Coverage for python/lsst/meas/astrom/matcher_probabilistic.py: 25%
205 statements
« prev ^ index » next coverage.py v7.2.2, created at 2023-03-23 23:46 +0000
1# This file is part of meas_astrom.
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (https://www.lsst.org).
6# See the COPYRIGHT file at the top-level directory of this distribution
7# for details of code ownership.
8#
9# This program is free software: you can redistribute it and/or modify
10# it under the terms of the GNU General Public License as published by
11# the Free Software Foundation, either version 3 of the License, or
12# (at your option) any later version.
13#
14# This program is distributed in the hope that it will be useful,
15# but WITHOUT ANY WARRANTY; without even the implied warranty of
16# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17# GNU General Public License for more details.
18#
19# You should have received a copy of the GNU General Public License
20# along with this program. If not, see <https://www.gnu.org/licenses/>.
22__all__ = ['ConvertCatalogCoordinatesConfig', 'MatchProbabilisticConfig', 'MatcherProbabilistic']
24import lsst.pex.config as pexConfig
26from dataclasses import dataclass
27import logging
28import numpy as np
29import pandas as pd
30from scipy.spatial import cKDTree
31import time
32from typing import Callable, Set
34logger_default = logging.getLogger(__name__)
37def _mul_column(column: np.array, value: float):
38 if value is not None and value != 1:
39 column *= value
40 return column
43def _radec_to_xyz(ra, dec):
44 """Convert input ra/dec coordinates to spherical unit vectors.
46 Parameters
47 ----------
48 ra, dec: `numpy.ndarray`
49 Arrays of right ascension/declination in degrees.
51 Returns
52 -------
53 vectors : `numpy.ndarray`, (N, 3)
54 Output unit vectors.
55 """
56 if ra.size != dec.size:
57 raise ValueError('ra and dec must be same size')
58 ras = np.radians(ra)
59 decs = np.radians(dec)
60 vectors = np.empty((ras.size, 3))
62 sin_dec = np.sin(np.pi / 2 - decs)
63 vectors[:, 0] = sin_dec * np.cos(ras)
64 vectors[:, 1] = sin_dec * np.sin(ras)
65 vectors[:, 2] = np.cos(np.pi / 2 - decs)
67 return vectors
70@dataclass
71class CatalogExtras:
72 """Store frequently-reference (meta)data relevant for matching a catalog.
74 Parameters
75 ----------
76 catalog : `pandas.DataFrame`
77 A pandas catalog to store extra information for.
78 select : `numpy.array`
79 A numpy boolean array of the same length as catalog to be used for
80 target selection.
81 """
82 n: int
83 indices: np.array
84 select: np.array
86 coordinate_factor: float = None
88 def __init__(self, catalog: pd.DataFrame, select: np.array = None, coordinate_factor: float = None):
89 self.n = len(catalog)
90 self.select = np.ones(self.n, dtype=bool) if select is None else select
91 self.indices = np.flatnonzero(select) if select is not None else np.arange(self.n)
92 self.coordinate_factor = coordinate_factor
95@dataclass(frozen=True)
96class ComparableCatalog:
97 """A catalog with sources with coordinate columns in some standard format/units.
99 catalog : `pandas.DataFrame`
100 A catalog with comparable coordinate columns.
101 column_coord1 : `str`
102 The first spatial coordinate column name.
103 column_coord2 : `str`
104 The second spatial coordinate column name.
105 coord1 : `numpy.array`
106 The first spatial coordinate values.
107 coord2 : `numpy.array`
108 The second spatial coordinate values.
109 extras : `CatalogExtras`
110 Extra cached (meta)data for the `catalog`.
111 """
112 catalog: pd.DataFrame
113 column_coord1: str
114 column_coord2: str
115 coord1: np.array
116 coord2: np.array
117 extras: CatalogExtras
120class ConvertCatalogCoordinatesConfig(pexConfig.Config):
121 """Configuration for the MatchProbabilistic matcher.
