Skip to content

overview.md

Latest commit

 

History

History
104 lines (79 loc) · 3.57 KB

overview.md

File metadata and controls

104 lines (79 loc) · 3.57 KB
Raw

Coadds: overview

A coadd is the sum of many depth-1 maps, for example all maps of one band in one week. A coadd has less noise than one depth-1 map. Thus, the pipeline can find fainter sources on longer time scales.

There are two stages:

  1. Make the coadds. The sotrp-coadd command reads depth-1 maps from mapcat and makes one coadd. It writes the coadd to FITS files and registers it in mapcat.
  2. Analyze the coadds. The sotrp command reads the registered coadds from mapcat. It runs the time-resolved pipeline on each coadd, as it does on a depth-1 map.

Mapcat connects the two stages. The second stage uses the coadds that the first stage registered.

flowchart LR

d1db[(mapcat:<br/>depth-1 maps)]
d1fits@{ shape: procs, label: "depth-1 FITS"}
coadd["sotrp-coadd"]
cfits@{ shape: procs, label: "coadd FITS"}
cdb[(mapcat:<br/>coadds + links)]
trp["sotrp"]
out@{ shape: procs, label: "pipeline outputs"}

d1db --> coadd
d1fits --> coadd
coadd --> cfits
coadd --> cdb
cdb --> trp
cfits --> trp
trp --> out
Loading

Stage 1: make the coadds

sotrp-coadd does these steps for each depth-1 map in the time window, one map at a time:

  1. It reads the map from disk.
  2. It applies the preprocessors: masks, the matched filter and cleaning.
  3. It merges the map into the coadd.
  4. It removes the map from memory.

There are two reasons for this sequence:

  • The preprocessors must see each observation separately. A source that moves (an asteroid or a planet) crosses many pixels in a sum of several days. The matched filter must use the noise properties of each observation.
  • Only one input map is in memory at a time. Thus, memory use does not increase with the number of maps.

The sotrp command also has a map_coadder, but it makes the coadd first and applies the preprocessors after. Thus, use sotrp-coadd for coadds of raw depth-1 maps.

When the coadd is complete, sotrp-coadd writes the coadd to FITS files. It then registers the coadd in mapcat, with a link to each depth-1 map in the coadd.

Stage 2: analyze the coadds

sotrp reads the registered coadds from mapcat. The coadds already have the preprocessors applied. Thus, sotrp starts with the postprocessors. Then it does the same steps as for a depth-1 map: forced photometry, source subtraction, blind search, the sifter and the outputs.

The library

The two commands use functions and classes in the sotrplib library. You can also use them directly in Python:

Function or class Module What it does
stream_coadd() sotrplib.maps.map_coadding Builds, preprocesses and merges maps one at a time.
RhoKappaMapCoadder sotrplib.maps.map_coadding Merges rho/kappa maps.
IntensityMapReader sotrplib.maps.database Selects depth-1 maps from mapcat.
register_coadd() sotrplib.maps.database Registers a coadd and its map links in mapcat.
CoaddRhoKappaMapReader sotrplib.maps.database Reads registered coadds from mapcat.
CoaddRhoAndKappaMap sotrplib.maps.core A registered coadd, read from disk.
MapOutputSerializer sotrplib.outputs.core Writes map fields to FITS files.

More information