Added loading and saving annotation projects

Makefile

@@ -51,10 +51,10 @@ watch:            ## Run tests on every change.
 
 .PHONY: clean
 clean:            ## Clean unused files.
-	@find ./ -name '*.pyc' -exec rm -f  \;
-	@find ./ -name '__pycache__' -exec rm -rf {} \;
-	@find ./ -name 'Thumbs.db' -exec rm -f {} \;
-	@find ./ -name '*~' -exec rm -f {} \;
+	@find src/ -name '*.pyc' -exec rm -f  \;
+	@find src/ -name '__pycache__' -exec rm -rf {} \;
+	@find src/ -name 'Thumbs.db' -exec rm -f {} \;
+	@find src/ -name '*~' -exec rm -f {} \;
 	@rm -rf .cache
 	@rm -rf .pytest_cache
 	@rm -rf .mypy_cache

src/soundevent/__init__.py

@@ -18,7 +18,6 @@
     ModelRun,
     MultiLineString,
     MultiPoint,
-    Sequence,
     MultiPolygon,
     Note,
     Point,
@@ -27,6 +26,7 @@
     PredictedTag,
     ProcessedClip,
     Recording,
+    Sequence,
     SoundEvent,
     Tag,
     Time,

src/soundevent/audio/__init__.py

@@ -1,7 +1,7 @@
 """Soundevent functions for handling audio files and arrays."""
 
 from .io import load_clip, load_recording
-from .media_info import get_media_info, MediaInfo
+from .media_info import MediaInfo, get_media_info
 from .spectrograms import compute_spectrogram
 
 __all__ = [

src/soundevent/data/__init__.py

@@ -36,7 +36,6 @@
 from soundevent.data.clips import Clip
 from soundevent.data.dataset import Dataset
 from soundevent.data.features import Feature
-from soundevent.data.sequences import Sequence
 from soundevent.data.geometries import (
     MAX_FREQUENCY,
     BoundingBox,
@@ -60,6 +59,7 @@
 from soundevent.data.predicted_tags import PredictedTag
 from soundevent.data.processed_clip import ProcessedClip
 from soundevent.data.recordings import Recording
+from soundevent.data.sequences import Sequence
 from soundevent.data.sound_events import SoundEvent
 from soundevent.data.tags import Tag
 

src/soundevent/data/annotation_projects.py

@@ -48,14 +48,17 @@
 class AnnotationProject(BaseModel):
     """Annotation collection."""
 
-    id: UUID = Field(default_factory=uuid4)
+    id: UUID = Field(default_factory=uuid4, repr=False)
     """Unique identifier for the annotation collection."""
 
     name: str
     """Name of the annotation project."""
 
-    description: Optional[str] = None
+    description: Optional[str] = Field(None, repr=False)
     """Description of the annotation collection."""
 
-    tasks: List[AnnotationTask] = Field(default_factory=list)
+    tasks: List[AnnotationTask] = Field(default_factory=list, repr=False)
     """List of annotation tasks in the project."""
+
+    instructions: Optional[str] = Field(default=None, repr=False)
+    """Instructions for annotators."""

src/soundevent/data/annotation_tasks.py

@@ -47,7 +47,9 @@
 research projects.
 
 """
+import datetime
 from typing import List, Optional
+from uuid import UUID, uuid4
 
 from pydantic import BaseModel, Field
 
@@ -60,6 +62,9 @@
 class AnnotationTask(BaseModel):
     """Annotation task."""
 
+    id: UUID = Field(default_factory=uuid4)
+    """Unique identifier for the annotation task."""
+
     clip: Clip
     """The annotated clip."""
 
@@ -69,6 +74,9 @@ class AnnotationTask(BaseModel):
     completed_by: Optional[str] = None
     """The user who completed the annotation task."""
 
+    completed_on: Optional[datetime.datetime] = None
+    """The date and time when the annotation task was completed."""
+
     notes: List[Note] = Field(default_factory=list)
     """Notes associated with the annotation task."""
 
@@ -77,3 +85,7 @@ class AnnotationTask(BaseModel):
 
     completed: bool = False
     """Whether the annotation task has been completed."""
+
+    def __hash__(self):
+        """Compute the hash value for the annotation task."""
+        return hash(self.id)

src/soundevent/data/annotations.py

@@ -77,10 +77,14 @@ class Annotation(BaseModel):
     tags: List[Tag] = Field(default_factory=list)
     """User provided tags to the annotated sound event."""
 
-    created_on: datetime.datetime = Field(
+    created_on: Optional[datetime.datetime] = Field(
         default_factory=datetime.datetime.now
     )
     """The time at which the annotation was created.
 
     Important for tracking the progress of an annotation task.
     """
+
+    def __hash__(self):
+        """Compute the hash of the annotation."""
+        return hash(self.id)

src/soundevent/data/clips.py

@@ -57,6 +57,9 @@
 class Clip(BaseModel):
     """Clip model."""
 
+    uuid: UUID = Field(default_factory=uuid4)
+    """The unique identifier of the clip."""
+
     recording: Recording
     """The recording that the clip belongs to."""
 
@@ -66,9 +69,6 @@ class Clip(BaseModel):
     end_time: float
     """The end time of the clip in seconds."""
 
-    uuid: UUID = Field(default_factory=uuid4)
-    """The unique identifier of the clip."""
-
     tags: List[Tag] = Field(default_factory=list)
     """List of tags associated with the clip."""
 