122 """
123 column_ref_coord1 = pexConfig.Field(
124 dtype=str,
125 default='ra',
126 doc='The reference table column for the first spatial coordinate (usually x or ra).',
127 )
128 column_ref_coord2 = pexConfig.Field(
129 dtype=str,
130 default='dec',
131 doc='The reference table column for the second spatial coordinate (usually y or dec).'
132 'Units must match column_ref_coord1.',
133 )
134 column_target_coord1 = pexConfig.Field(
135 dtype=str,
136 default='coord_ra',
137 doc='The target table column for the first spatial coordinate (usually x or ra).'
138 'Units must match column_ref_coord1.',
139 )
140 column_target_coord2 = pexConfig.Field(
141 dtype=str,
142 default='coord_dec',
143 doc='The target table column for the second spatial coordinate (usually y or dec).'
144 'Units must match column_ref_coord2.',
145 )
146 coords_spherical = pexConfig.Field(
147 dtype=bool,
148 default=True,
149 doc='Whether column_*_coord[12] are spherical coordinates (ra/dec) or not (pixel x/y)',
150 )
151 coords_ref_factor = pexConfig.Field(
152 dtype=float,
153 default=1.0,
154 doc='Multiplicative factor for reference catalog coordinates.'
155 'If coords_spherical is true, this must be the number of degrees per unit increment of '
156 'column_ref_coord[12]. Otherwise, it must convert the coordinate to the same units'
157 ' as the target coordinates.',
158 )
159 coords_target_factor = pexConfig.Field(
160 dtype=float,
161 default=1.0,
162 doc='Multiplicative factor for target catalog coordinates.'
163 'If coords_spherical is true, this must be the number of degrees per unit increment of '
164 'column_target_coord[12]. Otherwise, it must convert the coordinate to the same units'
165 ' as the reference coordinates.',
166 )
167 coords_ref_to_convert = pexConfig.DictField( 167 ↛ exitline 167 didn't jump to the function exit
168 default=None,
169 keytype=str,
170 itemtype=str,
171 dictCheck=lambda x: len(x) == 2,
172 doc='Dict mapping sky coordinate columns to be converted to pixel columns',
173 )
174 mag_zeropoint_ref = pexConfig.Field(
175 dtype=float,
176 default=31.4,
177 doc='Magnitude zeropoint for reference catalog.',
178 )
180 def format_catalogs(
181 self,
182 catalog_ref: pd.DataFrame,
183 catalog_target: pd.DataFrame,
184 select_ref: np.array = None,
185 select_target: np.array = None,
186 radec_to_xy_func: Callable = None,
187 return_converted_columns: bool = False,
188 **kwargs,
189 ):
190 """Format matched catalogs that may require coordinate conversions.
192 Parameters
193 ----------
194 catalog_ref : `pandas.DataFrame`
195 A reference catalog for comparison to `catalog_target`.
196 catalog_target : `pandas.DataFrame`
197 A target catalog with measurements for comparison to `catalog_ref`.
198 select_ref : `numpy.ndarray`, (Nref,)
199 A boolean array of len `catalog_ref`, True for valid match candidates.
200 select_target : `numpy.ndarray`, (Ntarget,)
201 A boolean array of len `catalog_target`, True for valid match candidates.
202 radec_to_xy_func : `typing.Callable`
203 Function taking equal-length ra, dec arrays and returning an ndarray of
204 - ``x``: current parameter (`float`).
205 - ``extra_args``: additional arguments (`dict`).
206 return_converted_columns : `bool`
207 Whether to return converted columns in the `coord1` and `coord2`
208 attributes, rather than keep the original values.
