This document will guide you through the installation process and will help you configure the library.

It is assumed that you have already installed Python 3.6.

There are two ways to install the project for Windows. The easy way is to use the install.ps1 script, which handles all the program requirements. If you don't want to use this script, an alternative method is described in the Installation without script section.

• Open PowerShell terminal as Administrator
• Go to qf-lib\win_dependencies directory
• Run the install.ps1 script
• Script will install all the dependencies and create a Python virtual environment with all the Python dependencies installed (later on you might need to activate the virtual environment; see: https://docs.python.org/3.6/library/venv.html)
• To check if everything installed correctly, check the Testing section.

NOTICE: If possible use precompiled packages (wheels) provided in the win_dependencies directory to install project's dependencies. However the provided wheels are only suitable for Windows (64-bit architecture) with Python 3.6 (64-bit).

• Install numpy by running pip install numpy-1.15.4+mkl-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install scipy by running pip install scipy-1.1.0-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install cvxopt by running pip install cvxopt-1.2.2-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install statsmodels by running pip install statsmodels-0.9.0-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install TA Lib (optional):
  • extract TA Lib SDK archive (zip) to C:\ta-lib (unfortunately this is the fixed path)
  • run pip install TA_Lib-0.4.17-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install certifi by running pip install certifi==2017.4.17
• Download and install GTK3+.
• Install other requirements using pip. Run pip install -r requirements.txt in the main directory of the QuantFin Lib.

• Download and install Bloomberg C++ SDK:
  • extract the files from the installation archive (zip) from the win_dependencies directory
  • create a new environment variable called BLPAPI_ROOT and set it to the path of the /bin directory inside the Bloomberg's directory (extracted from the zip achive)
  • add /bin directory to the PATH environment variable
• Install blpapi by running pip install blpapi-3.9.1-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install Haver by running Haver-1.1.0-cp36-cp36m-win_amd64.whl in the win_dependencies directory.
• Install Interactive Brokers platform :
  • install TWS by running TWS API Install 973.06.msi in the win_dependencies
  • go to C:\TWS API\source\pythonclient and run python setup.py install

The project has been developed using PyCharm and this is a suggested IDE for further development.

Starting directory is used to turn relative paths into absolute paths. In many places you'll be required to specify a path which should be relative to the starting directory. To set up starting directory one needs to either:

• call set_starting_dir_abs_path("C:\abs\path\to\starting\directory")
• or set QF_STARTING_DIRECTORY environment variable and put as value path to the project.
• For example the path might look like this: C:\Users\user_name\workspace\qf-lib

Many components from qf-lib require the Settings object as a dependency. To create it one needs to put in their Python code:

from qf_lib.settings import Settings


settings_path = ...
secret_settings_path = ...
settings = Settings(settings_path, secret_settings_path)

where settings_path is an absolute path of your file containing settings (JSON, described later) used by the application and the secret_settings_path which is complementary to the settings file but contains secret data (e.g. passwords) and thus mustn't be added to the CVS. When Settings object is created, it:

• loads settings defined in the settings.json file
• loads settings defined in the secret_settings.json file (if the file exists)
• loads settings defined in the QUANTFIN_SECRET environment variable (if the secret_settings.json file doesn't exist)
• merges both loaded sets of settings together.

Sample content of the settings.json:

{
  "some_setting": "value of that setting",
  "another_setting": "value of another setting",
  "some_connection_settings": {
    "username": "john.smith"
  }
}

Sample content of the secret_settings.json:

{
  "some_connection_settings": {
    "password": "my_secret_pass"
  }
}

You may define your own settings and later on have them loaded into Settings object. However there are some, which are required for some QF-Lib components to work correctly.

• Sample settings can be found in demo configuration \qf-lib\demo_scripts\demo_configuration\config_files\demo_settings.json
• Sample secret settings can be found in demo configuration \qf-lib\demo_scripts\demo_configuration\config_files\demo_secret_settings.json

NOTICE: all paths used in the settings should be relative to the starting directory. (which can be set either by using set_starting_dir_abs_path(path) function or by setting QF_STARTING_DIRECTORY environment variable)

Below you'll find a list of examples that can be put into settings

company_name and logo_path (optional)

settings.json:

    "company_name": "Sample Org Name",
    "logo_path": "path/to/logo.jpg"

Used by components producing tearsheets (in a form of PDFs). The company name and a logo is put in the header of those tearsheets.

document_css_directory (optional)

settings.json:

    "document_css_directory": "input/elements_css"

Setting used by the PDFExporter component to style the elements put in PDFs (e.g. tables, paragraphs, etc.). If document_css_directory is not specified then the default style will be applied.

bloomberg (optional)

settings.json:

  "bloomberg": {
    "host": "localhost",
    "port": 8194
  }

Used by the BloombergDataProvider. To have this component running, first you need to have a Bloomberg subscription. Then you need to have the BLPAPI running somewhere. Then you need to specify where the API is running by specifying its host and port.

email_templates_directory (optional)

settings.json:

  "email_templates_directory": "input/email_templates"

Setting used by the EmailPublisher. Email templates are HTML templates with placeholders (e.g. {{user.name}}).

output_directory

settings.json:

  "output_directory": "output"

A relative path to the directory into which different components will put their output (e.g. generated tearsheets).

smtp (optional)

settings.json:

  "smtp": {
    "host": "smtp.server.pl",
    "port": 587,
    "domain": "SOME_DOMAIN",
    "tls": true,
    "sender": "sample_user@some.domain.com"
  }

secret_settings.json:

  "smtp": {
    "username": "sample_user",
    "password": "VeryStrong P4ssw0rd with s0me Polish special characters (to confuse the hacker)"
  }

SMTP settings used by the EmailPublisher.

After the installation process you may test if everything was setup fine. In order to do so you need to download the QuantFin Lib from git and then:

1. Add path of QuantFin Lib to the PYTHONPATH environment variable (create it if it doesn't exist).
2. Go to the qf_lib_tests directory.
3. Run python run_unit_tests.py script. There should be no skipped tests and no errors if all the requirements were installed correctly.
4. Run python run_integration_tests.py script. It's better if you have established Bloomberg, Haver and Quandl connections. Otherwise some tests might be skipped. However there should be no errors.

• GTK3+
• Bloomberg C++ SDK
• Bloomberg Python
• Unofficial Windows Binaries for Python Extension Packages
• Haver
• cvxopt
• TA Lib