EPAM Indigo projects
Copyright (c) 2009-2022 EPAM Systems, Inc.
Licensed under the Apache License version 2.0
Introduction
This repository includes:
- Bingo: Chemistry search engine for Oracle, Microsoft SQL Server and PostgreSQL databases
- Bingo-Elastic: Set of APIs for efficient chemistry search in Elasticsearch
- Java API. Full README is available here
- Python API. Full README is available here
- Indigo: Universal cheminformatics library with bindings to .NET, Java, Python, R and WebAssembly, and the following tools:
- Legio: GUI application for combinatorial chemistry
- ChemDiff: Visual comparison of two SDF or SMILES files
- indigo-depict: Molecule and reaction rendering utility
- indigo-cano: Canonical SMILES generator
- indigo-deco: R-Group deconvolution utility
Detailed documentation is available at http://lifescience.opensource.epam.com
Changelog could be found in CHANGELOG.md.
Download
https://lifescience.opensource.epam.com/download/indigo.html
Bindings in public repositories:
Source code organization
Main directory structure layout:
api: Indigo API sourcesbingo: Bingo sourcescore: Core algorithms and data structures sourcesthird_party: sources for third-party librariesutils: utilities sources
Each project is placed in the corresponding directory with CMakeList.txt configuration
file, that does not include other projects. In order to build the whole project with the
correct references you need to use CMake configurations from the build_scripts directory.
Develop using dev container
This is an alternative setup solution next to have it on your PC directly. It builds on having the tools in a docker container and VSCode acts as a client.
Prerequisites
- Docker Desktop (or docker engine if you prefer)
- VSCode
Installation Steps
- Clone this repository
- Open in VSCode
- Follow recommendation to install the Dev Container Extension
- Open in Dev Container
The first time takes some minutes but afterwards you can directly develop from VSCode.
Remark for Windows Users: Its a known limitation that on windows the overlay driver is super slow, so if you want to have
a fast IDE, clone your container into a volume rather than natively on the client or in your WSL distribution.
Preinstalled build tools
To build the project from the sources, the following tools should be installed:
Required:
- GIT 1.8.2+
- C/C++ compilers with C++14 support (GCC, Clang and MSVC are officially supported)
- CMake 3.4+
- Python 3.6+
Required to build all targets:
- JDK 1.8+
- .NET Standard 2.0+
Required to build Indigo-WASM:
Dependencies:
Python
- wheel package installed. Command:
python -m pip install wheel
- setuptools version less than 72.0.0 package installed. Command:
python -m pip install setuptools==68.0.0
- waitress package installed (to run backend API test). Command:
python -m pip install waitress
- flasgger package installed (to run backend API test). Command:
python -m pip install flasgger
- psycopg2 package installed (to run backend API test). Command:
python -m pip install psycopg2
- sqlalchemy package installed (to run backend API test). Command:
python -m pip install sqlalchemy
- numpy package installed (to run backend API test). Command:
python -m pip install numpy
- celery package installed (to run backend API test). Command:
python -m pip install celery
- marshmallow package installed (to run backend API test). Command:
python -m pip install marshmallow
- redis package installed (to run backend API test). Command:
python -m pip install redis
- flask_httpauth package installed (to run backend API test). Command:
python -m pip install flask_httpauth
- pyparsing package installed (to run backend API test). Command:
python -m pip install pyparsing
- requests package installed (to run backend API test). Command:
python -m pip install requests
- pillow package installed (required for rendering tests). Command:
python -m pip install pillow
On Linux use python3 insted of python. Using virtual environment might be required as well.
Build instruction
On Windows use cmd to run the commands.
-
Create build folder
mkdir build
-
Move to build folder
cd build
-
Run CMake to configure the project with desired options. For instance:
cmake .. -DBUILD_INDIGO=ON -DBUILD_INDIGO_WRAPPERS=ON -DBUILD_INDIGO_UTILS=ON
-
Build Indigo from console:
cmake --build . --config Release --target <target name>
Replace <target name> with any of the following targets you need:
- ALL_BUILD (on Windows)
- all (on Linux)
- indigo-dotnet
- indigo-java
- indigo-python
Build results could be collected from Indigo/dist folder.