209 kwargs
211 Returns
212 -------
213 compcat_ref, compcat_target : `ComparableCatalog`
214 Comparable catalogs corresponding to the input reference and target.
216 """
217 convert_ref = self.coords_ref_to_convert
218 if convert_ref and not callable(radec_to_xy_func):
219 raise TypeError('radec_to_xy_func must be callable if converting ref coords')
221 # Set up objects with frequently-used attributes like selection bool array
222 extras_ref, extras_target = (
223 CatalogExtras(catalog, select=select, coordinate_factor=coord_factor)
224 for catalog, select, coord_factor in zip(
225 (catalog_ref, catalog_target),
226 (select_ref, select_target),
227 (self.coords_ref_factor, self.coords_target_factor),
228 )
229 )
231 compcats = []
233 # Retrieve coordinates and multiply them by scaling factors
234 for catalog, extras, (column1, column2), convert in (
235 (catalog_ref, extras_ref, (self.column_ref_coord1, self.column_ref_coord2), convert_ref),
236 (catalog_target, extras_target, (self.column_target_coord1, self.column_target_coord2), False),
237 ):
238 coord1, coord2 = (
239 _mul_column(catalog[column], extras.coordinate_factor)
240 for column in (column1, column2)
241 )
242 if convert:
243 xy_ref = radec_to_xy_func(coord1, coord2, self.coords_ref_factor, **kwargs)
244 for idx_coord, column_out in enumerate(self.coords_ref_to_convert.values()):
245 coord = np.array([xy[idx_coord] for xy in xy_ref])
246 catalog[column_out] = coord
247 if convert_ref and return_converted_columns:
248 column1, column2 = self.coords_ref_to_convert.values()
249 coord1, coord2 = catalog[column1], catalog[column2]
250 if isinstance(coord1, pd.Series):
251 coord1 = coord1.values
252 if isinstance(coord2, pd.Series):
253 coord2 = coord2.values
255 compcats.append(ComparableCatalog(
256 catalog=catalog, column_coord1=column1, column_coord2=column2,
257 coord1=coord1, coord2=coord2, extras=extras,
258 ))
260 return tuple(compcats)
263class MatchProbabilisticConfig(pexConfig.Config):
264 """Configuration for the MatchProbabilistic matcher.
265 """
266 column_ref_order = pexConfig.Field(
267 dtype=str,
268 default=None,
269 optional=True,
270 doc="Name of column in reference catalog specifying order for matching."
271 " Derived from columns_ref_flux if not set.",
272 )
274 @property
275 def columns_in_ref(self) -> Set[str]:
276 columns_all = [
277 self.coord_format.column_ref_coord1,
278 self.coord_format.column_ref_coord2,
279 ]
280 for columns in (
281 self.columns_ref_flux,
282 self.columns_ref_meas,
283 self.columns_ref_copy,
284 ):
285 columns_all.extend(columns)
287 return set(columns_all)
289 @property
290 def columns_in_target(self) -> Set[str]:
291 columns_all = [
292 self.coord_format.column_target_coord1,
293 self.coord_format.column_target_coord2,
294 ]
295 for columns in (
296 self.columns_target_meas,
297 self.columns_target_err,
298 self.columns_target_select_false,
299 self.columns_target_select_true,
300 self.columns_target_copy,
301 ):
302 columns_all.extend(columns)
303 return set(columns_all)
305 columns_ref_copy = pexConfig.ListField( 305 ↛ exitline 305 didn't jump to the function exit
306 dtype=str,
307 default=[],
308 listCheck=lambda x: len(set(x)) == len(x),
309 optional=True,
310 doc='Reference table columns to copy unchanged into both match tables',
311 )
312 columns_ref_flux = pexConfig.ListField( 312 ↛ exitline 312 didn't jump to the function exit
313 dtype=str,
314 default=[],
315 listCheck=lambda x: len(set(x)) == len(x),
316 optional=True,
317 doc="List of reference flux columns to nansum total magnitudes from if column_order is None",
318 )
319 columns_ref_meas = pexConfig.ListField(
320 dtype=str,
321 doc='The reference table columns to compute match likelihoods from '
322 '(usually centroids and fluxes/magnitudes)',
323 )
324 columns_target_copy = pexConfig.ListField( 324 ↛ exitline 324 didn't jump to the function exit
325 dtype=str,
326 default=[],
327 listCheck=lambda x: len(set(x)) == len(x),
328 optional=True,
329 doc='Target table columns to copy unchanged into both match tables',
330 )
331 columns_target_meas = pexConfig.ListField(
332 dtype=str,
333 doc='Target table columns with measurements corresponding to columns_ref_meas',
334 )
335 columns_target_err = pexConfig.ListField(
336 dtype=str,
337 doc='Target table columns with standard errors (sigma) corresponding to columns_ref_meas',
338 )
339 columns_target_select_true = pexConfig.ListField(
340 dtype=str,
341 default=('detect_isPrimary',),
342 doc='Target table columns to require to be True for selecting sources',
343 )
344 columns_target_select_false = pexConfig.ListField(
345 dtype=str,
346 default=('merge_peak_sky',),
347 doc='Target table columns to require to be False for selecting sources',
348 )
349 coord_format = pexConfig.ConfigField(
350 dtype=ConvertCatalogCoordinatesConfig,
351 doc="Configuration for coordinate conversion",
352 )
353 mag_brightest_ref = pexConfig.Field(
354 dtype=float,
355 default=-np.inf,
356 doc='Bright magnitude cutoff for selecting reference sources to match.'
