Skip to main content

HTTP-based interface to Stan, a package for Bayesian inference.

Project description

Stan logo

pypi version travis-ci build status test coverage report

HTTP-based REST interface to Stan, a package for Bayesian inference.

An HTTP 1.1 interface to the Stan C++ package, httpstan is a shim that allows users to interact with the Stan C++ library using a REST API. The package is intended for use as a universal backend for frontends which know how to make HTTP requests. The primary audience for this package is developers.

In addition to providing the essential functionality of the command-line interface to Stan (CmdStan) over HTTP, httpstan provides the following features:

  • Automatic caching of compiled Stan models

  • Automatic caching of samples from Stan models

  • Parallel sampling

Documentation: https://httpstan.readthedocs.org.

Requirements

  • Python version 3.6 or higher.

  • macOS or Linux.

Background

httpstan is a shim allowing clients able to make HTTP-based requests to call functions in the Stan C++ library’s stan::services namespace. httpstan was originally developed as a “backend” for a Stan interface written in Python, PyStan.

Stability and maintainability are two overriding goals of this software package.

Install

$ python3 -m pip install httpstan

Usage

After installing httpstan, running the module will begin listening on localhost, port 8080:

python3 -m httpstan

In a different terminal, make a POST request to http://localhost:8080/v1/models with Stan program code to compile the program:

curl -H "Content-Type: application/json" \
    --data '{"program_code":"parameters {real y;} model {y ~ normal(0,1);}"}' \
    http://localhost:8080/v1/models

This request will return a model name along with all the compiler output:

{"name": "models/xc2pdjb4", "compiler_output": "In file included from …"}

(The model name depends on the platform and the version of Stan.)

Drawing samples from this model using default settings requires two steps: (1) launching the sampling operation and (2) retrieving the output of the operation (once it has finished).

First we make a request to launch the sampling operation:

curl -H "Content-Type: application/json" \
    --data '{"function":"stan::services::sample::hmc_nuts_diag_e_adapt"}' \
    http://localhost:8080/v1/models/xc2pdjb4/fits

This request instructs httpstan to draw samples from the normal distribution described in the model. The function name picks out a specific function in the stan::services namespace found in the Stan C++ library (see the Stan C++ documentation for details). This request will return immediately with a reference to a long-running fit operation:

{"name": "operations/gkf54axb", "done": false, "metadata": {"fit": {"name": "models/xc2pdjb4/fits/gkf54axb"}}}

Once the operation is complete, the “fit” can be retrieved. The name of the fit, models/xc2pdjb4/fits/gkf54axb, is included in the metadata field of the operation. The fit is saved as sequence of Protocol Buffer messages. These messages are strung together using length-prefix encoding. To retrieve these messages, saving them locally in the file myfit.bin, make the following request:

curl http://localhost:8080/v1/models/xc2pdjb4/fits/gkf54axb > myfit.bin

To read the messages you will need a library for reading the encoding that Protocol Buffer messages use. In this example we will read the first message in the stream using the Protocol Buffer compiler tool protoc. (On Debian-based Linux you can find this tool in the protobuf-compiler package.) The following command skips the message length (one byte) and then decodes the message (which is 13 bytes in length):

dd bs=1 skip=1 if=myfit.bin 2>/dev/null | head -c 13 | \
  protoc --decode stan.WriterMessage protos/callbacks_writer.proto

Running the command above decodes the first message in the stream. The decoded message should resemble the following:

topic: LOGGER
feature {
  string_list {
    value: "info:"
  }
}

This empty logger message is not particularly informative. The next message is more interesting. It reads:

topic: LOGGER
feature {
  string_list {
    value: "info:Gradient evaluation took 4.7e-05 seconds"
  }
}

The Stan “fit”, saved in myfit.bin, aggregates all messages. By reading them one by one you can recover all messages sent by the Stan C++ library.

Contribute

Contribution guidelines are described in CONTRIBUTING.rst.

Citation

We appreciate citations as they let us discover what people have been doing with the software. Citations also provide evidence of use which can help in obtaining grant funding.

Allen Riddell, and Ari Hartikainen. 2019. Stan-Dev/Httpstan: V1.0.0. Zenodo. https://doi.org/10.5281/zenodo.3546351.

License

ISC 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 Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

httpstan-2.0.5-cp38-cp38-manylinux2014_x86_64.whl (50.8 MB view hashes)

Uploaded CPython 3.8

httpstan-2.0.5-cp38-cp38-macosx_10_9_x86_64.whl (48.6 MB view hashes)

Uploaded CPython 3.8 macOS 10.9+ x86-64

httpstan-2.0.5-cp37-cp37m-manylinux2014_x86_64.whl (50.8 MB view hashes)

Uploaded CPython 3.7m

httpstan-2.0.5-cp37-cp37m-macosx_10_9_x86_64.whl (48.6 MB view hashes)

Uploaded CPython 3.7m macOS 10.9+ x86-64

Supported by

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