marklogic-python-batch-insert.py

batch = DocumentBatch()
# Document with every type of metadata supplied
batch.add(
    Document(
        uri="/example1.json",
        content={"some": "data"},
        content_type="application/json",
        metadata=Metadata(
            quality=1.2,
            collections=["coll1", "coll2"],
            permissions={
                "rest-reader": ["read", "execute"],
                "rest-writer": "update",
            },
            properties={
                "prop1": "simple value",
                "prop2": "<nested><xml>will work</xml></nested>",
            },
            metadataValues={"key1": "value1", "key2": "value2"},
        ),
    )
)
# Minimal example of creating a document
batch.add(
    Document(
        uri="/example2.xml",
        content="<some>xml</some>",
        content_type="application/xml",
    ),
)
data, headers = batch.data_and_headers()

response = requests.post(
    "http://localhost:8000/v1/documents",
    auth=HTTPDigestAuth("admin", "admin"),
    data=data,
    headers=headers,
)

Implementing this can be done by copying the excellent work at marklogic/python_api - specifically in https://github.com/marklogic/python_api/blob/master/marklogic/client/bulkloader.py and https://github.com/marklogic/python_api/blob/master/marklogic/client/documents.py

Note that that project has a much larger scope, covering parts of the Manage API and Client API. Additionally, the Documents class serves two purposes - one as a facade on /v1/documents, and also for holding bits about a single document.