endoreg-db 0.8.6.4__py3-none-any.whl

endoreg_db/models/medical/patient/patient_finding.py

@@ -0,0 +1,357 @@
+from django.db import models
+from django.core.exceptions import ValidationError
+from django.utils import timezone
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from endoreg_db.models import (
+        PatientExamination,
+        Finding,
+        PatientFindingIntervention,
+        PatientFindingClassification,
+        LabelVideoSegment,
+    )
+    from endoreg_db.utils.links.requirement_link import RequirementLinks
+    
+class PatientFinding(models.Model):
+    patient_examination = models.ForeignKey('PatientExamination', on_delete=models.CASCADE, related_name='patient_findings')  # type: ignore[assignment]
+    finding = models.ForeignKey('Finding', on_delete=models.CASCADE, related_name='finding_patient_findings')  # type: ignore[assignment]
+    
+    # Audit-Felder für medizinische Nachverfolgung
+    created_at = models.DateTimeField(auto_now_add=True)
+    updated_at = models.DateTimeField(auto_now=True)
+    created_by = models.ForeignKey('auth.User', on_delete=models.PROTECT, related_name='created_findings', null=True, blank=True)
+    updated_by = models.ForeignKey('auth.User', on_delete=models.PROTECT, related_name='updated_findings', null=True, blank=True)
+    
+    # Soft Delete für historische Daten
+    is_active = models.BooleanField(default=True, help_text="Deaktiviert statt gelöscht für Audit-Trail")
+    deactivated_at = models.DateTimeField(null=True, blank=True)
+    deactivated_by = models.ForeignKey('auth.User', on_delete=models.PROTECT, related_name='deactivated_findings', null=True, blank=True)
+    
+    if TYPE_CHECKING:
+        patient_examination: "PatientExamination"
+        finding: "Finding"
+        classifications: models.QuerySet["PatientFindingClassification"]
+        interventions: models.QuerySet["PatientFindingIntervention"]
+        video_segments: models.QuerySet["LabelVideoSegment"]
+
+    class Meta:
+        verbose_name = 'Patient Finding'
+        verbose_name_plural = 'Patient Findings'
+        ordering = ['patient_examination', 'finding']
+        
+        # Wichtige Constraints für Datenintegrität
+        constraints = [
+            models.UniqueConstraint(
+                fields=['patient_examination', 'finding'],
+                condition=models.Q(is_active=True),
+                name='unique_active_finding_per_examination'
+            ),
+            models.CheckConstraint(
+                check=models.Q(
+                    deactivated_at__isnull=True,
+                    deactivated_by__isnull=True
+                ) | models.Q(
+                    deactivated_at__isnull=False,
+                    deactivated_by__isnull=False,
+                    is_active=False
+                ),
+                name='deactivation_fields_consistency'
+            )
+        ]
+        
+        # Performance-optimierte Indizes
+        indexes = [
+            models.Index(fields=['patient_examination', 'finding']),
+            models.Index(fields=['patient_examination', 'is_active']),
+            models.Index(fields=['created_at']),
+            models.Index(fields=['finding', 'is_active']),
+        ]
+
+    @property
+    def patient(self):
+        """Returns the patient associated with this patient finding."""
+        return self.patient_examination.patient
+
+    def __str__(self):
+        status = " (deaktiviert)" if not self.is_active else ""
+        return f"{self.patient_examination} - {self.finding}{status}"
+
+    def clean(self):
+        """
+        Validates the patient finding against business rules before saving.
+        
+        Ensures that the selected finding is allowed for the associated examination and that all required findings are present. Raises a ValidationError if the finding is not permitted or required findings are missing.
+        """
+        super().clean()
+        
+        # Prüfe ob Finding für diese Examination erlaubt ist
+        if self.finding and self.patient_examination:
+            available_findings = self.patient_examination.examination.get_available_findings()
+            if self.finding not in available_findings:
+                raise ValidationError({
+                    'finding': f'Finding "{self.finding.name}" ist nicht für Examination "{self.patient_examination.examination.name}" erlaubt.'
+                })
+        
+        # Prüfe Required Findings Logic
+        if self.finding and self.patient_examination:
+            self._validate_required_findings()
+    
+    # This avoids validation errors on partial updates 
+    def save(self, *args, **kwargs):
+        """
+        Validates the model instance before saving, unless performing a partial update.
