This code allows easy exploration of NGTS lightcurves, by picking objects from their bulk noise properties.

It is made of two components, the interactive web-based data viewer, and an IPython session for data manipulation.

### Usage

It runs a web server which listens on a specifiable port. You connect your browser to this port, and the landing page opens:

The fractional rms plot (yellow points) can be clicked on, filling in the empty lightcurve and coordinate plots. When this occurs the objects' id appears near the top and is a link to the detail page, which shows the same information but with a url that is fixed:

The url of this example page is /view/14289, matching the object id at the top.

Also available are the objects coordinates, and a link to a Simbad search in the nearby area.

As the software runs in the browser, the user must be able to connect to the port the server is listening on. By default this is 5000, but this can be configured by command line arguments.

1. Forward your port
2. Find the pipeline output file you wish to analyse
   • e.g. /ngts/pipedev/ParanalOutput/nightly_data/20150909-ng2000-802-custom-flat-high-quality.fits
3. Run the command: bash /home/sw/work/NGTS/lightcurve-visualiser/run_from_anywhere.sh <data file> -b <bin number> -p <port>

#### Running the server locally

The main server is written with Flask, and requires:

• flask
• fitsio
• numpy
• scipy
• astropy

I recommend the conda package manager. The javascript packages come from a CDN so do not need to be installed.

There are three methods for this. In approximate order of ease:

In the ssh command used to get to ngtshead, add the following flag:

-L 5000:localhost:5000

This forwards port 5000 on the remote machine (the port that the server is listening on) to localhost:5000 i.e. on your local machine. If port 5000 is not available, for example if another user is hosting on port 5000 then an alternative remote port can be specified:

-L 5001:localhost:5000

This ports remote port 5001 to local port 5000.

If you forget to forward the correct port when you ssh in to run the server, another ssh session can be started, with the -L flag mentioned in Forward on connection.

ssh understands a magic key combination which is a little tricky to do. There must be a new line (so I usually mash enter about 5 times) and press:

~C

This gives a little prompt ssh> under your cursor and you can forward ports from inside the active session. This control is running outside your logged in shell so behaves a little strange, for example shell history and tab completion do not work, and you must press escape to leave.

In this shell you can then add a -L flag as per Forward on connection.

Once an interesting object has been found, it is worth exploring in an interactive python session. The script get_long_lightcurves.py opens the main data file and extracts the useful data for the lightcurve. It defines some handy helper functions and tools for binning, sigma clipping and periodic searches.

Please be sure you have read the section on X forwarding!

Run the script get_long_lightcurves.py with the data file argument (same as the web viewer) and -a <aperture number> e.g. 14289. This launches an ipython session with a short help text on what variables are available.

• hjd
• rawflux (may be None)
• tamflux
• casuflux (may be None)
• skybkg
• imagelist
• cat (catalogue entry)
• sc (sigma clipped tamflux, 2 iterations)
• nights (list of nights)

bin: function returning binned data. Supply the y values (e.g. flux) and (optionally) x values and a number of bins, and get back the binned x and binned y nightly: generator, taking x and y data and the array of nights and yielding an (x/y) pair for each night nightly_bin: return the binned nightly data nightly_average: nightly normalised data period_fit: create a gatspy model for finding periodic signals through Lomb Scargle. Supply the x, y and period range. See the gatspy docs for more information periodogram: given a gatspy model, compute the period and power information for rendering a periodogram

In a similar way that the server requires port forwarding, this interactive tool when run on a remote machine requires X forwarding. This is achieved with the -X flag to ssh, but cannot be set up from another terminal or through the ssh magic key combination.

If you wish to run this code locally, e.g. if you have the data already then the following packages are required, in addition to the packages required for the server:

• ipython
• gatspy
• supersmoother