@@ -81,3 +81,7 @@ def validate_times(cls, values):
         if values["start_time"] > values["end_time"]:
             raise ValueError("start_time must be less than end_time")
         return values
+
+    def __hash__(self):
+        """Hash clip object."""
+        return hash((self.recording, self.start_time, self.end_time))

src/soundevent/data/geometries.py

@@ -53,9 +53,9 @@
 """
 import sys
 from abc import ABC, abstractmethod
-from typing import List, Tuple
+from typing import List, Tuple, Union
 
-from pydantic import BaseModel, ConfigDict, Field, PrivateAttr, field_validator
+from pydantic import BaseModel, Field, PrivateAttr, field_validator
 from shapely import geometry
 from shapely.geometry.base import BaseGeometry
 
@@ -112,14 +112,16 @@ class Geometry(BaseModel, ABC):
     geometry object is always in sync with the Shapely geometry object.
     """
 
-    model_config = ConfigDict(
-        arbitrary_types_allowed=True,
+    type: str = Field(
+        description="the type of geometry used to locate the sound event.",
         frozen=True,
+        include=True,
     )
 
-    type: GeometryType = Field(
-        ...,
-        description="the type of geometry used to locate the sound event.",
+    coordinates: Union[float, List, Tuple] = Field(
+        description="the coordinates of the geometry.",
+        frozen=True,
+        include=True,
     )
 
     _geom: BaseGeometry = PrivateAttr()
@@ -182,7 +184,7 @@ class TimeStamp(Geometry):
     time interval.
     """
 
-    type: Literal["TimeStamp"] = "TimeStamp"
+    type: str = "TimeStamp"
 
     coordinates: Time = Field(
         ...,
@@ -216,10 +218,9 @@ class TimeInterval(Geometry):
     not have a clear frequency range.
     """
 
-    type: Literal["TimeInterval"] = "TimeInterval"
+    type: str = "TimeInterval"
 
     coordinates: Tuple[Time, Time] = Field(
-        ...,
         description="The time interval of the sound event.",
     )
     """The time interval of the sound event.
@@ -267,7 +268,7 @@ class Point(Geometry):
     time and frequency.
     """
 
-    type: Literal["Point"] = "Point"
+    type: str = "Point"
 
     coordinates: Tuple[Time, Frequency] = Field(
         ...,
@@ -297,7 +298,7 @@ class LineString(Geometry):
     trajectory.
     """
 
-    type: Literal["LineString"] = "LineString"
+    type: str = "LineString"
 
     coordinates: List[Tuple[Time, Frequency]] = Field(
         ...,
@@ -345,7 +346,7 @@ class Polygon(Geometry):
     frequency trajectory and that are bounded by a clear start and end time.
     """
 
-    type: Literal["Polygon"] = "Polygon"
+    type: str = "Polygon"
 
     coordinates: List[List[Tuple[Time, Frequency]]] = Field(
         ...,
@@ -384,7 +385,7 @@ class BoundingBox(Geometry):
     range and start and stop times.
     """
 
-    type: Literal["BoundingBox"] = "BoundingBox"
+    type: str = "BoundingBox"
 
     coordinates: Tuple[Time, Frequency, Time, Frequency] = Field(
         ...,
@@ -431,7 +432,7 @@ class MultiPoint(Geometry):
     together form a sound event.
     """
 
-    type: Literal["MultiPoint"] = "MultiPoint"
+    type: str = "MultiPoint"
 
     coordinates: List[Tuple[Time, Frequency]] = Field(
         ...,
@@ -460,7 +461,7 @@ class MultiLineString(Geometry):
     harmonics.
     """
 
-    type: Literal["MultiLineString"] = "MultiLineString"
+    type: str = "MultiLineString"
 
     coordinates: List[List[Tuple[Time, Frequency]]] = Field(
         ...,
@@ -520,7 +521,7 @@ class MultiPolygon(Geometry):
     polygons.
     """
 
-    type: Literal["MultiPolygon"] = "MultiPolygon"
+    type: str = "MultiPolygon"
 
     coordinates: List[List[List[Tuple[Time, Frequency]]]] = Field(
         ...,

src/soundevent/data/recordings.py

@@ -51,9 +51,9 @@
 
 import datetime
 import os
-from uuid import UUID, uuid4
 from pathlib import Path
 from typing import List, Optional
+from uuid import UUID, uuid4
 
 from pydantic import BaseModel, Field
 
@@ -125,7 +125,7 @@ class Recording(BaseModel):
 
     def __hash__(self):
         """Hash function."""
-        return hash(self.hash)
+        return hash(self.id)
 
     @classmethod
     def from_file(

src/soundevent/data/sound_events.py

@@ -36,8 +36,8 @@
 of the sound events.
 """
 
+from typing import List, Optional
 from uuid import UUID, uuid4
-from typing import List
 
 from pydantic import BaseModel, Field
 
@@ -60,11 +60,15 @@ class SoundEvent(BaseModel):
     recording: Recording
     """The recording containing the sound event."""
 
-    geometry: Geometry
+    geometry: Optional[Geometry]
     """The geometry locating the sound event within the recording."""
 
     tags: List[Tag] = Field(default_factory=list)
     """The tags associated with the sound event."""
 
     features: List[Feature] = Field(default_factory=list)
     """The features associated with the sound event."""
+
+    def __hash__(self):
+        """Compute the hash of the sound event."""
+        return hash(self.uuid)

src/soundevent/io/__init__.py

@@ -6,7 +6,6 @@
 
 from soundevent.io.datasets import load_dataset, save_dataset
 
-
 __all__ = [
     "load_dataset",
     "save_dataset",