+        
+        Performs full model validation with `full_clean()` before saving, except when `update_fields` is specified for a partial update.
+        """
+        if not kwargs.get('update_fields'):
+            self.full_clean()
+        super().save(*args, **kwargs)
+
+    
+    def _validate_required_findings(self):
+        """Validiert Required vs Optional Finding Constraints"""
+        examination = self.patient_examination.examination
+        
+        # Hole Required Findings für diese Examination
+        required_findings = getattr(examination, 'required_findings', None)
+        if required_findings and required_findings.exists():
+            # Prüfe ob alle Required Findings vorhanden sind
+            existing_findings = self.patient_examination.patient_findings.filter(
+                is_active=True
+            ).values_list('finding', flat=True)
+            
+            missing_required = required_findings.exclude(id__in=existing_findings)
+            if missing_required.exists() and self.finding not in required_findings.all():
+                missing_names = ', '.join([f.name for f in missing_required])
+                raise ValidationError(
+                    f'Erforderliche Findings fehlen: {missing_names}'
+                )
+
+    def deactivate(self, user=None, reason=None):
+        """Soft Delete mit Audit-Trail"""
+        self.is_active = False
+        self.deactivated_at = timezone.now()
+        self.deactivated_by = user
+        self.save(update_fields=['is_active', 'deactivated_at', 'deactivated_by'])
+        
+        # Deaktiviere auch abhängige Objekte
+        self.locations.update(is_active=False, deactivated_at=timezone.now())
+        self.morphologies.update(is_active=False, deactivated_at=timezone.now())
+        self.interventions.update(is_active=False, deactivated_at=timezone.now())
+
+    def reactivate(self, user=None):
+        """
+        Reactivates a previously deactivated patient finding after validating its state.
+        
+        If validation passes, sets the finding as active, clears deactivation fields, updates the user, and saves changes. Raises a ValidationError if reactivation is not allowed.
+        """
+        if not self.is_active:
+            # Prüfe ob Reaktivierung erlaubt ist
+            try:
+                self.clean()
+                self.is_active = True
+                self.deactivated_at = None
+                self.deactivated_by = None
+                self.updated_by = user
+                self.save(update_fields=['is_active', 'deactivated_at', 'deactivated_by', 'updated_by'])
+            except ValidationError as e:
+                raise ValidationError(f'Reaktivierung nicht möglich: {e}')
+
+
+    def get_interventions(self):
+        """
+        Retrieve all active interventions associated with this patient finding.
+        
+        Returns:
+            QuerySet: Active related interventions with intervention details prefetched.
+        """
+        return self.interventions.filter(is_active=True).select_related('intervention')
+
+    def add_classification(
+            self,
+            classification_id,
+            classification_choice_id,
+            user=None,
+        ) -> "PatientFindingClassification":
+        """
+            Add a classification choice to this patient finding after validating its association.
+            
+            Parameters:
+            	classification_id: The ID of the classification to add.
+            	classification_choice_id: The ID of the classification choice to associate.
+            	user: Optional user performing the action.
+            
+            Returns:
+            	PatientFindingClassification: The created or existing active classification association.
+            
+            Raises:
+            	ValidationError: If the classification does not exist or the choice is not valid for the classification.
+            """
+        from .patient_finding_classification import PatientFindingClassification
+        from ..finding import FindingClassification, FindingClassificationChoice
+
+        try:
+            classification = FindingClassification.objects.get(id=classification_id)
+            classification_choice = FindingClassificationChoice.objects.filter(id=classification_choice_id).first()
+
+            if not classification.choices.filter(id=classification_choice_id).exists():
+                raise ValidationError(
+                    f'Classification Choice {classification_choice_id} gehört nicht zu Classification {classification_id}'
+                )
+
+            existing = self.classifications.filter(
+                classification=classification,
+                classification_choice=classification_choice,
+                is_active=True
+            ).first()
+
+            if existing:
+                return existing
+
+            patient_finding_classification = PatientFindingClassification.objects.create(
+                finding=self,
+                classification_id=classification_id,
+                classification_choice_id=classification_choice_id,
+            )
+
+            return patient_finding_classification
+
+        except FindingClassification.DoesNotExist:
+            raise ValidationError(f'Classification {classification_id} nicht gefunden')
+
+    def add_intervention(self, intervention_id, state="pending", date=None, user=None):
+        """
+        Add an intervention to the patient finding after validating its existence.
