tess-tiler

Turns big FFI data into small FFI data.

tess-tiler is a fast and memory-efficient package designed to convert collections of TESS Full Frame Images (FFI) into sub-sections which can fit into memory. This package uses a pre-defined tiling scheme which allows users to refer to a sub-sections of FFIs in a standardized way. It is based upon the map tiling scheme in use by Google Maps.

The tess-tiler tiling system

Each tile is uniquely identified by six parameters:

>>> import tess_tiler as tt
>>> tile = tt.TessTile(sector, camera, ccd, zoom, x, y, margin=0)

These parameters are defined as follows:

• sector, camera, and ccd are the standard TESS full-frame image identifiers.
• The zoom level is defined such that an FFI fits entirely into a single tile at zoom level 0. All subsequent zoom levels divide the image into a grid of 2^zoom x 2^zoom tiles.
• The (x, y) parameters identify the position of the tile along the column and row direction of the FFI. Their values range between 0 and 2^zoom - 1. Tile (x, y) = (0, 0) is the bottom left corner of the FFI.
• An optional margin parameter can be specified to allow tiles to overlap by a specified number of pixels.

Examples

tile definition	description
TessTile(1, 2, 3, 0, 0, 0)	Full FFI for Sector #1, Camera #2, CCD #3.
TessTile(1, 2, 3, 1, 0, 0)	Bottom-left quadrant (1/4th) of the same FFI.
TessTile(1, 2, 3, 1, 1, 1)	Top-right quadrant.
TessTile(1, 2, 3, 2, 2, 0)	Bottom-right octant (1/8th).
TessTile(1, 2, 3, 2, 0, 2)	Top-left octant.

Memory requirements

The memory required to store a single tile as a function of the zoom level is as follows:

zoom level	tiles/ccd	tile_size (px)	sky area	bytes/sector
0	1	2048 x 2048	143 deg²	64 GB
1	4	1024 x 1024	36 deg²	16 GB
2	16	512 x 512	9 deg²	4 GB
3	32	256 x 256	2 deg²	1 GB
4	64	128 x 128	0.6 deg²	256 MB
5	128	64 x 64	0.1 deg²	64 MB
6	256	32 x 32	0.04 deg²	16 MB

In this table, bytes/sector denotes the amount of memory required to cut out 27 days worth of 10-minute FFI data at 32 bits per pixel.

Example use

tess-tiler provides a fast and memory-efficient way to extract the image data for an entire sector of FFIs, without utilizing more than the memory required by the specific tile.

>>> import tess_tiler as tt
>>> tile = tt.TessTile(sector=1, camera=2, ccd=3, zoom=4, x=2, y=3)
>>> loader = tt.TileLoader(ffi_filenames)
>>> tpf = loader.create_tpf(tile)

To create a list of TessTile objects that cover a specific coordinate in the sky, you can use the sister tess-fov package:

>>> import tess_fov as tf
>>> tiles = tf.radec_to_tiles(ra, dec, zoom=3)  # returns a list of TessTile objects

To obtain the parameters of all the tiles that observed a known Solar System object, you can use the sister tess-ephem package:

>>> import tess_ephem as te
>>> tiles = te.asteroid_to_tiles("Vesta", zoom=2)  # returns a list of TessTile objects

How does tess-tiler compare to other tools?

Several excellent tools already exist which focus on enabling users to extract or download sub-regions from FFI images via the web, for example:

• TessCut enables users to download regions up to 100x100 pixels in size centered on an arbitrary coordinate, utilizing pre-computed data cubes stored in the cloud. In contrast, tess-tiler focuses on efficiently extracting tiles from a locally-stored collection of FFI files, according to a pre-defined tiling scheme.
• eleanor enables users to create or download 104-by-148 pixel cutout regions called postcards. tess-tiler expands this concept by providing a generic tool to create such postcards at different zoom levels from a local collection of FFI files.