Work with PROV-XML

PROV-XML needs the xml extra, which installs lxml:

python -m pip install "prov[xml]"

Serialize to a file

import prov.model as pm

document = pm.ProvDocument()
document.set_default_namespace("http://example.org/")
document.entity("e1")

document.serialize("document.xml", format="xml")

Serialize to a string

xml_str = document.serialize(format="xml")
print(xml_str)

Write xsi:type on every attribute

By default xsi:type is written only for prov:type, prov:location and prov:value, matching the examples in the PROV-XML specification. Pass force_types=True to write it for every non-prov: attribute too:

xml_str_typed = document.serialize(format="xml", force_types=True)

Deserialize from a file or stream

loaded = pm.ProvDocument.deserialize("document.xml", format="xml")
assert loaded == document

Deserialize from a string

loaded = pm.ProvDocument.deserialize(content=xml_str, format="xml")

Auto-detect the format with prov.read()

prov.read() tries PROV-XML fourth, after PROV-JSON, PROV-O and PROV-N. Valid PROV-XML auto-detects from a file path, raw content or a seekable stream. See Auto-detect the format with prov.read() in the PROV-JSON guide for how detection works and when to pass format= instead.

import prov

loaded = prov.read("document.xml")
assert loaded == document

Common errors

Malformed XML raises lxml’s own syntax error:

try:
    pm.ProvDocument.deserialize(content="not xml", format="xml")
except Exception as e:
    print(f"{type(e).__name__}: {e}")

XMLSyntaxError: Start tag expected, '<' not found, line 1, column 1 (<string>, line 1)

The parser does not resolve external DTD entities and never touches the network. A document that relies on entity expansion reads back with those values empty.