+        
+        Parameters:
+            intervention_id (int): The ID of the intervention to add.
+            state (str, optional): The state of the intervention. Defaults to "pending".
+            date (datetime, optional): The date of the intervention. Defaults to the current time if not provided.
+            user (User, optional): The user creating the intervention.
+        
+        Returns:
+            PatientFindingIntervention: The created intervention instance.
+        
+        Raises:
+            ValidationError: If the specified intervention does not exist.
+        """
+        from .patient_finding_intervention import PatientFindingIntervention
+        from ..finding import FindingIntervention
+        
+        try:
+            intervention = FindingIntervention.objects.get(id=intervention_id)
+            
+            patient_finding_intervention = PatientFindingIntervention.objects.create(
+                patient_finding=self,
+                intervention=intervention,
+                state=state,
+                date=date or timezone.now(),
+                created_by=user
+            )
+            
+            return patient_finding_intervention
+            
+        except FindingIntervention.DoesNotExist:
+            raise ValidationError(f'Intervention {intervention_id} nicht gefunden')
+
+    def add_video_segment(self, video_segment):
+        """
+        Associates a video segment with this patient finding.
+        
+        Parameters:
+        	video_segment: The video segment instance to add.
+        
+        Returns:
+        	The added video segment instance.
+        """
+        self.video_segments.add(video_segment) #TODO
+        return video_segment
+
+    # Manager für active/inactive Objekte
+    @property
+    def active_classifications(self):
+        """
+        Return all active classifications associated with this patient finding.
+        
+        Returns:
+        	QuerySet: Active related classifications where `is_active` is True.
+        """
+        return self.classifications.filter(is_active=True)
+    
+    @property
+    def locations(self):
+        """
+        Return all classifications of this patient finding that are of type "location".
+        
+        Returns:
+        	QuerySet: Classifications related to this finding filtered by the "location" classification type.
+        """
+        classifications = self.classifications.filter(
+            classification__classification_types__name__iexact="location"
+        )
+        return classifications
+
+    @property
+    def morphologies(self):
+        """
+        Return all classifications of this patient finding that are of type "morphology".
+        
+        Returns:
+        	QuerySet: Classifications related to this finding filtered by the "morphology" classification type.
+        """
+        classifications = self.classifications.filter(
+            classification__classification_types__name__iexact="morphology"
+        )
+        return classifications
+
+    @property
+    def active_interventions(self):
+        return self.interventions.filter(is_active=True)
+
+    @property
+    def links(self) -> "RequirementLinks":
+        """
+        Aggregates and returns all related model instances relevant for requirement evaluation
+        as a RequirementLinks object.
+        
+        This property provides access to:
+        - The finding associated with this patient finding
+        - All active finding classifications and their choices
+        - All active finding interventions
+        - The patient examination and patient
+        """
+        from endoreg_db.utils.links.requirement_link import RequirementLinks
+        from typing import cast, List
+        
+        # Get the base finding
+        findings_list = [self.finding] if self.finding else []
+        
+        # Get all active finding classifications and their choices
+        finding_classifications_list = []
+        finding_classification_choices_list = []
+        
+        for pf_classification in self.active_classifications:
+            if pf_classification.classification:
+                finding_classifications_list.append(pf_classification.classification)
+            if pf_classification.classification_choice:
+                finding_classification_choices_list.append(pf_classification.classification_choice)
+        
+        # Get all active finding interventions
+        finding_interventions_list = []
+        for pf_intervention in self.active_interventions:
+            if pf_intervention.intervention:
+                finding_interventions_list.append(pf_intervention.intervention)
+        
+        # Include patient examination and patient for context
+        patient_examinations_list = [self.patient_examination] if self.patient_examination else []
+        patient_findings_list = cast("List[PatientFinding]", [self])  # Include self for direct patient finding evaluations
+        
+        return RequirementLinks(
+            findings=findings_list,
+            finding_classifications=finding_classifications_list,
+            finding_classification_choices=finding_classification_choices_list,
+            finding_interventions=finding_interventions_list,
+            patient_examinations=patient_examinations_list,
+            patient_findings=patient_findings_list,
+        )

endoreg_db/models/medical/patient/patient_finding_classification.py

@@ -0,0 +1,207 @@
+from django.db import models
+from typing import TYPE_CHECKING, Dict
+import random
+import numpy as np
+
+
+# Corrected imports for type hints
+if TYPE_CHECKING:
+    from ..finding import (
+        FindingClassification, 
+        FindingClassificationChoice, 
+    )
+    from .patient_finding import PatientFinding
+
+class PatientFindingClassification(models.Model):
+    """Represents basic classifications for specific findings in a patient context.
