sverify

Prove a data pipeline's contracts before it runs, get the same result on DuckDB, Polars, DataFusion and Spark, and keep it up to date from Delta Lake and Iceberg change feeds.

You write a transformation in Ibis or SQL and say what it assumes about its inputs (requires) and what it guarantees about its output (ensures). A checker written in Lean proves the guarantees for every possible input, or says why it can't. Then:

• Contracts are proved, not tested. not_null, unique, ranges and relations between columns hold on every input that meets the preconditions, and division by zero, overflow and bad casts are ruled out before the job runs. Preconditions are checked on the data, or proved from Delta and Iceberg metadata.
• One plan gives the same answer on every engine. The checked plan runs on DuckDB, Polars or DataFusion, or is written as Spark SQL, with NULLs, integer division, NaN, ties and overflow spelled out, or it is refused where an engine can't match.
• Results are kept up to date from the changes, by delta plans proved equal to a full recompute, read from Delta's Change Data Feed or Iceberg's appends, and written to sinks exactly once, with the contract attached.

sverify is a young, independent project, and for now it is purely observational: it tells you what it can and can't prove, and describes how each engine behaves as observed at pinned versions, without certifying or ranking them. It isn't affiliated with any engine's project or vendor. Use it as a careful second pair of eyes rather than a gate production depends on, and expect the API to keep changing between minor releases. What it's for is making data engineering more joyful: fewer 3 a.m. pages, fewer surprises between engines, more trust in the tables you hand on.

Snowflake support is mostly aspirational for now. sverify writes Snowflake SQL, but has only checked it on an emulator, fakesnow, not on a real account (#63).

pip install sverify                # or: uv add sverify
pip install 'sverify[all]'         # with every engine and table format

The core runs on DuckDB and writes DuckDB, Snowflake and Spark SQL. Extras: datafusion (also compiles SQL strings), polars, spark (needs Java 17 or 21), delta, iceberg, or all. Wheels bundle the checker as a small native binary, for Linux (x86_64, aarch64) and macOS (arm64, x86_64), Python 3.12 and later; elsewhere pip builds from the sdist, which needs elan.

The API reference documents every public function and class, for the released version. For coding agents, the site also publishes skills for writing contracts, pipelines and engine code with sverify (npx skills add https://alexandernenninger.github.io/substrait-lean/); they ship in the wheel too, under sverify/skills/.

In thirty lines

import pyarrow as pa
import sverify as sv
from ibis import _
from sverify.io import StaticSource

orders = StaticSource(  # in memory; DeltaSource and IcebergSource work the same way
    pa.table({"id": [1, 2], "customer": ["ada", "bob"], "amount": [30, 5]}), name="orders"
)


@sv.transform(
    requires={"orders": [_.amount >= 0, sv.unique("id")]},
    ensures=[_.doubled >= 22, _.doubled.notnull(), sv.unique("id")],
)
def big_orders(orders):
    return orders.filter(orders.amount > 10).select(orders.id, doubled=orders.amount * 2)


big = orders.pipe(big_orders)  # proved here, before any data runs
big.refresh()  # computed, on DuckDB by default
orders.apply(sv.Changes(inserts=[(3, "cy", 50)], deletes=[(1, "ada", 30)]))
big.refresh(engine="polars")  # updated from the changes alone, on another engine
assert big.to_arrow().to_pylist() == [{"id": 3, "doubled": 100}]
print(big_orders.last_report)
# requires orders: (_.amount >= 0)
# requires orders: unique(id)
# ensures (_.doubled >= 22): proved

A postcondition that doesn't follow is refused before anything runs:

@sv.transform(requires={"orders": [_.amount >= 0]}, ensures=[_.total > 5])
def plus_one(orders):
    return orders.select(orders.id, total=orders.amount + 1)


from sverify.errors import VerificationError

try:
    plus_one(orders)
except VerificationError as e:
    print(e)
# could not prove: (_.total > 5) (it is TRUE or FALSE)

Testing with pytest

sverify.testing has assertions for your test suite, and installing sverify registers a pytest plugin with two fixtures: sv_source makes an in-memory source from rows, and a test that takes sv_engine runs once per installed engine.

from sverify.testing import assert_contract, assert_engines_agree, source


def test_big_orders():
    orders = source([], {"id": "int64", "customer": "string", "amount": "int64"}, name="orders")
    node = orders.pipe(big_orders)
    assert_contract(node)  # the contract is proved
    data = {"orders": [(1, "ada", 30), (2, "bob", 5)]}
    assert_contract(node, data=data, expected=[(1, 60)])  # and holds on these rows
    assert_engines_agree(node, data)  # DuckDB, Polars and DataFusion agree


test_big_orders()

A failure lists the rows: those of the test data that break a precondition, or the expected rows that are missing and the ones that were not expected. sverify.testing.cases(node) is a Hypothesis strategy of inputs that meet the preconditions, for property tests of the code around a node (pip install 'sverify[testing]').

