Create the Python Project

In this section we shall create a Python file based on the standard template that solves the example business problem.

We shall edit the python file named application.py to contain the optimization problem. To begin the process, complete the following steps:

Open Xpress Workbench.
The landing page is displayed. Select CREATE INSIGHT (PYTHON) PROJECT.
Select a folder in which to store the project.
A new template project is created containing client_rescources, model_resources, and python_source folders, as well as an application.xml file.
Click File > New > Python File. Save the new file in the python_source folder as application.py to overwrite the existing file.
We shall start by importing all the libraries that will be used in this tutorial.
```
import xpressinsight as xi
import xpress as xp
import pandas as pd
import sys
```

Below that block, enter the following:

@xi.AppConfig(name="Portfolio Optimization", version=xi.AppVersion(1, 0, 0))
class InsightApp(xi.AppBase):
    # Input entities
    MaxHighRisk: xi.Scalar(default=0.33, alias="Maximum investment into high-risk values")
    MaxPerShare: xi.Scalar(default=0.25, alias="Maximum investment per share")
    MinNorthAmerica: xi.Scalar(default=0.45, alias="Minimum investment into North-American values")
    ShareIds: xi.Index(dtype=xi.string, alias="Shares")

This section declares the parameters that will be used as constraints when performing optimization. The specified default values can be changed for every model execution without any need to edit the model source.

Below the constraints block, enter the following declarations:

# Input and result entities indexed over ShareIds
Shares: xi.DataFrame(index="ShareIds", columns=[
    xi.Column("Return", dtype=xi.real, format="$ 0.00", alias="Expected Return on Investment"),
    xi.Column("HighRisk", dtype=xi.boolean, alias="High-risk value"),
    xi.Column("NorthAmerica", dtype=xi.boolean, alias="Issued in North America"),
    xi.Column("fraction", dtype=xi.real, format="0.0%", alias="Fraction used", manage=xi.Manage.RESULT)
])

This section of code declares the entities that will be used to store the different data categories.

Below the inputs block, enter the following code snippet:

# Result entities
TotalReturn: xi.Scalar(dtype=xi.real, alias="Total expected return on investment", manage=xi.Manage.RESULT)
SummaryIds: xi.Index(dtype=xi.string, alias="Summary", manage=xi.Manage.RESULT)
SummaryValues: xi.Series(index="SummaryIds", dtype=xi.real, manage=xi.Manage.RESULT)

This section declares and names the data series that will be used to hold the result.

The final declaration specifies the source data file.

# Constant class attribute
    DATAFILE = "shares.csv"

Below the result line, add the two required mode functions-These must be included in any Python Insight file.

First declare the LOAD mode function.

@xi.ExecModeLoad(descr="Load input data and initialize all input entities.")
    def load(self):
        print("Loading data.")
        self.Shares = pd.read_csv(InsightApp.DATAFILE, index_col=['ShareIds'], squeeze=True)
        self.ShareIds = self.Shares.index
        print("Loading finished.")

The largest block of code is the RUN mode.

@xi.ExecModeRun(descr="Solve problem and initialize all result entities.")
def run(self):
    print('Starting optimization.')

    # Create Xpress problem and variables
    p = xp.problem("portfolio")
    self.Shares['fractionVar'] = pd.Series(xp.vars(self.ShareIds, vartype=xp.continuous, name='fractionVar'))
    p.addVariable(self.Shares.fractionVar)

    # Objective: expected total return
    objective = xp.Sum(self.Shares.Return * self.Shares.fractionVar)
    p.setObjective(objective, sense=xp.maximize)

    # Limit the percentage of high-risk values
    limit_high_risk = \
        xp.Sum(self.Shares.loc[self.Shares.HighRisk, 'fractionVar']) \
        <= self.MaxHighRisk
    p.addConstraint(limit_high_risk)

    # Minimum amount of North-American values
    limit_north_america = \
        xp.Sum(self.Shares.loc[self.Shares.NorthAmerica, 'fractionVar']) \
        >= self.MinNorthAmerica
    p.addConstraint(limit_north_america)

    # Spend all the capital
    p.addConstraint(xp.Sum(self.Shares.fractionVar) == 1)

    # Upper bounds on the investment per share
    for share_id, share in self.Shares.iterrows():
        p.addConstraint(share.fractionVar <= self.MaxPerShare)

    # Solve optimization problem
    p.solve()

    # Save results and key indicator values for GUI display
    self.Shares["fraction"] = pd.Series(p.getSolution(), index=self.ShareIds)
    self.TotalReturn = p.getObjVal()

    self.SummaryValues = pd.Series({
        "Expected total return": self.TotalReturn,
        "Total high risk shares": limit_high_risk.ub - p.getSlack(limit_high_risk),
        "Total North-American": limit_north_america.lb - p.getSlack(limit_north_america),
        "Largest position": self.Shares.fraction.max()
    })
    self.SummaryIds = self.SummaryValues.index
    print('\n', self.SummaryValues, '\n', sep='')
    print('Optimization finished.')

Finally, at the end of the file, add the main class declaration that specifies which instructions Python should initially run.
```
if __name__ == "__main__":
    app = xi.create_app(InsightApp)
    sys.exit(app.call_exec_modes(["LOAD", "RUN"]))
```

If you attempt to publish to Xpress Insight at this point, it will fail, because the compiler will recognize there is a missing file. We must add in the data file that is required by the Python script.

Create a new file. In Xpress Workbench, click File > New > Blank File, or use the keyboard shortcut Ctrl-N. Name the new file shares.csv. Save the file to the model_resources folder.

Double click shares.csv to open it in Xpress Workbench. Copy and paste the following text into the file, then save and close it.

ShareIds,Return,HighRisk,NorthAmerica
Bank,6.0,False,False
Brewery,8.0,False,False
Cars,7.0,False,False
Electronics,21.0,True,False
Hardware,17.0,True,True
Highways,9.0,False,False
Software,31.0,True,False
Telecom,12.0,True,True
Theater,26.0,True,True
Treasury,5.0,False,True

Now you can publish the application to Xpress Insight.

Contents

Index

Glossary

Search Results

Create the Python Project