This is the Python version, which is a generalized, optimized and more up-to-date version of the C++ demultiplexer named "findIndexes" which you can find here https://github.com/pelinakan/UBD.
See http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0057521 for the peer-reviewed reference to the program.
The main idea is to extract the cDNA barcodes from the input file (FASTQ, FASTA or BAM) and then try to find a match in a file with a list of reference barcodes using a kmer-based approach. All the reads that match will be outputted and the barcode and spatial information will be added to each record.
TagGD requires a file containing the indexes(barcodes) to be used to demultiple. The file should look like this:
BARCODE X Y
TagGD could potentially be used to demultiplex any type of index if the refered file is provided (one culd just create fake X,Y coordinates).
TagGD allows to demultiplex reads that contain several indexes.
See Wiki section
TagGD requires PySam and Numpy.
(Assuming that you have a virtual environment installed such as Anaconda 2.7)
cd <taggd demultiplexer root>
python setup.py build
python setup.py install
Alternatively, you can use PIP:
pip install taggd
taggd_demultiplex.py -h