Skip to content

config.py

PipelineConfig

Pipeline run configuration.

Attributes:

Name Type Description
SCHEMA

class attribute containing the YAML schema for the run config.

TEMPLATE_PATH str

class attribute containing the path to the default Jinja2 run config template file.

epoch_based bool

boolean indicating if the original run config inputs were provided with user-defined epochs.

Raises:

Type Description
PipelineConfigError

the input YAML config violates the schema.

Source code in vast_pipeline/pipeline/config.py
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
class PipelineConfig:
    """Pipeline run configuration.

    Attributes:
        SCHEMA: class attribute containing the YAML schema for the run config.
        TEMPLATE_PATH: class attribute containing the path to the default Jinja2 run
            config template file.
        epoch_based: boolean indicating if the original run config inputs were provided
            with user-defined epochs.

    Raises:
        PipelineConfigError: the input YAML config violates the schema.
    """

    # key: config input type, value: boolean indicating if it is required
    _REQUIRED_INPUT_TYPES: Dict[str, bool] = {
        "image": True,
        "selavy": True,
        "noise": True,
        "background": False,
    }
    # Inputs may be optional. All inputs will be either a unique list or a mapping (epoch
    # mode and/or glob expressions). These possibilities cannot be validated at once, so
    # it will accept Any and then revalidate later.
    _SCHEMA_INPUTS = {
        (k if v else yaml.Optional(k)): yaml.MapPattern(yaml.Str(), yaml.Any())
        | yaml.UniqueSeq(yaml.Str())
        for k, v in _REQUIRED_INPUT_TYPES.items()
    }
    _SCHEMA_GLOB_INPUTS = {"glob": yaml.Str() | yaml.Seq(yaml.Str())}
    _VALID_ASSOC_METHODS: List[str] = ["basic", "advanced", "deruiter"]
    SCHEMA = yaml.Map(
        {
            "run": yaml.Map(
                {
                    "path": yaml.Str(),
                    "suppress_astropy_warnings": yaml.Bool(),
                }
            ),
            "inputs": yaml.Map(_SCHEMA_INPUTS),
            "source_monitoring": yaml.Map(
                {
                    "monitor": yaml.Bool(),
                    "min_sigma": yaml.Float(),
                    "edge_buffer_scale": yaml.Float(),
                    "cluster_threshold": yaml.Float(),
                    "allow_nan": yaml.Bool(),
                }
            ),
            "source_association": yaml.Map(
                {
                    "method": yaml.Enum(_VALID_ASSOC_METHODS),
                    "radius": yaml.Float(),
                    "deruiter_radius": yaml.Float(),
                    "deruiter_beamwidth_limit": yaml.Float(),
                    "parallel": yaml.Bool(),
                    "epoch_duplicate_radius": yaml.Float(),
                }
            ),
            "new_sources": yaml.Map(
                {
                    "min_sigma": yaml.Float(),
                }
            ),
            "measurements": yaml.Map(
                {
                    "source_finder": yaml.Enum(["selavy"]),
                    "flux_fractional_error": yaml.Float(),
                    "condon_errors": yaml.Bool(),
                    "selavy_local_rms_fill_value": yaml.Float(),
                    "write_arrow_files": yaml.Bool(),
                    "ra_uncertainty": yaml.Float(),
                    "dec_uncertainty": yaml.Float(),
                }
            ),
            "variability": yaml.Map(
                {
                    "source_aggregate_pair_metrics_min_abs_vs": yaml.Float(),
                }
            ),
        }
    )
    # path to default run config template
    TEMPLATE_PATH: str = os.path.join(
        settings.BASE_DIR, "vast_pipeline", "config_template.yaml.j2"
    )

    def __init__(self, config_yaml: yaml.YAML):
        """Initialises PipelineConfig with parsed (but not necessarily validated) YAML.

        Args:
            config_yaml (yaml.YAML): Input YAML, usually the output of `strictyaml.load`.

        Raises:
            PipelineConfigError: The input YAML config violates the schema.
        """
        self._yaml: yaml.YAML = config_yaml
        # The epoch_based parameter below is for if the user has entered just lists we
        # don't have access to the dates until the Image instances are created. So we
        # flag this as true so that we can reorder the epochs once the date information
        # is available. It is also recorded in the database such that there is a record
        # of the fact that the run was processed in an epoch based mode.
        self.epoch_based: bool

        # Determine if epoch-based association should be used based on input files.
        # If inputs have been parsed to dicts, then the user has defined their own epochs.
        # If inputs have been parsed to lists, we must convert to dicts and auto-fill
        # the epochs.

