Please select your preferred option, perhaps the one that suits best your existing workflow:
Tomographer provides an interface for easy invocation from Python code. Here are the steps to get started and analyze data from your experiment:
Install the necessary Python packages.
First, install tomographer’s dependencies. Depending on whether you’re
using conda or pip by
itself, you should run the following commands:
# for *conda* users:
> conda install numpy gcc libgcc
> conda install -c conda-forge "pybind11>=2.2"
# OR, for *pip* users:
> pip install numpy "pybind11>=2.2"
All pip install commands (e.g. pip install SOMEPACKAGE) might have to be
prefixed by sudo -H (e.g. sudo -H pip install SOMEPACKAGE) if you need
administrator priviledges, or you can use the option --user to install
into a user directory (e.g. pip install SOMEPACKAGE --user).
Now, install tomographer itself, using pip (this should also work within
a conda environment):
> pip install tomographer
In case of trouble, you should seek additional information and instructions in the README file.
Open your favorite Python environment (e.g. edit a Python script or open an IPython or Jupyter notebook session), and start off with this one-qubit example.
Note: The script is designed to display well within a Jupyter notebook, including real-time feedback on the progress of the random walk.
Edit at wish to get a hang of how the Python call works. Read the documentation of
the tomographer.tomorun.tomorun() Python function, the
main entry point to our procedure from Python code.
As in the example, the histogram fit, as well as the calculation of the quantum error bars can be handled by some provided python routines
Write up and publish paper. We will be very happy if you cite us :)
Philippe Faist and Renato Renner. Practical and Reliable Error Bars in Quantum Tomography. Phys. Rev. Lett. 117:1, 010404 (2016). arXiv:1509.06763
Philippe Faist. The Tomographer Project. Available at: https://github.com/Tomographer/tomographer/
tomorunThe Tomographer project provides the tomorun program which analyses data from a quantum
tomography experiment.
Download and install Tomographer/tomorun. You do not need to set up the Python interface.
Get familiar with tomorun. Get an idea of what the program does and how to use it with:
> tomorun --help
Create a MATLAB data file with your experimental data, as specified in tomorun’s help
text, encoded in the MATLAB variables dim, Emn, and Nm.
Don’t forget to properly model noise in your measurement device into effective POVM effects.
The MAT file format is widely recognized by Mathematica, numpy/scipy, … so that it shouldn’t be an issue to generate the data file.
Don’t forget to save in the appropriate MATLAB file version so that it can be read by
tomorun. Just to be safe, in MATLAB, you should use the save command with an
explicit version option:
>> save('my_data_file.mat', ..., '-v6')
(This depends on the version and configuration of the
MatIO library tomorun was compiled
with.)
Create a configuration file for tomorun for your experiment (or edit this example
configuration). This is
simply a list of tomorun options with corresponding values, and optional comments.
Use the --data-file-name option to specify the name of the file where you saved your
data (step above). To keep things simple, put the configuration file and data file in
the same directory.
The option --value-type allows you to specify the figure of merit you want to
track. Set the histogram range and number of bins using --value-hist.
Fine-tune the Metropolis-Hastings random walk with the options --step-size,
--n-sweep, --n-therm, and --n-run. Specify the number of random walk instances to
run with --n-repeats (typically one per available CPU). Use a binning analysis
because that’s better (it’s on by default).
The output is controlled by the options --write-histogram and
--write-histogram-from-config-file-name. When using the option
--write-histogram=<filename>, then the file written will be
<filename>-histogram.csv, and will be overwritten if it already exists.
Run tomorun.
> tomorun --config my_config_file
This will run the Metropolis-Hastings random walk according to the given settings.
Hit Ctrl-C to get an intermediate progress report, and to see if the parameters of
the random walk are appropriate. You should have an acceptance ratio roughly in between
0.2 and 0.4. Hit Ctrl-C twice in short succession to abort, for instance if you want
to change the parameters of the random walk and start again.
The output consists in a tab-separated table of values as documented in tomorun
--help. The output file is specified with the options --write-histogram and
--write-histogram-from-config-file-name.
Analyse the output. (Read Ph. Faist and R. Renner, Phys. Rev. Lett. 117, 010404 (2016) (arxiv:1509.06763) to understand the theory behind this method.)
Depending on what your figure of merit is, and what you indend to do, fit the numerical
estimate of the distribution of the figure of merit with the appropriate model. A
MATLAB tool is provided in the source tomographer distribution to help you
(tools/analyze_tomorun_histogram.m),
which uses the Curve Fitting Toolbox to analyze the output from tomorun.
A Python script which is capable of analyzing output from tomorun is also
provided
(tools/analyze_tomorun_histogram_simple.py)
(note this requires the tomographer python package to be installed, see
instructions above). Run the script with --help to get information on its
possible command-line options.
Write up and publish paper. We will be very happy if you cite us :)
Philippe Faist and Renato Renner. Practical and Reliable Error Bars in Quantum Tomography. Phys. Rev. Lett. 117:1, 010404 (2016). arXiv:1509.06763
Philippe Faist. The Tomographer Project. Available at: https://github.com/Tomographer/tomographer/