model-connector 0.0.4b1

Library for connection of TransCAD data files to Python

model-connector: Library for accessing travel model data

Installation:

Install using pip

pip install model-connector

Use Cases

• TransCAD fixed format binary tables: fast read and write using pandas dataframes.
• TransCAD matrix files: read a TransCAD matrix into a numpy array (beta)

TransCAD Binary Files

Reading from a TransCAD binary file

import model_connector as mc

# Read a TransCAD fixed format binary file into a dataframe, 
# retaining null values
df = mc.read_ffb("myfile.bin")

# Read a TransCAD fixed format binary file into a dataframe, 
# converting null values to zeros
dfz = mc.read_ffb("myfile.bin", null_to_zero=True)

Writing to a TransCAD binary file

import model_connector as mc
import pandas as pd

df = pd.DataFrame({"ID":[1,2,3], 
                   "RealField":[1.1, 2.2, 3.3, ],
                   "IntField":[1, 2, 3],
                   "DateField":[20210101, 20210115, 20210130],
                   "DateTimeField":[pd.Timestamp("2021-01-01 00:00:05"),
                                    pd.Timestamp("2021-01-15 00:10:05"),
                                    pd.Timestamp("2021-01-30 00:20:05")]})

mc.write_ffb(df, "sample_output.bin")

TransCAD Matrix Files

Matrix access requires TransCAD 9 or greater with a valid license.

Reading a TransCAD Matrix File

import model_connector as mc
import numpy as np

# Reading a TransCAD matrix into a dictionary of numpy arrays, 
# using the default behavior of converting null to zero.
matrix = mc.read_matrix_tc("myfile.mtx")
for core in matrix.keys():
    print(f"Sum of {core} = {matrix[core].sum()}" )

# Optionally Convert null to NAN. This will fail for integer matrices since
# numpy doesn't support NAN in integer arrays.
matrix_withnan = mc.read_matrix_tc("myfile.mtx", missing=mc.ConvertMissing.NAN)
for core in matrix_withnan.keys():
    print(f"Sum of {core} = {matrix_withnan[core].sum()}" )

# By default, all matrix cores are loaded. Use the cores parameter to load only
# some matrix tables.
matrix_autos = mc.read_matrix_tc("myfile.mtx", cores=["DA", "SR2", "SR3p"])
for core in matrix_autos.keys():
    print(f"Sum of {core} = {matrix_autos[core].sum()}" )

# By default, data is read using the base index. Use the row_index and 
# col_index parameters to read data using a different index.
# Indices must be specified as an integer. the base index is 0, and
# the remaining indices are ordered as returned by GetMatrixIndexNames()
# in GISDK.

matrix_subset = mc.read_matrix_tc("myfile.mtx", row_index=1, col_index=1)
for core in matrix_subset.keys():
    print(f"Sum of {core} = {matrix_subset[core].sum()}" )

print(f"Shape of matrix is {matrix_subset["DA"].shape}")

Writing a TransCAD Matrix File

This functionality has not yet been implemented.

Specifying the TransCAD Installation

The TransCAD matrix functions search the system registry to locate a TransCAD installation. If the matrix functions fail even with a valid installation, this may be due to multiple versions or licensing methods being present on the computer.

The TransCAD program location can be specified using set_tcpath(). The current path can be obtained using get_tcpath(). If the path is not set manually, get_tcpath() will report the detected version of TransCAD.

# Set the TransCAD location as shown below
import model_connector as mc
mc.set_tcpath("C:/Program Files/TransCAD 10.0/tcw.exe")
print("TransCAD Path: ", mc.get_tcpath())