        # ensure the inputs are valid in case .from_file(..., validate=False) was used
        try:
            self._validate_inputs()
        except yaml.YAMLValidationError as e:
            raise PipelineConfigError(e)

        # detect simple list inputs and convert them to epoch-mode inputs
        for input_file_type in self._REQUIRED_INPUT_TYPES:
            # skip missing optional input types, e.g. background
            if (
                not self._REQUIRED_INPUT_TYPES[input_file_type]
                and input_file_type not in self["inputs"]
            ):
                continue

            input_files = self["inputs"][input_file_type]

            # resolve glob expressions if present
            if isinstance(input_files, dict):
                # must be either a glob expression, list of glob expressions, or epoch-mode
                if "glob" in input_files:
                    # resolve the glob expressions
                    self.epoch_based = False
                    file_list = self._resolve_glob_expressions(
                        self._yaml["inputs"][input_file_type]
                    )
                    self._yaml["inputs"][input_file_type] = self._create_input_epochs(
                        file_list
                    )
                else:
                    # epoch-mode with either a list of files or glob expressions
                    self.epoch_based = True
                    for epoch in input_files:
                        if "glob" in input_files[epoch]:
                            # resolve the glob expressions
                            file_list = self._resolve_glob_expressions(
                                self._yaml["inputs"][input_file_type][epoch]
                            )
                            self._yaml["inputs"][input_file_type][epoch] = file_list
            else:
                # Epoch-based association not requested and no globs present. Replace
                # input lists with dicts where each input file has it's own epoch.
                self.epoch_based = False
                self._yaml["inputs"][input_file_type] = self._create_input_epochs(
                    input_files
                )

    def __getitem__(self, name: str):
        """Retrieves the requested YAML chunk as a native Python object."""
        return self._yaml[name].data

    @staticmethod
    def _create_input_epochs(input_files: List[str]) -> Dict[str, List[str]]:
        """Convert a list of input files into a dict where each list element is placed
        into its own list of length 1 and mapped to by a unique key, a string that is a
        0-padded integer. For example, ["A", "B", "C", ..., "Z"] would be converted to
        {
            "01": ["A"],
            "02": ["B"],
            "03": ["C"],
            ...
            "26": ["Z"],
        }
        The keys are 0-padded to ensure the strings are sortable regardless of the
        length of `input_files`.
        This conversion is required for run configs that are not defined in "epoch mode"
        as after config validation, the pipeline assumes that there will be defined
        epochs.

        Args:
            input_files (List[str]): the list of input file paths.

        Returns:
            Dict[str, List[str]]: the input file paths mapped to by unique epoch keys.
        """
        pad_width = len(str(len(input_files)))
        input_files_dict = {
            f"{i + 1:0{pad_width}}": [val] for i, val in enumerate(input_files)
        }
        return input_files_dict

    @classmethod
    def from_file(
        cls,
        yaml_path: str,
        label: str = "run config",
        validate: bool = True,
        add_defaults: bool = True,
    ) -> "PipelineConfig":
        """Create a PipelineConfig object from a run configuration YAML file.

        Args:
            yaml_path: Path to the run config YAML file.
            label: A label for the config object that will be used in error messages.
                Default is "run config".
            validate: Perform config schema validation immediately after loading
                the config file. If set to False, the full schema validation
                will not be performed until PipelineConfig.validate() is
                explicitly called. The inputs are always validated regardless.
                Defaults to True.
            add_defaults: Add missing configuration parameters using configured
                defaults. The defaults are read from the Django settings file.
                Defaults to True.

        Raises:
            PipelineConfigError: The run config YAML file fails schema validation.

        """
        schema = PipelineConfig.SCHEMA if validate else yaml.Any()
        with open(yaml_path) as fh:
            config_str = fh.read()
        try:
            config_yaml = yaml.load(config_str, schema=schema, label=label)
        except yaml.YAMLValidationError as e:
            raise PipelineConfigError(e)

        if add_defaults:
            # make a template config based on defaults
            config_defaults_str = make_config_template(
                cls.TEMPLATE_PATH,
                **settings.PIPE_RUN_CONFIG_DEFAULTS,
            )
            config_defaults_dict: Dict[str, Any] = yaml.load(config_defaults_str).data

            # merge configs
            config_dict = dict_merge(config_defaults_dict, config_yaml.data)
            config_yaml = yaml.as_document(config_dict, schema=schema, label=label)
        return cls(config_yaml)

    @staticmethod
    def _resolve_glob_expressions(input_files: yaml.YAML) -> List[str]:
        """Resolve glob expressions in a YAML chunk, returning a list of sorted file
        paths.