'indigo-python' target is commonly used.
Run tests
Befor running any test you have to build and install indigo-python
- Build indigo-python using
--target indigo-python or --target ALL_BUILD(on Windows) or --target all(on Linux). See Build instruction above.
- On Windows the package should be in 'Indigo/api/python/dist' folder
- On Linux it is located in 'Indigo\dist' folder
Package will be named like 'epam.indigo-<version-arch>.whl'. For instance: epam.indigo-1.29.0.dev2-py3-none-win_amd64.whl
-
Install package using pip
- If Indigo package has been already installed, uninstall it with the following command:
python -m pip uninstall <path-to-.whl-file> -y
python -m pip install <path-to-.whl-file>
Replace <path-to-.whl-file> with the right path to .whl package. For instance: python -m pip install ../api/python/dist/epam.indigo-1.29.0.dev2-py3-none-win_amd64.whl
-
Run integration test
to run all test
python api/tests/integration/test.py -t 1
to run tests by mask use test_name
python api/tests/integration/test.py -t 1 -p test_name
to update all tests "ref/" using -u flag
python api/tests/integration/test.py -u
to update all tests "ref/" without -u flag
export INDIGO_UPDATE_TESTS="True"
python api/tests/integration/test.py
to update a single test's "ref/" by mask use test_name and -u flag
python api/tests/integration/test.py -t 1 -p test_name -u
To run backend API test:
-
Build and install indigo-python
-
Set environment variable by running this command:
for Linux
export INDIGO_SERVICE_URL=http://localhost:5000/v2
for Windows in PowerShell
$env:INDIGO_SERVICE_URL="http://localhost:5000/v2"
-
Run backend service :
cd utils/indigo-service/backend/service
cp v2/common/config.py .
waitress-serve --listen="127.0.0.1:5000 [::1]:5000" app:app
you may use any port instead of 5000
-
Run backend API test:
python utils/indigo-service/backend/service/tests/api/indigo_test.py`
use -k test_name to run test by pattern.
How to build Indigo-WASM
Build tools prerequisites
Make sure git is running from path:
>git --version
git version 2.26.2.windows.1
Make sure python is running from path:
>python --version
Python 3.9.0
Make sure cmake is running from path:
>cmake --version
cmake version 3.18.4
Download corresponding ninja-xxx.zip and unpack to folder on path.
Make sure it's running from path:
>ninja --version
1.10.2
git clone https://github.com/emscripten-core/emsdk.git
cd emsdk
./emsdk install latest
./emsdk activate latest
source ./emsdk_env.sh
Note: On Windows, run emsdk instead of ./emsdk, and emsdk_env.bat instead of source ./emsdk_env.sh, use cmd instead of powershell.
Get Indigo sources
Clone (or checkout) Indigo repository
>git clone https://github.com/epam/Indigo.git
Build Indigo
For each new session, set environment anew:
>cd emsdk
>./emsdk activate latest
If fresh build:
>mkdir build
>cd build
Now build:
>emcmake cmake .. -DCMAKE_BUILD_TYPE=Debug -G Ninja
>ninja indigo-ketcher-js-test
How to build conda package
Prepare conda build environment as described at https://docs.conda.io/projects/conda-build/en/stable/install-conda-build.html
Change directory to ```recipe-conda``, set INDIGO_VERSION to existing Indigo version(already published at PyPi), run conda build:
Linux/Mac:
>cd conda-recipe
>INDIGO_VERSION=1.29.0 conda build .
Windows(using cmd):
>cd conda-recipe
>set INDIGO_VERSION=1.29.0
>conda build .
To upload packages: install anaconda-client, login, and upload packet using command provided by conda-build in output
>conda install anaconda-client
>anaconda login
Enter username:
Enter password:
>anaconda upload /conda-build-dir/packet-name
