Fast cheminformatics fingerprint search, at your fingertips.
Chemfp is a set of command-line tools and a Python library for fingerprint generation and high-performance similarity search.
- Do you want single-threaded search of 1M 1024-bit fingerprints in under 10 milliseconds?
- Do you want to make a sparse similarity matrix from 1M 2048-bit fingerprints in less than 30 minutes on a four core machine?
- Do you want to include fingerprint similarity results in your Python web application?
- ... with fast reload times during development, and without the complexity of using a dedicated search server?
- Do you work with fingerprints from multiple chemistry toolkits, or have custom fingerprint types?
- Do you want command-line tools including sub-second similarity search times?
- Do you program in Python and want to write new fingerprint analysis programs?
- Do you want a perpetual license along with the source code?
If that sounds interesting
There are two ways to try out chemfp. From the download page page you can request an evaluation copy of the most recent version of chemfp, or you can download an earlier version for no cost under the MIT license.
Some of the many improvements in the newest version of chemfp are: higher performance, support for the FPB binary format for fast loading times, support for more than 4GB of fingerprint data, sublinear Tversky search in addition to sublinear Tanimoto search, API improvements for web development, and support for both Python 2.7 and Python 3.5+.