        Args:
            input_files (yaml.YAML): A validated YAML chunk of input files that is a
                mapping of "glob" to either a single glob expression or a sequence of
                glob expressions. e.g.
                ---
                glob: /foo/*.fits
                ---
                or
                ---
                glob:
                - /foo/A/*.fits
                - /foo/B/*.fits
                ---

        Returns:
            List[str]: The resolved file paths in lexicographical order.
        """
        file_list: List[str] = []
        if input_files["glob"].is_sequence():
            for glob_expr in input_files["glob"]:
                file_list.extend(sorted(list(glob(glob_expr.data))))
        else:
            file_list.extend(sorted(list(glob(input_files["glob"].data))))
        return file_list

    def _validate_inputs(self):
        """Validate the input files. Each input type (i.e. image, selavy, noise,
        background) may be given as one of the following:
            1. A list of files.
            2. A glob expression.
            3. A list of glob expressions.
            4. A mapping of epochs to any of the above.
        Each input type is validated individually. Extra input validation steps, e.g. to
        ensure each input type has the same number of files, are performed in
        `validate()`.

        Raises:
            PipelineConfigError: The run config inputs fail schema validation.
        """
        try:
            # first pass validation
            self._yaml["inputs"].revalidate(yaml.Map(self._SCHEMA_INPUTS))

            for input_type in self._yaml["inputs"]:
                input_yaml = self._yaml["inputs"][input_type]
                if input_yaml.is_mapping():
                    # inputs are either epoch-mode, glob expressions, or both
                    if "glob" in input_yaml:
                        # validate globs
                        input_yaml.revalidate(yaml.Map(self._SCHEMA_GLOB_INPUTS))
                    else:
                        # validate epoch mode which may also contain glob expressions
                        input_yaml.revalidate(
                            yaml.MapPattern(
                                yaml.Str(),
                                yaml.UniqueSeq(yaml.Str())
                                | yaml.Map(self._SCHEMA_GLOB_INPUTS),
                            )
                        )
        except yaml.YAMLValidationError as e:
            raise PipelineConfigError(e)

    def validate(self, user: User = None):
        """Perform extra validation steps not covered by the default schema validation.
        The following checks are performed in order. If a check fails, an exception is
        raised and no further checks are performed.

        1. All input files have the same number of epochs and the same number of files
            per epoch.
        2. The number of input files does not exceed the configured pipeline maximum.
            This is only enforced if a regular user (not staff/admin) created the run.
        3. There are at least two input images.
        4. Background input images are required is source monitoring is turned on.
        5. All input files exist.

        Args:
            user: Optional. The User of the request if made through the UI. Defaults to
                None.

        Raises:
            PipelineConfigError: a validation check failed.
        """
        # run standard base schema validation
        try:
            self._yaml.revalidate(self.SCHEMA)
        except yaml.YAMLValidationError as e:
            raise PipelineConfigError(e)

        # epochs defined for images only, used as the reference list of epochs
        epochs_image = self["inputs"]["image"].keys()
        # map input type to a set of epochs
        epochs_by_input_type = {
            input_type: set(self["inputs"][input_type].keys())
            for input_type in self["inputs"].keys()
        }
        # map input type to total number of files from all epochs
        n_files_by_input_type = {}
        for input_type, epochs_set in epochs_by_input_type.items():
            n_files_by_input_type[input_type] = 0
            for epoch in epochs_set:
                n_files_by_input_type[input_type] += len(self["inputs"][input_type][epoch])
        n_files = 0  # total number of input files
        # map input type to a mapping of epoch to file count
        epoch_n_files: Dict[str, Dict[str, int]] = {}
        for input_type in self["inputs"].keys():
            epoch_n_files[input_type] = {}
            for epoch in self["inputs"][input_type].keys():
                n = len(self["inputs"][input_type][epoch])
                epoch_n_files[input_type][epoch] = n
                n_files += n

        # Note by this point the input files have been converted to a mapping regardless
        # of the user's input format.
        # Ensure all input file types have the same epochs.
        try:
            for input_type in self["inputs"].keys():
                self._yaml["inputs"][input_type].revalidate(
                    yaml.Map({epoch: yaml.Seq(yaml.Str()) for epoch in epochs_image})
                )
        except yaml.YAMLValidationError:
            # number of epochs could be different or the name of the epochs may not match
            # find out which by counting the number of unique epochs per input type
            n_epochs_per_input_type = [
                len(epochs_set) for epochs_set in epochs_by_input_type.values()
            ]
            if len(set(n_epochs_per_input_type)) > 1:
                if self.epoch_based:
                    error_msg = "The number of epochs must match for all input types.\n"
                else:
                    error_msg = "The number of files must match for all input types.\n"
            else:
                error_msg = "The name of the epochs must match for all input types.\n"
            counts_str = ""
            if self.epoch_based:
                for input_type in epoch_n_files.keys():
                    n = len(epoch_n_files[input_type])
                    counts_str += (
                        f"{input_type} has {n} epoch{'s' if n > 1 else ''}:"
                        f" {', '.join(epoch_n_files[input_type].keys())}\n"
                    )
            else:
                for input_type, n in n_files_by_input_type.items():
                    counts_str += f"{input_type} has {n} file{'s' if n > 1 else ''}\n"

            counts_str = counts_str[:-1]
            raise PipelineConfigError(error_msg + counts_str)

