Add trials metadata

src/neuroconv/datainterfaces/__init__.py

@@ -94,6 +94,7 @@
 # Text
 from .text.csv.csvtimeintervalsinterface import CsvTimeIntervalsInterface
 from .text.excel.exceltimeintervalsinterface import ExcelTimeIntervalsInterface
+from .text.timeintervalsinterface import TimeIntervalsInterface
 
 interface_list = [
     # Ecephys
@@ -162,6 +163,7 @@
     # Text
     CsvTimeIntervalsInterface,
     ExcelTimeIntervalsInterface,
+    TimeIntervalsInterface,
 ]
 
 interfaces_by_category = dict(

src/neuroconv/datainterfaces/text/csv/csvtimeintervalsinterface.py

@@ -1,10 +1,10 @@
 import pandas as pd
 
-from ..timeintervalsinterface import TimeIntervalsInterface
+from ..timeintervalsinterface import BaseTimeIntervalsInterface
 from ....utils.types import FilePathType
 
 
-class CsvTimeIntervalsInterface(TimeIntervalsInterface):
+class CsvTimeIntervalsInterface(BaseTimeIntervalsInterface):
     """Interface for adding data from a .csv file as a TimeIntervals object."""
 
     display_name = "CSV time interval table"

src/neuroconv/datainterfaces/text/excel/exceltimeintervalsinterface.py

@@ -2,11 +2,11 @@
 
 import pandas as pd
 
-from ..timeintervalsinterface import TimeIntervalsInterface
+from ..timeintervalsinterface import BaseTimeIntervalsInterface
 from ....utils.types import FilePathType
 
 
-class ExcelTimeIntervalsInterface(TimeIntervalsInterface):
+class ExcelTimeIntervalsInterface(BaseTimeIntervalsInterface):
     """Interface for adding data from an Excel file to NWB as a TimeIntervals object."""
 
     display_name = "Excel time interval table"

src/neuroconv/datainterfaces/text/time_intervals.schema.json

@@ -0,0 +1,50 @@
+{
+  "type": "object",
+  "patternProperties": {
+    "^[a-zA-Z0-9_]+$": {
+      "type": "object",
+      "properties": {
+        "description": {
+          "type": "string"
+        },
+        "columns": {
+          "type": "array",
+          "items": {
+            "type": "object",
+            "properties": {
+              "name": {
+                "type": "string"
+              },
+              "description": {
+                "type": "string"
+              }
+            },
+            "required": [
+              "name"
+            ],
+            "additionalProperties": false
+          }
+        },
+        "data": {
+          "type": "array",
+          "items": {
+            "type": "object",
+            "properties": {
+              "start_time": {
+                "type": "number"
+              },
+              "stop_time": {
+                "type": "number"
+              }
+            },
+            "required": [
+              "start_time",
+              "stop_time"
+            ],
+            "additionalProperties": true
+          }
+        }
+      }
+    }
+  }
+}

src/neuroconv/datainterfaces/text/timeintervalsinterface.py

@@ -4,14 +4,15 @@
 
 import numpy as np
 from pynwb import NWBFile
+from pynwb.epoch import TimeIntervals
 
 from ...basedatainterface import BaseDataInterface
 from ...tools.text import convert_df_to_time_intervals
-from ...utils.dict import load_dict_from_file
+from ...utils.dict import DeepDict, load_dict_from_file
 from ...utils.types import FilePathType
 
 
-class TimeIntervalsInterface(BaseDataInterface):
+class BaseTimeIntervalsInterface(BaseDataInterface):
     """Abstract Interface for time intervals."""
 
