implement track count and generate appropriate number of audio tracks in the revtmd technical section

^^

Right now, multchoicebox is auto selecting the first item. You have to click CLEAR ALL. This must be fixed as it's annoying and will lead to bad data.

https://wiki.archivematica.org/PREMIS_metadata:_events

Possibly add metadata from mediainfo like the FITS example here: https://gist.github.com/mjordan/8672407

As these have children, the final xmlstarlet call doesn't affect them. it only deletes the children, not the parent. Figure out how to add as many as needed. Could also be handy for a later QC/fixity check/migration phase.

Start counter at x (equal to number of definite fields) - add counter after each option to get the correct number of coding process histories

Investigate other repositories and see how they do it, then do that.

Argparse will automatically improve documentation.

If ffmpeg decides the names, a first process log can be overwritten if it completes in the same second as the first second.
Use the report environment variable, which should be easy on unix, not s much on windows :{

check if similar situation on unix.

Right now, I'm using indexes[1] so perhaps something involving XPATH like '//16mm Telecine' or '//Quality Assesment' would be better..

Part of cleaning/maintanence protocol should be recording the date of deck cleaning somewhere in the revtmd coding process history. Not sure where, settings sounds wonky, but not sure where else it could go.

With the possibility of some source drives not allowing manifests to be written to them it is suggested that the destination for these files should be somewhere on the desktop.

For bad HDCAM->Digi dubs that maintained the pillarboxes: make a script that

1. removes the pillarboxes from the input
2. exports as PAL

Perhaps ask user which codec they want?

^^

Add multiple possibilities for FPS.
Exposure pushed to the point of noise
condition info - how to represent?
Specify how much audio level was adjusted by.

Perhaps do a verify first, rather than create a new manifest?

As these fixity checks can take a long time, it should email the csv once it has completed.
This also means that we can open the reports on our office computers.

Right now it's baffling and meaningless when checksums fail.

tried dcpfixity on ingest1 (windows). error states: 'System cannot find file system' . seems there is a need to install 'openssl' for windows for the script to be working.

Codec is currently broken - requires codec.type children rather than just the codec name.

^^ it's be a shame to do a very long fixity check and then you can't copy the file file because of write access.

md5deep could still work, but I think a counter that shows progress of maneifest generation would be very helpful, especially when running a 24 hour move with 150k files.

move is a system command on windows, and with the changes to IFIscripts that allow us to update the scripts easier, this is suddenly an issue.

mover.py would work..

^^

check if cp/gcp/robocopy have log modes - otherwise, find a tradeoff between capturing stdout and having terminal updates for the user.

It sucks.
Adding an option for output could be interesting though.

We should be able to just use sys.argv[0] to copy the script wherever. Thanks to @vphill for the idea.

^^

DCDMs are often just in audio&video folder without subfolder. eg E:/video/tiff tiff tiff and E:/ audio/ wav wav. To select E:/ as source would be great.

For DCDM's to save time can we have an option to option to create ~without destination manifest ~ just detination manifest?

Figure out how to get regular sha1s out of the base64/binary hashes and store as traditional sha1sum manifest.

This was a terrible idea. Remove as soon as possible. Find a better way. Intermediate workaround is to just take in ffmpeg logs but even that is pretty bad.

What to do?

1. Possibly make new fields based around revtmd/videoMD.
2. Create temporary xml file in submission package, this is harvested and added to the final inmagic xml during the accessions phase?
3. Easygui might be the only user friendly way to enter user info?

^^

Make clear that the script requires a PKL, but change so that it accepts folder input.

You should probably add that.

Requires uuid extraction from fonts. Is this possible? Ask someone who knows.

Not sure what the issue is exactly, perhaps it's something to do with my flags, perhaps the recursive flag?

Seeing as we already have the filenames and the md5s, we should generate a manifest using the md5sum standard, something like:
path/to/file md5checksum43890jksdjikloj49s

final message = ' files transferred ok and checksums match' as it's just checking the manifests.

https://wiki.archivematica.org/PREMIS_metadata:_events

Implement this in logs and database cos it's probably a good idea to do so.

Figure out if these are created as a result of the script, in particular the actual .tmp file creation. I don't think that this is the issue as these are created in the parent directory, while the .fuses are in the main file dir.

ltopers just deletes all hidden files, so maybe we could delete all files that .startswith(.fuse) . Hopefully if we delete them before the manifest check then all will be well. Further tests required.