        # Ensure all input file type epochs have the same number of files per epoch.
        # This could be combined with the number of epochs validation above, but we want
        # to give specific feedback to the user on failure.
        try:
            for input_type in self["inputs"].keys():
                self._yaml["inputs"][input_type].revalidate(
                    yaml.Map(
                        {
                            epoch: yaml.FixedSeq(
                                [
                                    yaml.Str()
                                    for _ in range(epoch_n_files["image"][epoch])
                                ]
                            )
                            for epoch in epochs_image
                        }
                    )
                )
        except yaml.YAMLValidationError:
            # map input type to a mapping of epoch to file count
            file_counts_str = ""
            for input_type in self["inputs"].keys():
                file_counts_str += f"{input_type}:\n"
                for epoch in sorted(self["inputs"][input_type].keys()):
                    file_counts_str += (
                        f"  {epoch}: {len(self['inputs'][input_type][epoch])}\n"
                    )
            file_counts_str = file_counts_str[:-1]
            raise PipelineConfigError(
                "The number of files per epoch does not match between input types.\n"
                + file_counts_str
            )

        # ensure the number of input files is less than the user limit
        if user and n_files > settings.MAX_PIPERUN_IMAGES:
            if user.is_staff:
                logger.warning(
                    "Maximum number of images"
                    f" ({settings.MAX_PIPERUN_IMAGES}) rule bypassed with"
                    " admin status."
                )
            else:
                raise PipelineConfigError(
                    f"The number of images entered ({n_files})"
                    " exceeds the maximum number of images currently"
                    f" allowed ({settings.MAX_PIPERUN_IMAGES}). Please ask"
                    " an administrator for advice on processing your run."
                )

        # ensure at least two inputs are provided
        check = [n_files_by_input_type[input_type] < 2 for input_type in self["inputs"].keys()]
        if any(check):
            raise PipelineConfigError(
                "Number of image files must to be larger than 1"
            )

        # ensure background files are provided if source monitoring is requested
        if self["source_monitoring"]["monitor"]:
            inputs_schema = yaml.Map(
                {
                    k: yaml.UniqueSeq(yaml.Str())
                    | yaml.MapPattern(yaml.Str(), yaml.UniqueSeq(yaml.Str()))
                    for k in self._REQUIRED_INPUT_TYPES
                }
            )
            try:
                self._yaml["inputs"].revalidate(inputs_schema)
            except yaml.YAMLValidationError:
                raise PipelineConfigError(
                    "Background files must be provided if source monitoring is enabled."
                )

        # ensure the input files all exist
        for input_type in self["inputs"].keys():
            for epoch, file_list in self["inputs"][input_type].items():
                for file in file_list:
                    if not os.path.exists(file):
                        raise PipelineConfigError(f"{file} does not exist.")

    def check_prev_config_diff(self) -> bool:
        """
        Checks if the previous config file differs from the current config file. Used in
        add mode. Only returns true if the images are different and the other general
        settings are the same (the requirement for add mode). Otherwise False is returned.

        Returns:
            True if images are different but general settings are the same, otherwise
            False is returned.
        """
        prev_config = PipelineConfig.from_file(
            os.path.join(self["run"]["path"], "config_prev.yaml"),
            label="previous run config",
        )
        if self._yaml == prev_config._yaml:
            return True

        # are the input image files different?
        images_changed = self["inputs"]["image"] != prev_config["inputs"]["image"]

        # are all the non-input file configs the same?
        config_dict = self._yaml.data
        prev_config_dict = prev_config._yaml.data
        _ = config_dict.pop("inputs")
        _ = prev_config_dict.pop("inputs")
        settings_check = config_dict == prev_config_dict

        if images_changed and settings_check:
            return False
        return True

__getitem__(self, name)

Retrieves the requested YAML chunk as a native Python object.

Source code in vast_pipeline/pipeline/config.py
192
193
194
def __getitem__(self, name: str):
    """Retrieves the requested YAML chunk as a native Python object."""
    return self._yaml[name].data

__init__(self, config_yaml)

Initialises PipelineConfig with parsed (but not necessarily validated) YAML.

Parameters:

Name Type Description Default
config_yaml yaml.YAML

Input YAML, usually the output of strictyaml.load.

required

Raises:

Type Description
PipelineConfigError

The input YAML config violates the schema.