357 ' Ignored if column_ref_order is None.'
358 )
359 mag_faintest_ref = pexConfig.Field(
360 dtype=float,
361 default=np.Inf,
362 doc='Faint magnitude cutoff for selecting reference sources to match.'
363 ' Ignored if column_ref_order is None.'
364 )
365 match_dist_max = pexConfig.Field(
366 dtype=float,
367 default=0.5,
368 doc='Maximum match distance. Units must be arcseconds if coords_spherical, '
369 'or else match those of column_*_coord[12] multiplied by coords_*_factor.',
370 )
371 match_n_max = pexConfig.Field(
372 dtype=int,
373 default=10,
374 optional=True,
375 doc='Maximum number of spatial matches to consider (in ascending distance order).',
376 )
377 match_n_finite_min = pexConfig.Field(
378 dtype=int,
379 default=3,
380 optional=True,
381 doc='Minimum number of columns with a finite value to measure match likelihood',
382 )
383 order_ascending = pexConfig.Field(
384 dtype=bool,
385 default=False,
386 optional=True,
387 doc='Whether to order reference match candidates in ascending order of column_ref_order '
388 '(should be False if the column is a flux and True if it is a magnitude.',
389 )
392def default_value(dtype):
393 if dtype == str:
394 return ''
395 elif dtype == np.signedinteger:
396 return np.Inf
397 elif dtype == np.unsignedinteger:
398 return -np.Inf
399 return None
402class MatcherProbabilistic:
403 """A probabilistic, greedy catalog matcher.
405 Parameters
406 ----------
407 config: `MatchProbabilisticConfig`
408 A configuration instance.
409 """
410 config: MatchProbabilisticConfig
412 def __init__(
413 self,
414 config: MatchProbabilisticConfig,
415 ):
416 self.config = config
418 def match(
419 self,
420 catalog_ref: pd.DataFrame,
421 catalog_target: pd.DataFrame,
422 select_ref: np.array = None,
423 select_target: np.array = None,
424 logger: logging.Logger = None,
425 logging_n_rows: int = None,
426 **kwargs
427 ):
428 """Match catalogs.
430 Parameters
431 ----------
432 catalog_ref : `pandas.DataFrame`
433 A reference catalog to match in order of a given column (i.e. greedily).
434 catalog_target : `pandas.DataFrame`
435 A target catalog for matching sources from `catalog_ref`. Must contain measurements with errors.
436 select_ref : `numpy.array`
437 A boolean array of the same length as `catalog_ref` selecting the sources that can be matched.
438 select_target : `numpy.array`
439 A boolean array of the same length as `catalog_target` selecting the sources that can be matched.
440 logger : `logging.Logger`
441 A Logger for logging.
442 logging_n_rows : `int`
443 The number of sources to match before printing a log message.
444 kwargs
445 Additional keyword arguments to pass to `format_catalogs`.
447 Returns
448 -------
449 catalog_out_ref : `pandas.DataFrame`
450 A catalog of identical length to `catalog_ref`, containing match information for rows selected by
451 `select_ref` (including the matching row index in `catalog_target`).
452 catalog_out_target : `pandas.DataFrame`
453 A catalog of identical length to `catalog_target`, containing the indices of matching rows in
454 `catalog_ref`.
