Skip to main content

Python loader for doltlite — Dolt version control through SQLite's drop-in API.

Project description

doltlite (Python)

A Python loader for doltlite — Dolt's version control through SQLite's drop-in API.

import doltlite          # one-time bootstrap
import sqlite3

conn = sqlite3.connect("repo.db")
conn.execute("CREATE TABLE t(x INT)")
conn.execute("INSERT INTO t VALUES (1)")
conn.execute("SELECT dolt_commit('-A', '-m', 'init')")

import doltlite makes libdoltlite the active SQLite engine for the current interpreter, so anything that uses the standard sqlite3 module — including SQLAlchemy with sqlite:///... URLs — transparently gains dolt_commit, dolt_branch, dolt_merge, dolt_log, dolt_diff_<table>, and the other Dolt SQL functions and virtual tables.

Install

pip install doltlite

Wheels bundle a precompiled libdoltlite for macOS (arm64) and Linux (x86_64, aarch64). No system-level setup required.

Intel Mac wheels aren't shipped in v0.11.x — GitHub's free-tier macos-13 runners queue for hours and block releases. Intel Mac users should build libdoltlite locally and use the DOLTLITE_LIB path below.

For development against a local checkout of doltlite, point DOLTLITE_LIB at your built library instead:

DOLTLITE_LIB=/path/to/libdoltlite.dylib python3 your_script.py

Requirements

The package piggybacks on Python's stdlib sqlite3, which must load SQLite as a shared extension at runtime. The following Pythons work:

  • Distro / system Python (Linux)
  • Homebrew Python (macOS, Linux)
  • pyenv-built Python
  • Conda Python

These do not work because their _sqlite3 is statically linked into the interpreter:

  • python-build-standalone interpreters — the default for uv python install, mise, and Rye

If you use uv, target one of the supported Pythons explicitly:

uv venv --python /opt/homebrew/bin/python3   # or /usr/bin/python3

How the bootstrap works

Doltlite is a SQLite drop-in: it implements the same sqlite3_* C API and adds Dolt-specific functions and virtual tables on top. To use it from Python, libdoltlite has to be loaded ahead of the system libsqlite3 so its symbols win during the dynamic-link symbol-resolution pass that Python's _sqlite3 module triggers when it's first imported.

The mechanics differ by platform:

Linux

ELF flat-namespace symbol resolution: if libdoltlite is loaded into the process with RTLD_GLOBAL before _sqlite3 is loaded, its sqlite3_* symbols enter the global namespace and the later libsqlite3.so lookup finds them first.

import doltlite does ctypes.CDLL(libdoltlite_path, mode=ctypes.RTLD_GLOBAL) in this case — no re-exec required.

If sqlite3 was already imported before doltlite, that boat has sailed: we fall back to re-execing the interpreter with LD_PRELOAD set, so libdoltlite is loaded at process start.

macOS

macOS uses a two-level namespace: _sqlite3.so has an LC_LOAD_DYLIB command bound to a specific libsqlite3.dylib path (e.g. /opt/homebrew/opt/sqlite/lib/libsqlite3.dylib). Plain dlopen / ctypes.CDLL does not redirect that resolution, and DYLD_INSERT_LIBRARIES alone doesn't either — the inserted library has to have an install_name (LC_ID_DYLIB) that matches the path _sqlite3.so was linked against.

import doltlite does:

  1. Detect that path via otool -L $(python3 -c 'import _sqlite3; print(_sqlite3.__file__)').
  2. Copy libdoltlite to $TMPDIR/.../libsqlite3.dylib (cached per (lib, mtime, install_name)).
  3. Rewrite the shim's install_name with install_name_tool -id <path>.
  4. Re-exec the interpreter with DYLD_INSERT_LIBRARIES=<shim>.

The two-level lookup then accepts the shim because the install_name matches.

This requires otool and install_name_tool — install Xcode Command Line Tools (xcode-select --install) if missing.