+    Links a PatientFinding to a specific classification and choice, with optional subcategory values.
+    """
+    finding = models.ForeignKey( # type: ignore[assignment]
+        "PatientFinding", 
+        on_delete=models.CASCADE, 
+        related_name="classifications"
+    )
+    classification = models.ForeignKey( # type: ignore[assignment]
+        "FindingClassification", 
+        on_delete=models.CASCADE, 
+        related_name="patient_finding_classifications"
+    )
+    classification_choice = models.ForeignKey( # type: ignore[assignment]
+        "FindingClassificationChoice", 
+        on_delete=models.CASCADE, 
+        related_name="patient_finding_classifications"
+    )
+
+    is_active = models.BooleanField(default=True, help_text="Indicates if the classification is currently active.")
+    subcategories = models.JSONField(blank=True, null=True)
+    numerical_descriptors = models.JSONField(blank=True, null=True)
+
+    if TYPE_CHECKING:
+        finding: "PatientFinding"
+        classification: "FindingClassification"
+        classification_choice: "FindingClassificationChoice"
+
+    class Meta:
+        verbose_name = 'Patient Finding Classification'
+        verbose_name_plural = 'Patient Finding Classifications'
+        ordering = ['finding', 'classification', 'classification_choice']
+
+    def __str__(self):
+        """
+        Return a string representation combining the finding, classification, and classification choice.
+        """
+        return f"{self.finding} - {self.classification} - {self.classification_choice}"
+
+    def save(self, *args, **kwargs):
+        """
+        Saves the model instance after validating and initializing classification-related fields.
+        
+        Ensures that the selected classification choice is valid for the associated classification. If subcategories or numerical descriptors are unset, initializes them from the classification choice before saving.
+        """
+        if self.classification_choice not in self.classification.choices.all():
+            raise ValueError("classification_choice must be in classification.choices")
+
+        if not self.subcategories:
+            self.subcategories = self.classification_choice.subcategories
+
+        if not self.numerical_descriptors:
+            self.numerical_descriptors = self.classification_choice.numerical_descriptors
+
+        super().save(*args, **kwargs)
+
+    def initialize_and_get_subcategories(self):
+        """
+        Ensure the subcategories field is initialized and return its dictionary.
+        
+        Returns:
+            dict: The subcategories associated with this classification.
+        """
+        if not self.subcategories:
+            self.save()
+        return self.subcategories
+
+    def initialize_and_get_descriptors(self):
+        """
+        Return the numerical descriptors dictionary, initializing it if necessary.
+        
+        If the `numerical_descriptors` field is empty or uninitialized, the method triggers model initialization and returns the resulting dictionary.
+        """
+        if not self.numerical_descriptors:
+            self.save()
+        return self.numerical_descriptors
+
+    def set_subcategory(self, subcategory_name: str, subcategory_value: Dict):
+        """
+        Update the value of a specified subcategory and save the classification.
+        
+        Parameters:
+            subcategory_name (str): The name of the subcategory to update.
+            subcategory_value (dict): The value to assign to the subcategory.
+        
+        Returns:
+            dict: The updated subcategory dictionary.
+        """
+        assert subcategory_name in self.subcategories, "Subcategory must be in subcategories."
+        self.subcategories[subcategory_name]["value"] = subcategory_value
+        self.save()
+
+        return self.subcategories[subcategory_name]
+    
+    def set_random_subcategories(self):
+        """
+        Assign random values to all required subcategories that do not already have a value.
+        
+        For each required subcategory without a value, selects a random option from its available choices, updates the subcategory, saves the model, and returns the updated subcategories dictionary.
+        
+        Returns:
+            dict: The updated subcategories with random values assigned where needed.