     keywords = ("table", "trials", "epochs", "time intervals")
@@ -154,3 +155,58 @@ def add_to_nwbfile(
     @abstractmethod
     def _read_file(self, file_path: FilePathType, **read_kwargs):
         pass
+
+
+class TimeIntervalsInterface(BaseDataInterface):
+
+    display_name = "Time Intervals"
+    keywords = ("table", "trials", "epochs", "time intervals")
+    associated_suffixes = ()
+    info = "Interface for time intervals data added to metadata."
+
+    @classmethod
+    def add_to_nwbfile(cls, nwbfile: NWBFile, metadata: DeepDict) -> None:
+
+        for table_name, table_metadata in metadata["TimeIntervals"].items():
+            time_intervals = TimeIntervals(
+                name=table_name, description=table_metadata.get("description", "no description")
+            )
+
+            # add custom columns
+            for column_name, column_metadata in table_metadata["columns"].items():
+                if column_name not in ("start_time", "stop_time"):
+                    time_intervals.add_column(
+                        name=column_name, description=column_metadata.get("description", "no description")
+                    )
+            # handle any custom columns that were not defined in columns object
+            for column_name in table_metadata["data"][0]:
+                if column_name not in time_intervals.colnames + ("start_time", "stop_time"):
+                    time_intervals.add_column(name=key, description="no description")
+            # add data
+            for row in table_metadata["data"]:
+                time_intervals.add_interval(**row)
+
+            # add to nwbfile
+            if table_name == "trials":
+                nwbfile.trials = time_intervals
+            else:
+                nwbfile.add_time_intervals(time_intervals)
+
+    def get_metadata_schema(self) -> dict:
+        metadata_schema = super().get_metadata_schema()
+
+        time_intervals_schema = load_dict_from_file(Path(__file__).parent / "time_intervals.schema.json")
+        metadata_schema["properties"]["TimeIntervals"] = time_intervals_schema
+        return metadata_schema
+
+    def get_metadata(self) -> dict:
+        metadata = super().get_metadata()
+        metadata["TimeIntervals"]["trials"] = dict(
+            columns=dict(
+                start_time=dict(description="start time of the trial"),
+                stop_time=dict(description="stop time of the trial"),
+            ),
+            data=[],
+        )
+
+        return metadata

tests/test_minimal/test_tools/test_importing.py

@@ -27,5 +27,3 @@ def test_guide_attributes():
                 assert isinstance(
                     value, tuple
                 ), f"{name} incorrectly specified GUIDE-related attribute 'associated_suffixes' (must be tuple)."
-            if isinstance(value, tuple):
-                assert len(value) > 0, f"{name} is missing entries in GUIDE related attribute {key}."

tests/test_text/test_time_intervals.py

@@ -8,6 +8,7 @@
 from neuroconv.datainterfaces import (
     CsvTimeIntervalsInterface,
     ExcelTimeIntervalsInterface,
+    TimeIntervalsInterface,
 )
 from neuroconv.tools.nwb_helpers import make_nwbfile_from_metadata
 from neuroconv.tools.text import convert_df_to_time_intervals
@@ -96,3 +97,40 @@ def test_csv_round_trip_rename(tmp_path):
 def test_get_metadata_schema():
     interface = CsvTimeIntervalsInterface(trials_csv_path)
     interface.get_metadata_schema()
+
+
+def test_trials():
+    metadata = dict(
+        NWBFile=dict(
+            session_start_time=datetime.now().astimezone(),
+        ),
+        TimeIntervals=dict(
+            trials=dict(
+                columns=dict(
+                    start_time=dict(description="start time of the trial"),
+                    stop_time=dict(description="stop time of the trial"),
+                    correct=dict(description="correct or not"),
+                ),
+                data=[
+                    dict(start_time=0.0, stop_time=1.0, correct=True),
+                    dict(start_time=1.0, stop_time=2.0, correct=False),
+                ],
+            ),
+            new_table=dict(
+                columns=dict(
+                    stim_id=dict(description="stimulus ID"),
+                ),
+                data=[
+                    dict(start_time=0.0, stop_time=1.0, stim_id=0),
+                    dict(start_time=1.0, stop_time=2.0, stim_id=1),
+                ],
+            ),
+        ),
+    )
+
+    nwbfile = TimeIntervalsInterface().create_nwbfile(metadata)
+    assert nwbfile.trials.correct.description == "correct or not"
+    assert_array_equal(nwbfile.trials.correct[:], [True, False])
+    assert_array_equal(nwbfile.trials.start_time[:], [0.0, 1.0])
+
+    assert_array_equal(nwbfile.intervals["new_table"].stim_id[:], [0, 1])