Re-exec caveats

When a re-exec is required (macOS, or Linux-after-sqlite3-loaded), the bootstrap calls os.execvpe with sys.argv. That means the invocation must name a script file we can replay:

  • python3 my_script.py
  • python3 -m my_package
  • python3 -c "import doltlite; ..." — code string isn't in argv
  • ❌ Interactive REPL / Jupyter — there's no script to re-exec

In the unsupported cases the bootstrap raises DoltliteLoadError with a clear workaround: set DYLD_INSERT_LIBRARIES (macOS) or LD_PRELOAD (Linux) yourself before starting Python.

API

import doltlite

# Side-effect import does the bootstrap automatically. Subsequent
# imports are no-ops thanks to a process-env marker.

# If you want to bootstrap explicitly (e.g. inside a function):
doltlite.bootstrap()

# Find where the loaded libdoltlite came from:
doltlite.libdoltlite_path()  # absolute path

See also

License

Apache License 2.0. See LICENSE.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

doltlite-0.11.8.tar.gz (13.1 kB view details)

Uploaded Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

doltlite-0.11.8-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.4 MB view details)

Uploaded Python 3manylinux: glibc 2.17+ x86-64

doltlite-0.11.8-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (4.8 MB view details)

Uploaded Python 3manylinux: glibc 2.17+ ARM64

doltlite-0.11.8-py3-none-macosx_11_0_arm64.whl (1.0 MB view details)

Uploaded Python 3macOS 11.0+ ARM64

File details

Details for the file doltlite-0.11.8.tar.gz.

File metadata

  • Download URL: doltlite-0.11.8.tar.gz
  • Upload date:
  • Size: 13.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for doltlite-0.11.8.tar.gz
Algorithm Hash digest
SHA256 4130e3352495c79d997ea4ab52b38f22f805fae7db1c07bbe05b3d472cb9881f
MD5 a995f9b4030015adc0a7c220d26e1639
BLAKE2b-256 a683808077752ee106ca683d1f9f92fda917578c54f2ad26565861abb5b975c2

See more details on using hashes here.

Provenance

The following attestation bundles were made for doltlite-0.11.8.tar.gz:

Publisher: wheels.yml on dolthub/doltlite-python

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file doltlite-0.11.8-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for doltlite-0.11.8-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 f2174028a2baf136294174c44f6f6089d91a8fad1981536163eb0a33e5fdbde6
MD5 6e674d18199c12c8efcd68cff95f9f47
BLAKE2b-256 e47e1dd8e9b16644f0747671e0e4284ff439a8744b02c511ef7ad85426819c70

See more details on using hashes here.

Provenance

The following attestation bundles were made for doltlite-0.11.8-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl:

Publisher: wheels.yml on dolthub/doltlite-python

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file doltlite-0.11.8-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for doltlite-0.11.8-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 4be50d5ee953180e6555e84432b0b34d3a8fc566d7b351e8581eb8b0f51b1b17
MD5 91cec79ff6db922aaab09311e600832a
BLAKE2b-256 e0c08b539268e461bb30332664cc49957a7db63512f6f44c4b3ac432783adc15

See more details on using hashes here.

Provenance

The following attestation bundles were made for doltlite-0.11.8-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl:

Publisher: wheels.yml on dolthub/doltlite-python

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file doltlite-0.11.8-py3-none-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for doltlite-0.11.8-py3-none-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 d803af367fe3946e80e0b13f81178dd0d09b90dd93b57ffb9726870fac1585f2
MD5 9721b963399118fb05130f88bcb89c19
BLAKE2b-256 a21b26ae02d964e338cd5da21f104f3c41c925b44cbd10a6f73f87eaa8a7f491

See more details on using hashes here.

Provenance

The following attestation bundles were made for doltlite-0.11.8-py3-none-macosx_11_0_arm64.whl:

Publisher: wheels.yml on dolthub/doltlite-python

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page