455 exceptions : `dict` [`int`, `Exception`]
456 A dictionary keyed by `catalog_target` row number of the first exception caught when matching.
457 """
458 if logger is None:
459 logger = logger_default
461 config = self.config
463 # Transform any coordinates, if required
464 # Note: The returned objects contain the original catalogs, as well as
465 # transformed coordinates, and the selection of sources for matching.
466 # These might be identical to the arrays passed as kwargs, but that
467 # depends on config settings.
468 # For the rest of this function, the selection arrays will be used,
469 # but the indices of the original, unfiltered catalog will also be
470 # output, so some further indexing steps are needed.
471 ref, target = config.coord_format.format_catalogs(
472 catalog_ref=catalog_ref, catalog_target=catalog_target,
473 select_ref=select_ref, select_target=select_target,
474 **kwargs
475 )
477 # If no order is specified, take nansum of all flux columns for a 'total flux'
478 # Note: it won't actually be a total flux if bands overlap significantly
479 # (or it might define a filter with >100% efficiency
480 # Also, this is done on the original dataframe as it's harder to accomplish
481 # just with a recarray
482 column_order = (
483 catalog_ref.loc[ref.extras.select, config.column_ref_order]
484 if config.column_ref_order is not None else
485 np.nansum(catalog_ref.loc[ref.extras.select, config.columns_ref_flux], axis=1)
486 )
487 order = np.argsort(column_order if config.order_ascending else -column_order)
489 n_ref_select = len(ref.extras.indices)
491 match_dist_max = config.match_dist_max
492 coords_spherical = config.coord_format.coords_spherical
493 if coords_spherical:
494 match_dist_max = np.radians(match_dist_max / 3600.)
496 # Convert ra/dec sky coordinates to spherical vectors for accurate distances
497 func_convert = _radec_to_xyz if coords_spherical else np.vstack
498 vec_ref, vec_target = (
499 func_convert(cat.coord1[cat.extras.select], cat.coord2[cat.extras.select])
500 for cat in (ref, target)
501 )
503 # Generate K-d tree to compute distances
504 logger.info('Generating cKDTree with match_n_max=%d', config.match_n_max)
505 tree_obj = cKDTree(vec_target)
507 scores, idxs_target_select = tree_obj.query(
508 vec_ref,
509 distance_upper_bound=match_dist_max,
510 k=config.match_n_max,
511 )
513 n_target_select = len(target.extras.indices)
514 n_matches = np.sum(idxs_target_select != n_target_select, axis=1)
515 n_matched_max = np.sum(n_matches == config.match_n_max)
516 if n_matched_max > 0:
517 logger.warning(
518 '%d/%d (%.2f%%) selected true objects have n_matches=n_match_max(%d)',
519 n_matched_max, n_ref_select, 100.*n_matched_max/n_ref_select, config.match_n_max
520 )
522 # Pre-allocate outputs
523 target_row_match = np.full(target.extras.n, np.nan, dtype=np.int64)
524 ref_candidate_match = np.zeros(ref.extras.n, dtype=bool)
525 ref_row_match = np.full(ref.extras.n, np.nan, dtype=np.int64)
526 ref_match_count = np.zeros(ref.extras.n, dtype=np.int32)
527 ref_match_meas_finite = np.zeros(ref.extras.n, dtype=np.int32)
528 ref_chisq = np.full(ref.extras.n, np.nan, dtype=float)
530 # Need the original reference row indices for output
531 idx_orig_ref, idx_orig_target = (np.argwhere(cat.extras.select) for cat in (ref, target))
533 # Retrieve required columns, including any converted ones (default to original column name)
534 columns_convert = config.coord_format.coords_ref_to_convert
535 if columns_convert is None:
536 columns_convert = {}
537 data_ref = ref.catalog[
538 [columns_convert.get(column, column) for column in config.columns_ref_meas]
539 ].iloc[ref.extras.indices[order]]
540 data_target = target.catalog[config.columns_target_meas][target.extras.select]
