Closed mjordan closed 6 years ago
We should break our the ingest_object()
function into a class per content model.
We now have an ingester for books. However, derivatives for book pages are not being generated. Investigating.
We now have a compound object ingester. Derivatives also now work for book pages and compound child objects.
Input for books:
bookstestinput/
├── 1
│ ├── 1
│ │ └── OBJ.tiff
│ ├── 2
│ │ └── OBJ.tiff
│ ├── 3
│ │ └── OBJ.tiff
│ ├── 4
│ │ └── OBJ.tiff
│ └── MODS.xml
└── 2
├── 1
│ └── OBJ.tiff
├── 2
│ └── OBJ.tiff
├── 3
│ └── OBJ.tiff
├── 4
│ └── OBJ.tiff
└── MODS.xml
For compounds:
cpdtestinput/
├── 1
│ ├── 1
│ │ ├── MODS.xml
│ │ └── OBJ.jpg
│ ├── 2
│ │ ├── MODS.xml
│ │ └── OBJ.jpg
│ └── MODS.xml
└── 2
├── 1
│ ├── MODS.xml
│ └── OBJ.tif
├── 2
│ ├── MODS.xml
│ └── OBJ.tif
└── MODS.xml
Closed with 6887205eb87fa2c952912c24b7015ea7efe1f4f2.
For each of these complex objects, the content model passed in
--cmodel
can determine how to assemble the object in each input subdirectory. Input data would include the parts of each child object, like this:Objects would need to created in a top-down traversal, since to ingest each child we need the PID of its parent.