Deposit left around when ingesting something a second time

[marcolarosa]

If you run the test:

  it.only("Does not increment version number if you add the same thing twice", async function() {
    await repository.createNewObjectContent("xx", makeContent);
    const object = await repository.createNewObjectContent("xx", makeContent);
    const inventory = await object.getInventory();
    assert.strictEqual(inventory.head, "v1");
  });

and then look at the content of the ocfl repo you'll see there are no errors but the deposit folder has been left around:

If you change the test to (createNewObjectContent 3 times):

  it.only("Does not increment version number if you add the same thing twice", async function() {
    await repository.createNewObjectContent("xx", makeContent);
    await repository.createNewObjectContent("xx", makeContent);
    const object = await repository.createNewObjectContent("xx", makeContent);
    const inventory = await object.getInventory();
    assert.strictEqual(inventory.head, "v1");
  });

You get an error:

[marcolarosa]

Hey,

I'm bumping in to this issue whilst trying to write out the pdsc catalog. Is there a way to work around it? Specifically, given an id, is there a way I can first see if the object has already been written?

Something like:

if (!ocflObject.exists(id)) {
  // write it out
fi

This is a show stopper for me ATM and I don't want to implement a check if object exists in my code as I think it makes sense in this lib but can't see one in there.

[spikelynch]

I don't think there's a method to check for existence yet and yeah, it makes sense to put it in the library, will have a look.

[marcolarosa]

FYI - looks like there's bits of this.

https://github.com/UTS-eResearch/ocfl-js/blob/master/lib/repository.js#L197 uses pairtree.path to convert an id to a filesystem path (this.objectIdToPath = pairtree.path in constructor)

Then that path is used in ocflObject.js isObject method to check if path exists https://github.com/UTS-eResearch/ocfl-js/blob/master/lib/repository.js#L212.

So the easy fix is as follows - change isObject to add an id to path conversion if an id is passed in:

  async isObject({id, aPath }) {
    if (id) aPath = pairtree.path(id)
    // TODO: Check if this content root with NAMASTE and returns ocfl version
    // 0=ocfl_object_1.0
    // looks at path and see if the content of the file is
    // TODO: Make this look for a namaste file beginning with 0=ocfl_object_ and extract the version
    const theFile = path.join(aPath, "0=" + this.nameVersion(this.ocflVersion));
    return await fs.pathExists(theFile);
  }

Note the signature change - looks like isObject is used in a few places

[marcolarosa]

Ok - see pull request #31

[spikelynch]

Ok have merged it - does that get around your current problem? Or do you still need an existence test in the library? (My coding brain is a bit slower than usual after my break)

[marcolarosa]

I'm not sure yet. I'm trying to get my code to use my fork and struggling... Will report back when I figure out what's happening but for now; don't worry about existence test. In any case, I can add one and send off a pull request if I do.