Source code in vast_pipeline/pipeline/config.py
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
def __init__(self, config_yaml: yaml.YAML):
    """Initialises PipelineConfig with parsed (but not necessarily validated) YAML.

    Args:
        config_yaml (yaml.YAML): Input YAML, usually the output of `strictyaml.load`.

    Raises:
        PipelineConfigError: The input YAML config violates the schema.
    """
    self._yaml: yaml.YAML = config_yaml
    # The epoch_based parameter below is for if the user has entered just lists we
    # don't have access to the dates until the Image instances are created. So we
    # flag this as true so that we can reorder the epochs once the date information
    # is available. It is also recorded in the database such that there is a record
    # of the fact that the run was processed in an epoch based mode.
    self.epoch_based: bool

    # Determine if epoch-based association should be used based on input files.
    # If inputs have been parsed to dicts, then the user has defined their own epochs.
    # If inputs have been parsed to lists, we must convert to dicts and auto-fill
    # the epochs.

    # ensure the inputs are valid in case .from_file(..., validate=False) was used
    try:
        self._validate_inputs()
    except yaml.YAMLValidationError as e:
        raise PipelineConfigError(e)

    # detect simple list inputs and convert them to epoch-mode inputs
    for input_file_type in self._REQUIRED_INPUT_TYPES:
        # skip missing optional input types, e.g. background
        if (
            not self._REQUIRED_INPUT_TYPES[input_file_type]
            and input_file_type not in self["inputs"]
        ):
            continue

        input_files = self["inputs"][input_file_type]

        # resolve glob expressions if present
        if isinstance(input_files, dict):
            # must be either a glob expression, list of glob expressions, or epoch-mode
            if "glob" in input_files:
                # resolve the glob expressions
                self.epoch_based = False
                file_list = self._resolve_glob_expressions(
                    self._yaml["inputs"][input_file_type]
                )
                self._yaml["inputs"][input_file_type] = self._create_input_epochs(
                    file_list
                )
            else:
                # epoch-mode with either a list of files or glob expressions
                self.epoch_based = True
                for epoch in input_files:
                    if "glob" in input_files[epoch]:
                        # resolve the glob expressions
                        file_list = self._resolve_glob_expressions(
                            self._yaml["inputs"][input_file_type][epoch]
                        )
                        self._yaml["inputs"][input_file_type][epoch] = file_list
        else:
            # Epoch-based association not requested and no globs present. Replace
            # input lists with dicts where each input file has it's own epoch.
            self.epoch_based = False
            self._yaml["inputs"][input_file_type] = self._create_input_epochs(
                input_files
            )

_create_input_epochs(input_files)

Convert a list of input files into a dict where each list element is placed into its own list of length 1 and mapped to by a unique key, a string that is a 0-padded integer. For example, ["A", "B", "C", ..., "Z"] would be converted to { "01": ["A"], "02": ["B"], "03": ["C"], ... "26": ["Z"], } The keys are 0-padded to ensure the strings are sortable regardless of the length of input_files. This conversion is required for run configs that are not defined in "epoch mode" as after config validation, the pipeline assumes that there will be defined epochs.

Parameters:

Name Type Description Default
input_files List[str]

the list of input file paths.

required

Returns:

Type Description
Dict[str, List[str]]

Dict[str, List[str]]: the input file paths mapped to by unique epoch keys.

Source code in vast_pipeline/pipeline/config.py
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
@staticmethod
def _create_input_epochs(input_files: List[str]) -> Dict[str, List[str]]:
    """Convert a list of input files into a dict where each list element is placed
    into its own list of length 1 and mapped to by a unique key, a string that is a
    0-padded integer. For example, ["A", "B", "C", ..., "Z"] would be converted to
    {
        "01": ["A"],
        "02": ["B"],
        "03": ["C"],
        ...
        "26": ["Z"],
    }
    The keys are 0-padded to ensure the strings are sortable regardless of the
    length of `input_files`.
    This conversion is required for run configs that are not defined in "epoch mode"
    as after config validation, the pipeline assumes that there will be defined
    epochs.

    Args:
        input_files (List[str]): the list of input file paths.

    Returns:
        Dict[str, List[str]]: the input file paths mapped to by unique epoch keys.
    """
    pad_width = len(str(len(input_files)))
    input_files_dict = {
        f"{i + 1:0{pad_width}}": [val] for i, val in enumerate(input_files)
    }
    return input_files_dict

_resolve_glob_expressions(input_files)

Resolve glob expressions in a YAML chunk, returning a list of sorted file paths.

Parameters:

Name Type Description Default
input_files yaml.YAML

A validated YAML chunk of input files that is a mapping of "glob" to either a single glob expression or a sequence of glob expressions. e.g.


glob: /foo/*.fits
or

glob: - /foo/A/.fits - /foo/B/.fits


required

Returns:

Type Description
List[str]

List[str]: The resolved file paths in lexicographical order.

