUID's from signature file...
Traceback (most recent call last):
  File "update_signatures.py", line 171, in <module>
    main(defaults)
  File "update_signatures.py", line 69, in main
    for node in tree.iter("{http://www.nationalarchives.gov.uk/pronom/SignatureFile}FileFormat"):
AttributeError: ElementTree instance has no attribute 'iter'

[mistydemeo]

Yep, these new signature files were generated from that script. I'm submitting the new ones here so a version of fido with the latest PRONOM signatures can be packaged.

[Kris-LIBIS]

I'm sorry, I overlooked the pull request.

Any idea why the update script isn't working in my case?

[mistydemeo]

Are you using an older Python release? The ElementTree.iter method was introduced in Python 2.7, so it wouldn't be present if you're using 2.5 or 2.6.

[Kris-LIBIS]

Thanks. That's it. Sorry, I'm a Python noob.

[mistydemeo]

No worries! Yeah, I think FIDO only works with 2.7, and 3.2+.

[mistydemeo]

Ping - any chance of getting a new release with the latest PRONOM signatures? We're prepping a new Archivematica release, and it would be great to have an up-to-date FIDO/PRONOM to include.

[mistydemeo]

Updating this branch with the new v75 signature files. A few changes are also required to make downloading the signature files work with the reorganization of the Nat Archives website.

[mistydemeo]

Updated to the v75 signature files.

Also includes a few updates and bugfixes for the updater, required to be able to do the new import.

I'm not 100% sure that 7852c1f is correct/the right solution, but it does appear to work. I don't know that the newline is meant to be significant given what appears to be the structure of byte signatures.

[anjackson]

Weird. In the v75 sig file looks like this:

...
                    <LeftFragment MaxOffset="0" MinOffset="0" Position="3">687474703A2F2F7777772E696E7465726C69732E63682F494E5445524C4953322E33</LeftFragment>
                    <LeftFragment MaxOffset="0" MinOffset="0" Position="4">22</LeftFragment>
                    <LeftFragment MaxOffset="0" MinOffset="0" Position="4">27</LeftFragment>
                    <LeftFragment MaxOffset="0" MinOffset="0" Position="5">
3C5452414E5346455220786D6C6E733D</LeftFragment>
...

i.e. the content of that last element starts with a newline.

From context, I think it's probably safe to assume that this whitespace can be ignored, but I'll alert the PRONOM folks anyway. (done)

[mistydemeo]

V77 should fix the stray newlines. Removing the two commits that deal with newlines, and testing an import of V77. If that works I'll update the branch.

[mistydemeo]

Should I keep the commits with the v74 and v75 signature files, or include only v77?

[mistydemeo]

Thanks!