Overview

ReadIM is a C++ wrapper to load DaVis Images and Vectors and is a 'low level' wrapper of C++ libraries provided by LaVision GMBH. ReadIMX source was latest updated by LaVision in Aug-2014.

A higher level module: "IM" exists to work with images and vectors. It isn't hosted on PyPi however it can still be installed with pip. It provides more convenient file read / write capability and is the recommended starting point to read and write IM7/VC7 files.

Installation

This module must be compiled to work correctly. If there isn't a binary on pip you'll need to have the appropriate build tools installed for it to compile and install properly.

pip install ReadIM

Usage

To load a .vc7 file run

import ReadIM
filename = ReadIM.extra.get_sample_vector_filenames()[0]
buffer, atts   =  ReadIM.extra.get_Buffer_andAttributeList(filename)
v_array, _ = ReadIM.extra.buffer_as_array(buffer)
v_array.shape
del(v_array)    # Required
ReadIM.DestroyBuffer(buffer)
ReadIM.DestroyAttributeListSafe(atts)

similarly for a .im7 file run

import ReadIM
filename = ReadIM.extra.get_sample_image_filenames()[0]
buffer, atts   =  ReadIM.extra.get_Buffer_andAttributeList(filename)
im_array, _ = ReadIM.extra.buffer_as_array(buffer)
im_array.shape
del(im_array)   # Required
ReadIM.DestroyBuffer(buffer)
ReadIM.DestroyAttributeListSafe(atts)

Writing files

(requires existing buffer and atts first)

atts = ReadIM.load_AttributeList({'attribute':'value'})
ReadIM.WriteIM7('saved_file.im7', True, buffer, atts.next)

Memory leaks

Memory cleanup of the buffer and attribute list is not automatic. Both the buffer and attribute list must be destroyed manually. All references to the buffer and attributes should be removed before attempting to destroy the memory, failing to do so will crash python.

del(im_array)
ReadIM.DestroyBuffer(buffer)
ReadIM.DestroyAttributeListSafe(atts)

VC7 files

Depending on the filetype, there could be several frames that make up the optimal vector field as decided by DaVis. For a full description of the buffer you should contact LaVision support. Below is a link for some code snippets. The higher level "IM" automatically reads the optimal result.

see the function "_get_vectors" at https://bitbucket.org/fleming79/im/src/master/IM/core.py