Source code in vast_pipeline/pipeline/config.py
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
@staticmethod
def _resolve_glob_expressions(input_files: yaml.YAML) -> List[str]:
    """Resolve glob expressions in a YAML chunk, returning a list of sorted file
    paths.

    Args:
        input_files (yaml.YAML): A validated YAML chunk of input files that is a
            mapping of "glob" to either a single glob expression or a sequence of
            glob expressions. e.g.
            ---
            glob: /foo/*.fits
            ---
            or
            ---
            glob:
            - /foo/A/*.fits
            - /foo/B/*.fits
            ---

    Returns:
        List[str]: The resolved file paths in lexicographical order.
    """
    file_list: List[str] = []
    if input_files["glob"].is_sequence():
        for glob_expr in input_files["glob"]:
            file_list.extend(sorted(list(glob(glob_expr.data))))
    else:
        file_list.extend(sorted(list(glob(input_files["glob"].data))))
    return file_list

_validate_inputs(self)

Validate the input files. Each input type (i.e. image, selavy, noise, background) may be given as one of the following: 1. A list of files. 2. A glob expression. 3. A list of glob expressions. 4. A mapping of epochs to any of the above. Each input type is validated individually. Extra input validation steps, e.g. to ensure each input type has the same number of files, are performed in validate().

Raises:

Type Description
PipelineConfigError

The run config inputs fail schema validation.

Source code in vast_pipeline/pipeline/config.py
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
def _validate_inputs(self):
    """Validate the input files. Each input type (i.e. image, selavy, noise,
    background) may be given as one of the following:
        1. A list of files.
        2. A glob expression.
        3. A list of glob expressions.
        4. A mapping of epochs to any of the above.
    Each input type is validated individually. Extra input validation steps, e.g. to
    ensure each input type has the same number of files, are performed in
    `validate()`.

    Raises:
        PipelineConfigError: The run config inputs fail schema validation.
    """
    try:
        # first pass validation
        self._yaml["inputs"].revalidate(yaml.Map(self._SCHEMA_INPUTS))

        for input_type in self._yaml["inputs"]:
            input_yaml = self._yaml["inputs"][input_type]
            if input_yaml.is_mapping():
                # inputs are either epoch-mode, glob expressions, or both
                if "glob" in input_yaml:
                    # validate globs
                    input_yaml.revalidate(yaml.Map(self._SCHEMA_GLOB_INPUTS))
                else:
                    # validate epoch mode which may also contain glob expressions
                    input_yaml.revalidate(
                        yaml.MapPattern(
                            yaml.Str(),
                            yaml.UniqueSeq(yaml.Str())
                            | yaml.Map(self._SCHEMA_GLOB_INPUTS),
                        )
                    )
    except yaml.YAMLValidationError as e:
        raise PipelineConfigError(e)

check_prev_config_diff(self)

Checks if the previous config file differs from the current config file. Used in add mode. Only returns true if the images are different and the other general settings are the same (the requirement for add mode). Otherwise False is returned.

Returns:

Type Description
bool

True if images are different but general settings are the same, otherwise

bool

False is returned.

Source code in vast_pipeline/pipeline/config.py
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
def check_prev_config_diff(self) -> bool:
    """
    Checks if the previous config file differs from the current config file. Used in
    add mode. Only returns true if the images are different and the other general
    settings are the same (the requirement for add mode). Otherwise False is returned.

    Returns:
        True if images are different but general settings are the same, otherwise
        False is returned.
    """
    prev_config = PipelineConfig.from_file(
        os.path.join(self["run"]["path"], "config_prev.yaml"),
        label="previous run config",
    )
    if self._yaml == prev_config._yaml:
        return True

    # are the input image files different?
    images_changed = self["inputs"]["image"] != prev_config["inputs"]["image"]

    # are all the non-input file configs the same?
    config_dict = self._yaml.data
    prev_config_dict = prev_config._yaml.data
    _ = config_dict.pop("inputs")
    _ = prev_config_dict.pop("inputs")
    settings_check = config_dict == prev_config_dict

    if images_changed and settings_check:
        return False
    return True

from_file(cls, yaml_path, label='run config', validate=True, add_defaults=True)

Create a PipelineConfig object from a run configuration YAML file.

Parameters:

Name Type Description Default
yaml_path str

Path to the run config YAML file.

required
label str

A label for the config object that will be used in error messages. Default is "run config".

'run config'
validate bool

Perform config schema validation immediately after loading the config file. If set to False, the full schema validation will not be performed until PipelineConfig.validate() is explicitly called. The inputs are always validated regardless. Defaults to True.

