Output from tests/test_demo_using_bagit_bags.py.
Imagine that we have a Bagit bag tests/testdata/bags/uaa_v1 that represents the initial state of an object. We can use --create to make a new OCFL object /tmp/obj with that content as the v1 state:
> python ocfl-object.py create --objdir tmp/obj --srcbag tests/testdata/bags/uaa_v1 -v
INFO:root:Created OCFL object info:bb123cd4567 in tmp/obj
Now that we have the object it is of course valid.
> python ocfl-validate.py tmp/obj
OCFL v1.1 Object at tmp/obj is VALID
Looking inside the object we see v1 with the expected 2 content files.
> python ocfl-object.py show --objdir tmp/obj
INFO:root:OCFL v1.1 Object at tmp/obj has VALID STRUCTURE (DIGESTS NOT CHECKED)
Object tree for None
[tmp/obj]
├── 0=ocfl_object_1.1
├── inventory.json
├── inventory.json.sha512
└── v1
├── content (2 files)
├── inventory.json
└── inventory.json.sha512
If we have a bag tests/testdata/bags/uaa_v2 with updated content we can --update the object to create v2.
> python ocfl-object.py update --objdir tmp/obj --srcbag tests/testdata/bags/uaa_v2 -v
INFO:root:Updated OCFL object info:bb123cd4567 by adding v2
Updated object info:bb123cd4567 to v2
Looking inside the object we now see v1 and v2. There are no content files inside v2 because although this update added two files they have identical content (and hence digest) as one of the files in v1
> python ocfl-object.py show --objdir tmp/obj
INFO:root:OCFL v1.1 Object at tmp/obj has VALID STRUCTURE (DIGESTS NOT CHECKED)
Object tree for None
[tmp/obj]
├── 0=ocfl_object_1.1
├── inventory.json
├── inventory.json.sha512
├── v1
│ ├── content (2 files)
│ ├── inventory.json
│ └── inventory.json.sha512
└── v2
├── inventory.json
└── inventory.json.sha512
Similarly we can --update with tests/testdata/bags/uaa_v3 to create v3.
> python ocfl-object.py update --objdir tmp/obj --srcbag tests/testdata/bags/uaa_v3 -v
INFO:root:Updated OCFL object info:bb123cd4567 by adding v3
Updated object info:bb123cd4567 to v3
Looking inside again we see that v3 does add another content file.
> python ocfl-object.py show --objdir tmp/obj
INFO:root:OCFL v1.1 Object at tmp/obj has VALID STRUCTURE (DIGESTS NOT CHECKED)
Object tree for None
[tmp/obj]
├── 0=ocfl_object_1.1
├── inventory.json
├── inventory.json.sha512
├── v1
│ ├── content (2 files)
│ ├── inventory.json
│ └── inventory.json.sha512
├── v2
│ ├── inventory.json
│ └── inventory.json.sha512
└── v3
├── content (1 files)
├── inventory.json
└── inventory.json.sha512
Finally, we can --update again with tests/testdata/bags/uaa_v4 to create v4.
> python ocfl-object.py update --objdir tmp/obj --srcbag tests/testdata/bags/uaa_v4 -v
INFO:root:Updated OCFL object info:bb123cd4567 by adding v4
Updated object info:bb123cd4567 to v4
Taking the newly created OCFL object /tmp/obj we can --extract the v4 content as a Bagit bag. The --set-bagging-date means that the created time for v4 will be used to generate the Bagging-Date in the Bagit bag’s metadata.
> python ocfl-object.py extract --objver v4 --objdir tmp/obj --dstbag tmp/extracted_v4 --set-bagging-date -v
INFO:root:Extracted v4 into tmp/extracted_v4
Extracted content for v4 saved as Bagit bag in tmp/extracted_v4
We note that the OCFL object had only one content file in v4 but the extracted object state for v4 includes 4 files, two of which have identical content (dracula.txt and another_directory/a_third_copy_of_dracula.txt). We can now compare the extracted bag /tmp/uaa_v4 that with the bag we used to create v4 tests/testdata/bags/uaa_v4 using a recursive diff.
> diff --ignore-matching-lines bag-info.txt -r tmp/extracted_v4 tests/testdata/bags/uaa_v4
diff --ignore-matching-lines ' bag-info.txt' -r tmp/extracted_v4/bag-info.txt tests/testdata/bags/uaa_v4/bag-info.txt
1d0
< Bag-Software-Agent: bagit.py v1.8.1 <https://github.com/LibraryOfCongress/bagit-python>
diff --ignore-matching-lines ' bag-info.txt' -r tmp/extracted_v4/tagmanifest-sha512.txt tests/testdata/bags/uaa_v4/tagmanifest-sha512.txt
1,2d0
< 5c2e2b9cacc93cb315d57f09fac6d199c3378313b6cf918bb0a70e1839c4e4c0c2e5a7f9ae869cf7755e09a196a835be1af7c510d3d5faa5d0c0b3f6be9f816a manifest-sha512.txt
< 6f2dbad07611c6c313392e6ae0676233b5a50379021ee93fb667c301fca5924a65792e23275cb4ee5f91265a570d2fbe4eba734397534717ffb7c555b6c1ab19 bag-info.txt
3a2,3
> 10624e6d45462def7af66d1a0d977606c7b073b01809c1d42258cfab5c34a275480943cbe78044416aee1f23822cc3762f92247b8f39b5c6ddc5ae32a8f94ce5 bag-info.txt
> 5c2e2b9cacc93cb315d57f09fac6d199c3378313b6cf918bb0a70e1839c4e4c0c2e5a7f9ae869cf7755e09a196a835be1af7c510d3d5faa5d0c0b3f6be9f816a manifest-sha512.txt
(last command exited with return code 1)
The only difference that shows in the output of the diff is the addition of the Bag-Software-Agent: line in the newly created bag that was no present in the input metadata. Because of this additional metadata, the digest for the bag is also different, but this is removed from the diff output with the --ignore-matching-lines ' bag-info.txt' parameter. The bag content matches.