+        """
+        if not self.subcategories or not self.numerical_descriptors:
+            self.save()
+        
+        self.refresh_from_db()
+
+        for subcategory_name, subcategory_dict in self.subcategories.items():
+            if subcategory_dict["required"] and not subcategory_dict.get("value", None):
+                subcategory_choice = random.choice(subcategory_dict["choices"])
+                self.subcategories[subcategory_name]["value"] = subcategory_choice
+
+        self.save()
+
+        return self.subcategories
+    
+    def get_random_value_for_numerical_descriptor(self, descriptor_name):
+
+        """
+        Generate a random value for the specified numerical descriptor using its defined distribution parameters.
+        
+        Parameters:
+            descriptor_name (str): The name of the numerical descriptor to generate a value for.
+        
+        Returns:
+            float: A randomly generated value based on the descriptor's distribution, clipped to its min and max range.
+        
+        Raises:
+            ValueError: If the descriptor's distribution type is not supported.
+        """  
+        assert descriptor_name in self.numerical_descriptors, "Descriptor must be in numerical descriptors."
+        descriptor = self.numerical_descriptors[descriptor_name]
+        min_val = descriptor.get("min", 0)
+        max_val = descriptor.get("max", 1)
+        distribution = descriptor.get("distribution", "normal")
+        if distribution == "normal":
+            mean = descriptor.get("mean", 0.5)
+            std = descriptor.get("std", 0.1)
+            value = np.random.normal(mean, std)
+            # clip value to min and max
+            value = np.clip(value, min_val, max_val)
+        elif distribution == "uniform":
+            value = np.random.uniform(min_val, max_val)
+        else:
+            raise ValueError("Distribution not supported")
+        
+        return value
+
+    def set_random_numerical_descriptor(self, descriptor_name, save=True):
+        """
+        Assigns a random value to the specified numerical descriptor and optionally saves the model.
+        
+        Parameters:
+            descriptor_name (str): The name of the numerical descriptor to update.
+            save (bool): If True, saves the model after updating the descriptor. Defaults to True.
+        
+        Returns:
+            dict: The updated numerical descriptor dictionary with the new random value.
+        
+        Raises:
+            ValueError: If the descriptor name is not present in the numerical descriptors.
+        """
+        if descriptor_name not in self.numerical_descriptors:
+            raise ValueError("Descriptor name must be in numerical descriptors.")
+
+        value = self.get_random_value_for_numerical_descriptor(descriptor_name)
+        
+        self.numerical_descriptors[descriptor_name]["value"] = value
+        if save:
+            self.save()
+
+        return self.numerical_descriptors[descriptor_name]
+    
+    def set_random_numerical_descriptors(self):
+        """
+        Assigns random values to all numerical descriptors and saves the model.
+        
+        Returns:
+            dict: The updated numerical_descriptors dictionary with assigned random values.
+        """
+        if not self.subcategories or not self.numerical_descriptors:
+            self.save()
+
+        numerical_descriptors = self.numerical_descriptors
+
+        for numerical_descriptor_name, _numerical_descriptor_dict in numerical_descriptors.items():
+            self.set_random_numerical_descriptor(numerical_descriptor_name, save=False)
+
+        self.save()
+
+        return self.numerical_descriptors

endoreg_db/models/medical/patient/patient_finding_intervention.py

@@ -0,0 +1,40 @@
+from operator import is_
+from django.db import models 
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from ..finding import FindingIntervention
+    from .patient_finding import PatientFinding
+
+class PatientFindingIntervention(models.Model):
+    finding = models.ForeignKey(
+        'PatientFinding', 
+        on_delete=models.CASCADE, 
+        related_name='interventions'
+    )
+    intervention = models.ForeignKey(
+        'FindingIntervention',
+        on_delete=models.CASCADE,
+        related_name='patient_finding_interventions'
+    )
+    is_active = models.BooleanField(
+        default=True, 
+        help_text="Indicates if the intervention is currently active."
+    )
+    state = models.CharField(max_length=100, blank=True, null=True)
+    time_start = models.DateTimeField(blank=True, null=True)
+    time_end = models.DateTimeField(blank=True, null=True)
+    date = models.DateField(blank=True, null=True)
+
+    if TYPE_CHECKING:
+        finding: "PatientFinding"
+        intervention: "FindingIntervention"
+
+    def __str__(self):
+        return self.intervention.name
+
+    def natural_key(self):
+        return (self.intervention.name,)
+    
+
+    objects = models.Manager()