541 errors_target = target.catalog[config.columns_target_err][target.extras.select]
543 exceptions = {}
544 # The kdTree uses len(inputs) as a sentinel value for no match
545 matched_target = {n_target_select, }
547 t_begin = time.process_time()
549 logger.info('Matching n_indices=%d/%d', len(order), len(ref.catalog))
550 for index_n, index_row_select in enumerate(order):
551 index_row = idx_orig_ref[index_row_select]
552 ref_candidate_match[index_row] = True
553 found = idxs_target_select[index_row_select, :]
554 # Select match candidates from nearby sources not already matched
555 # Note: set lookup is apparently fast enough that this is a few percent faster than:
556 # found = [x for x in found[found != n_target_select] if x not in matched_target]
557 # ... at least for ~1M sources
558 found = [x for x in found if x not in matched_target]
559 n_found = len(found)
560 if n_found > 0:
561 # This is an ndarray of n_found rows x len(data_ref/target) columns
562 chi = (
563 (data_target.iloc[found].values - data_ref.iloc[index_n].values)
564 / errors_target.iloc[found].values
565 )
566 finite = np.isfinite(chi)
567 n_finite = np.sum(finite, axis=1)
568 # Require some number of finite chi_sq to match
569 chisq_good = n_finite >= config.match_n_finite_min
570 if np.any(chisq_good):
571 try:
572 chisq_sum = np.zeros(n_found, dtype=float)
573 chisq_sum[chisq_good] = np.nansum(chi[chisq_good, :] ** 2, axis=1)
574 idx_chisq_min = np.nanargmin(chisq_sum / n_finite)
575 ref_match_meas_finite[index_row] = n_finite[idx_chisq_min]
576 ref_match_count[index_row] = len(chisq_good)
577 ref_chisq[index_row] = chisq_sum[idx_chisq_min]
578 idx_match_select = found[idx_chisq_min]
579 row_target = target.extras.indices[idx_match_select]
580 ref_row_match[index_row] = row_target
582 target_row_match[row_target] = index_row
583 matched_target.add(idx_match_select)
584 except Exception as error:
585 # Can't foresee any exceptions, but they shouldn't prevent
586 # matching subsequent sources
587 exceptions[index_row] = error
589 if logging_n_rows and ((index_n + 1) % logging_n_rows == 0):
590 t_elapsed = time.process_time() - t_begin
591 logger.info(
592 'Processed %d/%d in %.2fs at sort value=%.3f',
593 index_n + 1, n_ref_select, t_elapsed, column_order[order[index_n]],
594 )
596 data_ref = {
597 'match_candidate': ref_candidate_match,
598 'match_row': ref_row_match,
599 'match_count': ref_match_count,
600 'match_chisq': ref_chisq,
601 'match_n_chisq_finite': ref_match_meas_finite,
602 }
603 data_target = {
604 'match_candidate': target.extras.select if target.extras.select is not None else (
605 np.ones(target.extras.n, dtype=bool)),
606 'match_row': target_row_match,
607 }
609 for (columns, out_original, out_matched, in_original, in_matched, matches) in (
610 (
611 self.config.columns_ref_copy,
612 data_ref,
613 data_target,
614 ref,
615 target,
616 target_row_match,
617 ),
618 (
619 self.config.columns_target_copy,
620 data_target,
621 data_ref,
622 target,
623 ref,
624 ref_row_match,
625 ),
626 ):
627 matched = matches >= 0
628 idx_matched = matches[matched]
630 for column in columns:
631 values = in_original.catalog[column]
632 out_original[column] = values
633 dtype = in_original.catalog[column].dtype
634 if dtype == object:
635 types = list(set((type(x) for x in values)))
636 if len(types) != 1:
637 raise RuntimeError(f'Column {column} dtype={dtype} has multiple types={types}')
638 dtype = types[0]
640 column_match = np.full(in_matched.extras.n, default_value(dtype), dtype=dtype)
641 column_match[matched] = in_original.catalog[column][idx_matched]
642 out_matched[f'match_{column}'] = column_match
644 catalog_out_ref = pd.DataFrame(data_ref)
645 catalog_out_target = pd.DataFrame(data_target)
647 return catalog_out_ref, catalog_out_target, exceptions