This directory contains three datasets, one created by Svea ca. 2010 (q.rd),
another done by Jonas in 2018 (q2.rd), and then a third one by Jonas
from 2021 (q3.rd).

TODO: create relationships between the three things as q3 goes live.

All are lists of candidates for "positional" lensing.

eDR3 update
===========

As before:

* dachs imp q3 build_highpm (where "high proper motion" is now not
  to be taken too literally; this will do a seqscan through edr3 and
  hence takes a while)
* dachs imp q3 build_rawcands -- the result is plc3.rawcands.
* then some magic happens at Jonas' end (TODO: figure it out) yielding
  data3/amlensing.fits.
* I run the import.


Jonas' new data
===============

Jonas created his list as follows:

* dachs imp q2 build_highpm; this yields a table plc2.highpm of high proper
  motion stars from Gaia DR2 together with the areas we should check for
  stars (because they might be lensed)
* dachs imp q2 build_rawcands; this actually checks wether suitable
  stars are within the regions of interest established in step 1.  This
  results in a table plc2.rawcands.
* Jonas pulled the result and did processing on his box, which
  eventually encluded running creation/amlensing.py
* That script adds some quality information obtained from a full DR2 table
  (which is in data/hpms.vot; see the paper for how it came to be), and
  then computes the impact geometries to figure out which events might
  actually be observable.
* This gave data/amlensing2.fits, which we're then ingesting.


Svea's old data
===============

The final list is created in three steps:

* Select candiate lens stars using ``gavo imp q import_highpm``; this runs
  SQL scripts pulling in data in a common format into a table plc.highpm.
* Select candiate lens-lensed pairs doing a crossmatch with PPMXL with
  some tricks.  Use ``gavo imp q match``, data is left in
	plc.rawcands
* Compute predicted encounters, etc, for the object pairs in plc.rawcands
  doing ``gavo imp q compute``.  What's actually executed is
	defined in res/makecands.py.


To make the tests run on a development machine that doesn't have the
full source catalogs, transfer a dump:

alnilam: dachs dump create trans.dump amlensing/q#data
dev machine: dachs dump load trans.dump