True
add_defaults bool

Add missing configuration parameters using configured defaults. The defaults are read from the Django settings file. Defaults to True.

True

Raises:

Type Description
PipelineConfigError

The run config YAML file fails schema validation.

Source code in vast_pipeline/pipeline/config.py
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
@classmethod
def from_file(
    cls,
    yaml_path: str,
    label: str = "run config",
    validate: bool = True,
    add_defaults: bool = True,
) -> "PipelineConfig":
    """Create a PipelineConfig object from a run configuration YAML file.

    Args:
        yaml_path: Path to the run config YAML file.
        label: A label for the config object that will be used in error messages.
            Default is "run config".
        validate: Perform config schema validation immediately after loading
            the config file. If set to False, the full schema validation
            will not be performed until PipelineConfig.validate() is
            explicitly called. The inputs are always validated regardless.
            Defaults to True.
        add_defaults: Add missing configuration parameters using configured
            defaults. The defaults are read from the Django settings file.
            Defaults to True.

    Raises:
        PipelineConfigError: The run config YAML file fails schema validation.

    """
    schema = PipelineConfig.SCHEMA if validate else yaml.Any()
    with open(yaml_path) as fh:
        config_str = fh.read()
    try:
        config_yaml = yaml.load(config_str, schema=schema, label=label)
    except yaml.YAMLValidationError as e:
        raise PipelineConfigError(e)

    if add_defaults:
        # make a template config based on defaults
        config_defaults_str = make_config_template(
            cls.TEMPLATE_PATH,
            **settings.PIPE_RUN_CONFIG_DEFAULTS,
        )
        config_defaults_dict: Dict[str, Any] = yaml.load(config_defaults_str).data

        # merge configs
        config_dict = dict_merge(config_defaults_dict, config_yaml.data)
        config_yaml = yaml.as_document(config_dict, schema=schema, label=label)
    return cls(config_yaml)

validate(self, user=None)

Perform extra validation steps not covered by the default schema validation. The following checks are performed in order. If a check fails, an exception is raised and no further checks are performed.

  1. All input files have the same number of epochs and the same number of files per epoch.
  2. The number of input files does not exceed the configured pipeline maximum. This is only enforced if a regular user (not staff/admin) created the run.
  3. There are at least two input images.
  4. Background input images are required is source monitoring is turned on.
  5. All input files exist.

Parameters:

Name Type Description Default
user User

Optional. The User of the request if made through the UI. Defaults to None.

None

Raises:

Type Description
PipelineConfigError

a validation check failed.

Source code in vast_pipeline/pipeline/config.py
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
def validate(self, user: User = None):
    """Perform extra validation steps not covered by the default schema validation.
    The following checks are performed in order. If a check fails, an exception is
    raised and no further checks are performed.

    1. All input files have the same number of epochs and the same number of files
        per epoch.
    2. The number of input files does not exceed the configured pipeline maximum.
        This is only enforced if a regular user (not staff/admin) created the run.
    3. There are at least two input images.
    4. Background input images are required is source monitoring is turned on.
    5. All input files exist.

    Args:
        user: Optional. The User of the request if made through the UI. Defaults to
            None.

    Raises:
        PipelineConfigError: a validation check failed.
    """
    # run standard base schema validation
    try:
        self._yaml.revalidate(self.SCHEMA)
    except yaml.YAMLValidationError as e:
        raise PipelineConfigError(e)

    # epochs defined for images only, used as the reference list of epochs
    epochs_image = self["inputs"]["image"].keys()
    # map input type to a set of epochs
    epochs_by_input_type = {
        input_type: set(self["inputs"][input_type].keys())
        for input_type in self["inputs"].keys()
    }
    # map input type to total number of files from all epochs
    n_files_by_input_type = {}
    for input_type, epochs_set in epochs_by_input_type.items():
        n_files_by_input_type[input_type] = 0
        for epoch in epochs_set:
            n_files_by_input_type[input_type] += len(self["inputs"][input_type][epoch])
    n_files = 0  # total number of input files
    # map input type to a mapping of epoch to file count
    epoch_n_files: Dict[str, Dict[str, int]] = {}
    for input_type in self["inputs"].keys():
        epoch_n_files[input_type] = {}
        for epoch in self["inputs"][input_type].keys():
            n = len(self["inputs"][input_type][epoch])
            epoch_n_files[input_type][epoch] = n
            n_files += n