SQL models

A transform's body can be SQL, over its inputs by name. It is proved, composed and refreshed like any other:

@sv.transform(
    requires={"orders": ["amount >= 0", "amount IS NOT NULL"]},
    ensures=["total >= 0", sv.unique("customer")],
)
def per_customer(orders):
    return sv.sql(
        "SELECT customer, coalesce(sum(amount), 0) AS total FROM orders GROUP BY customer",
        orders=orders,
    )


totals = orders.pipe(per_customer)  # proved here, like an Ibis transform
totals.refresh()
assert sorted(totals.to_arrow().to_pylist(), key=str) == [
    {"customer": "bob", "total": 5},
    {"customer": "cy", "total": 50},
]

Several inputs are several keywords (sv.sql(query, orders=orders, customers=customers)). The SQL is planned by DataFusion (pip install 'sverify[datafusion]'), or by Isthmus without it.

Workflows

Three complete jobs from a catalogue of 139 real analytical queries, each from source tables to a sink, with what goes wrong without the checker. Open them in the browser:

• Duplicate card payments: a self-join over a Delta table, refreshed from its Change Data Feed, where a watermark job would miss a late duplicate and keep a reversed one.
• Sensor alerts: LAG over an Iceberg table, where ties and a NaN reading change the alerts, and Snowflake and Spark need proved tie-freedom.
• An orders mart: dbt tests as a proved contract on a point-in-time join to a type-2 dimension, in Ibis and in SQL.

The whole tour explains every feature in a chapter of its own; it runs locally with uvx marimo edit --sandbox notebooks from a clone.

The API in brief

import json

import ibis
from sverify.engines import SNOWFLAKE, to_sql  # also DUCKDB, SPARK
from sverify.io import DeltaSink, DeltaSource  # IcebergSource, IcebergSink likewise

sv.verify(
    "select id, amount from orders where amount > 0",
    tables={"orders": ibis.table({"id": "int64", "amount": "int64"}, name="orders")},
    ensures=["amount > 0"],
)  # SQL, checked; returns the report
to_sql(json.loads(big.plan), SNOWFLAKE)  # the checked plan, or sverify.engines.Unsupported

# p = DeltaSource("s3://.../orders").pipe(paid).write(DeltaSink("s3://.../paid"))  # stored
# r = p.pipe(revenue, regions=regions).write(DeltaSink("s3://.../revenue"))
# r.refresh(engine="polars"); sv.group(r, other).refresh(); r.to_pandas()

• @sv.transform(requires=..., ensures=..., on_unproved="error" | "runtime" | "warn") proves a contract on each call; .last_report says what was proved and what is known about every output column. sv.unique(...) states keys.
• source.pipe(paid).pipe(revenue, regions=regions) builds nodes left to right, in the order the data flows (the call form revenue(paid(source), regions=regions) is the same node). A step is a transform, SQL (sv.sql(query), which binds the chained node to the one table the keywords leave over) or a plain function. A downstream precondition is proved from the upstream contract (CompositionError names a broken link).
• Without .write in between, transforms fuse into one plan; .write(sink) stores a node, exactly once per state, and nodes downstream read it as a versioned source. refresh(engine=...) updates the whole lineage, and sv.group(a, b) several outputs against the same source versions.
• Filters, projections, inner and cross joins and UNION ALL are maintained from the changes; anything else is recomputed and diffed.
• sverify.io has Delta, Iceberg and in-memory sources and sinks, and base classes for your own: a Source implements read(since) (from_query, from_frame and from_files build one), a Sink implements write(change) (sink_from wraps a function).
• sverify.testing has assert_contract, assert_engines_agree and the Hypothesis strategy cases for pytest, with the plugin's sv_engine and sv_source fixtures.
• sverify.engines has the engines, SQL dialects and plans (run_sharded runs a plan per shard and combines the results, with a proof that the result is the same); sverify.errors the exceptions.

Not supported yet: DECIMAL columns (keep money as integer cents), framed window aggregates, RANK and NTILE. The repository's README has the coverage table, the full guide, and what is proved and what is trusted.

Configuration

• SVERIFY_CHECK: path to a substrait-check binary to use instead of the bundled one, for example your own lake build.
• SVERIFY_ISTHMUS: path to an Isthmus binary (otherwise isthmus on PATH), to compile SQL strings without DataFusion; sv.connect(sql_producer="isthmus") picks it, and it is the default when DataFusion is not installed. sverify doesn't ship it (~120 MB); scripts/fetch_isthmus.sh <dir> downloads the pinned release.
• DataFusion is pinned to 51, because Ibis 12's DataFusion backend doesn't connect to later versions.

Development

git clone https://github.com/AlexanderNenninger/substrait-lean
cd substrait-lean/python
uv sync                          # builds substrait-check with lake (needs elan) and links it in
uv run poe qa                    # proofs, lint, typecheck, golden, Python tests, notebooks

See the repository for setup details.

License

sverify is licensed under the Mozilla Public License 2.0 (MPL-2.0). You can use it in any product, open or proprietary, and your own code stays yours. If you distribute a modified version of sverify's own files, those files stay under the MPL and you share their source. The license applies file by file, so it never reaches the code around them.

python/src/sverify/ibis_substrait/ is derived from ibis-substrait and stays under the Apache License 2.0.