    # Note by this point the input files have been converted to a mapping regardless
    # of the user's input format.
    # Ensure all input file types have the same epochs.
    try:
        for input_type in self["inputs"].keys():
            self._yaml["inputs"][input_type].revalidate(
                yaml.Map({epoch: yaml.Seq(yaml.Str()) for epoch in epochs_image})
            )
    except yaml.YAMLValidationError:
        # number of epochs could be different or the name of the epochs may not match
        # find out which by counting the number of unique epochs per input type
        n_epochs_per_input_type = [
            len(epochs_set) for epochs_set in epochs_by_input_type.values()
        ]
        if len(set(n_epochs_per_input_type)) > 1:
            if self.epoch_based:
                error_msg = "The number of epochs must match for all input types.\n"
            else:
                error_msg = "The number of files must match for all input types.\n"
        else:
            error_msg = "The name of the epochs must match for all input types.\n"
        counts_str = ""
        if self.epoch_based:
            for input_type in epoch_n_files.keys():
                n = len(epoch_n_files[input_type])
                counts_str += (
                    f"{input_type} has {n} epoch{'s' if n > 1 else ''}:"
                    f" {', '.join(epoch_n_files[input_type].keys())}\n"
                )
        else:
            for input_type, n in n_files_by_input_type.items():
                counts_str += f"{input_type} has {n} file{'s' if n > 1 else ''}\n"

        counts_str = counts_str[:-1]
        raise PipelineConfigError(error_msg + counts_str)

    # Ensure all input file type epochs have the same number of files per epoch.
    # This could be combined with the number of epochs validation above, but we want
    # to give specific feedback to the user on failure.
    try:
        for input_type in self["inputs"].keys():
            self._yaml["inputs"][input_type].revalidate(
                yaml.Map(
                    {
                        epoch: yaml.FixedSeq(
                            [
                                yaml.Str()
                                for _ in range(epoch_n_files["image"][epoch])
                            ]
                        )
                        for epoch in epochs_image
                    }
                )
            )
    except yaml.YAMLValidationError:
        # map input type to a mapping of epoch to file count
        file_counts_str = ""
        for input_type in self["inputs"].keys():
            file_counts_str += f"{input_type}:\n"
            for epoch in sorted(self["inputs"][input_type].keys()):
                file_counts_str += (
                    f"  {epoch}: {len(self['inputs'][input_type][epoch])}\n"
                )
        file_counts_str = file_counts_str[:-1]
        raise PipelineConfigError(
            "The number of files per epoch does not match between input types.\n"
            + file_counts_str
        )

    # ensure the number of input files is less than the user limit
    if user and n_files > settings.MAX_PIPERUN_IMAGES:
        if user.is_staff:
            logger.warning(
                "Maximum number of images"
                f" ({settings.MAX_PIPERUN_IMAGES}) rule bypassed with"
                " admin status."
            )
        else:
            raise PipelineConfigError(
                f"The number of images entered ({n_files})"
                " exceeds the maximum number of images currently"
                f" allowed ({settings.MAX_PIPERUN_IMAGES}). Please ask"
                " an administrator for advice on processing your run."
            )

    # ensure at least two inputs are provided
    check = [n_files_by_input_type[input_type] < 2 for input_type in self["inputs"].keys()]
    if any(check):
        raise PipelineConfigError(
            "Number of image files must to be larger than 1"
        )

    # ensure background files are provided if source monitoring is requested
    if self["source_monitoring"]["monitor"]:
        inputs_schema = yaml.Map(
            {
                k: yaml.UniqueSeq(yaml.Str())
                | yaml.MapPattern(yaml.Str(), yaml.UniqueSeq(yaml.Str()))
                for k in self._REQUIRED_INPUT_TYPES
            }
        )
        try:
            self._yaml["inputs"].revalidate(inputs_schema)
        except yaml.YAMLValidationError:
            raise PipelineConfigError(
                "Background files must be provided if source monitoring is enabled."
            )

    # ensure the input files all exist
    for input_type in self["inputs"].keys():
        for epoch, file_list in self["inputs"][input_type].items():
            for file in file_list:
                if not os.path.exists(file):
                    raise PipelineConfigError(f"{file} does not exist.")

make_config_template(template_path, **kwargs)

Generate the contents of a run configuration file from on a Jinja2 template.

Parameters:

Name Type Description Default
template_path str

Path to a Jinja2 template.

required
**kwargs

keyword arguments passed to the template renderer to fill in template variables.

{}

Returns:

Type Description
str

Filled in template string.

Source code in vast_pipeline/pipeline/config.py
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
def make_config_template(template_path: str, **kwargs) -> str:
    """Generate the contents of a run configuration file from on a Jinja2 template.

    Args:
        template_path: Path to a Jinja2 template.
        **kwargs: keyword arguments passed to the template renderer to fill in template
            variables.

    Returns:
        Filled in template string.
    """
    with open(template_path, "r") as fp:
        template_str = fp.read()
    env = Environment(trim_blocks=True, lstrip_blocks=True)
    template = env.from_string(template_str)
    return template.render(**kwargs)

Last update: March 2, 2022
Created: March 2, 2022