How to Load Tabular Data

Custom TsvLoader/CsvLoader, column_specs typed fields, nested-JSON Field access, group_by child timelines

How to Load Tabular Data

The fastest way to get music data into TimeToAlign! is through tabular loaders. If your data is in CSV or TSV format, you’re just 3 lines of code away from analysis.

What you’ll learn: - Load music annotations from TSV/CSV files - Access event counts, coordinate ranges, and metadata - Create timelines from loaded data - Write a custom loader that maps your columns to event fields - Promote selected columns to typed fields with column_specs - Reach nested JSON columns with Field

Time: 15 minutes

This is the gentle introduction. For the full column-to-field mechanism — the Step-1 resolution chain, composite columns, and Step-2 field promotion — see the in-depth CSV/TSV how-to.

TL;DR

from timetoalign.loader.tabular import TsvLoader

loader = TsvLoader()
loader.load("beethoven.notes.tsv")

df = loader.events.to_dataframe()    # Get as DataFrame
timeline = loader.create_timeline()  # Create Timeline

Setup

from timetoalign.testdata import ensure_data

BEETHOVEN = ensure_data("score") / "beethoven_woo71"
THORESEN = ensure_data("thoresen")

# Available files
{
    "Beethoven files": [f.name for f in BEETHOVEN.glob("WoO71.*.tsv")],
    "Thoresen files": [f.name for f in THORESEN.glob("*.tsv")],
}

{'Beethoven files': ['WoO71.chords.tsv',
  'WoO71.measures.tsv',
  'WoO71.notes.tsv'],
 'Thoresen files': ['thoresen_test_h.tsv', 'thoresen_test.tsv']}

Loading Notes from TSV

TsvLoader is the generic starting point for a TSV whose columns you control. Declare the coordinate columns and promote the values you want as typed fields. For a standard ms3 score export, use the score Ms3Loader instead; it understands its notes, measures, chords, and harmonies facets.

Three lines of code:

from timetoalign.core import NumberType, TimeUnit  # noqa: E402
from timetoalign.loader.tabular import TsvLoader  # noqa: E402


class BeethovenNotesLoader(TsvLoader):
    """Generic TSV configuration for the coordinate and pitch columns used here."""

    start_column = "quarterbeats"
    duration_column = "duration_qb"
    _default_unit = TimeUnit.quarters
    coordinate_type = NumberType.fraction
    column_specs = {"midi": int, "staff": int, "voice": int}


loader = BeethovenNotesLoader()
loader.load(BEETHOVEN / "WoO71.notes.tsv")

f"{len(loader.events):,} notes loaded"

'4,753 notes loaded'

Converting to pandas

Use to_dataframe() to get a DataFrame with clean coordinate values:

loader.events.to_dataframe().head()

	id	name	temporal_type	event_type	start	end	duration	scalar	mc	tied	...	gracenote	nominal_duration	staff	volta	timesig	mn_onset	tpc	octave	voice
0	e000000	A3	interval	Event	0	1	1	1	1	NaN	...	NaN	1/4	2	NaN	2/4	1/4	3	3	1
1	e000001	E4	interval	Event	0	1	1	1	1	NaN	...	NaN	1/4	1	NaN	2/4	1/4	4	4	2
2	e000002	A4	interval	Event	0	1/2	1/2	1	1	NaN	...	NaN	1/8	1	NaN	2/4	1/4	3	4	1
3	e000003	C#5	interval	Event	0	1/2	1/2	1	1	NaN	...	NaN	1/8	1	NaN	2/4	1/4	7	5	1
4	e000004	E5	interval	Event	1/2	1	1/2	1	1	NaN	...	NaN	1/8	1	NaN	2/4	3/8	4	5	1

5 rows × 24 columns

Quick Statistics

The loader provides immediate access to summary information:

{
    "event_count": len(loader.events),
    "coordinate_range": loader.events.coordinate_range(),
    "unit": str(loader.unit),
    "number_type": str(loader.number_type),
}

{'event_count': 4753,
 'coordinate_range': (Fraction(-8016172774238891, 35184372088832),
  Fraction(1753, 2)),
 'unit': 'quarters',
 'number_type': 'fraction'}

Creating Timelines

TimeToAlign! represents temporal data as Timelines:

timeline = loader.create_timeline(uid="beethoven_notes")
timeline

ContinuousLogicalTimeline[beethoven_notes] (4753 events)
                      0 _______________________________ 876.5 quarters

Custom Loaders for Non-Standard Formats

For files that don’t match the ms3 format, create a custom loader by subclassing TsvLoader or CsvLoader. You point the canonical column attributes (start_column, duration_column, …) at the names in your file.

Let’s load the Thoresen annotations file, which has a different column structure:

import pandas as pd  # noqa: E402

pd.read_csv(THORESEN / "thoresen_test.tsv", sep="\t", nrows=3)

	event_id	alignment_group_id	start_time_sec	duration_sec	event_type	graphical_element_id	image_filename	rect_coords_json	text_content	text_anchor_xy_json	layer_order	description
0	annot_cue_001	NaN	0.0	5.0	rectangle	rect_a	thoresen_2010_form-building-patterns_p90-91_pa...	{"x": 10, "y": 90, "width": 148, "height": 55}	NaN	NaN	NaN	NaN
1	annot_cue_002	NaN	1.5	4.0	rectangle	rect_b	thoresen_2010_form-building-patterns_p90-91_pa...	{"x": 40, "y": 37, "width": 127, "height": 21}	NaN	NaN	NaN	NaN
2	annot_cue_003	NaN	3.5	2.0	rectangle	rect_c	thoresen_2010_form-building-patterns_p90-91_pa...	{"x": 111, "y": 60, "width": 57, "height": 23}	NaN	NaN	NaN	NaN

The Simplest Custom Loader

Map the core coordinate columns and nothing else. Every source column that you do not name survives as an opaque property column — carried alongside the event so you never lose data, but left untyped:

from timetoalign.loader.tabular import TsvLoader  # noqa: E402


class ThoresenLoader(TsvLoader):
    """Minimal loader — maps the core event fields."""

    id_column = "event_id"
    start_column = "start_time_sec"
    duration_column = "duration_sec"
    event_type_column = "event_type"
    name_column = "description"

    _default_unit = TimeUnit.seconds
    coordinate_type = NumberType.float


thoresen = ThoresenLoader()
thoresen.load(THORESEN / "thoresen_test.tsv")
thoresen.events.to_dataframe()

	id	name	temporal_type	event_type	start	end	duration	rect_coords_json	layer_order	text_content	graphical_element_id	image_filename	alignment_group_id	text_anchor_xy_json
0	annot_cue_001	NaN	interval	rectangle	0.0	5.00	5.00	{"x": 10, "y": 90, "width": 148, "height": 55}	NaN	NaN	rect_a	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
1	annot_cue_002	NaN	interval	rectangle	1.5	5.50	4.00	{"x": 40, "y": 37, "width": 127, "height": 21}	NaN	NaN	rect_b	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
2	annot_cue_003	NaN	interval	rectangle	3.5	5.50	2.00	{"x": 111, "y": 60, "width": 57, "height": 23}	NaN	NaN	rect_c	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
3	annot_cue_004	NaN	interval	rectangle	34.6	39.80	5.20	{"x": 145, "y": 90, "width": 160, "height": 58}	NaN	NaN	rect_a2	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
4	annot_cue_005	NaN	interval	rectangle	43.5	48.00	4.50	{"x": 385, "y": 46, "width": 139, "height": 20}	NaN	NaN	rect_h2	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
5	annot_cue_006	NaN	interval	rectangle	71.0	75.75	4.75	{"x": 310, "y": 93, "width": 154, "height": 18}	NaN	NaN	rect_d3	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
6	annot_cue_007	NaN	interval	rectangle	76.0	83.50	7.50	{"x": 456, "y": 69, "width": 229, "height": 18}	NaN	NaN	rect_b3	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
7	annot_cue_008	NaN	interval	rectangle	90.5	94.50	4.00	{"x": 14, "y": 115, "width": 127, "height": 31}	NaN	NaN	rect_i4	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
8	annot_cue_009	NaN	interval	rectangle	113.4	116.40	3.00	{"x": 663, "y": 82, "width": 97, "height": 23}	NaN	NaN	rect_a4	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
9	annot_cue_010	NaN	interval	rectangle	121.0	128.50	7.50	{"x": 19, "y": 119, "width": 251, "height": 29}	NaN	NaN	rect_i5	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
10	annot_cue_011	NaN	interval	rectangle	141.0	142.50	1.50	{"x": 595, "y": 45, "width": 64, "height": 21}	NaN	NaN	rect_f5	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN

Promoting Columns to Typed Fields with `column_specs`

When a property column should become a typed field, name it in column_specs. The keys are source-column names; the values are anything the loader can resolve to a field — a bare Python type (int / float / str) is the simplest form. Here we give two extra columns explicit types:

class ThoresenTypedLoader(TsvLoader):
    """Selected columns promoted to typed fields."""

    id_column = "event_id"
    start_column = "start_time_sec"
    duration_column = "duration_sec"
    event_type_column = "event_type"
    name_column = "description"

    _default_unit = TimeUnit.seconds
    coordinate_type = NumberType.float

    # Promote these source columns to typed fields.
    column_specs = {
        "image_filename": str,
        "graphical_element_id": str,
    }


typed = ThoresenTypedLoader()
typed.load(THORESEN / "thoresen_test.tsv")
typed.events.to_dataframe()

	id	name	temporal_type	event_type	start	end	duration	rect_coords_json	layer_order	text_content	graphical_element_id	image_filename	alignment_group_id	text_anchor_xy_json
0	annot_cue_001	NaN	interval	rectangle	0.0	5.00	5.00	{"x": 10, "y": 90, "width": 148, "height": 55}	NaN	NaN	rect_a	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
1	annot_cue_002	NaN	interval	rectangle	1.5	5.50	4.00	{"x": 40, "y": 37, "width": 127, "height": 21}	NaN	NaN	rect_b	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
2	annot_cue_003	NaN	interval	rectangle	3.5	5.50	2.00	{"x": 111, "y": 60, "width": 57, "height": 23}	NaN	NaN	rect_c	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
3	annot_cue_004	NaN	interval	rectangle	34.6	39.80	5.20	{"x": 145, "y": 90, "width": 160, "height": 58}	NaN	NaN	rect_a2	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
4	annot_cue_005	NaN	interval	rectangle	43.5	48.00	4.50	{"x": 385, "y": 46, "width": 139, "height": 20}	NaN	NaN	rect_h2	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
5	annot_cue_006	NaN	interval	rectangle	71.0	75.75	4.75	{"x": 310, "y": 93, "width": 154, "height": 18}	NaN	NaN	rect_d3	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
6	annot_cue_007	NaN	interval	rectangle	76.0	83.50	7.50	{"x": 456, "y": 69, "width": 229, "height": 18}	NaN	NaN	rect_b3	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
7	annot_cue_008	NaN	interval	rectangle	90.5	94.50	4.00	{"x": 14, "y": 115, "width": 127, "height": 31}	NaN	NaN	rect_i4	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
8	annot_cue_009	NaN	interval	rectangle	113.4	116.40	3.00	{"x": 663, "y": 82, "width": 97, "height": 23}	NaN	NaN	rect_a4	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
9	annot_cue_010	NaN	interval	rectangle	121.0	128.50	7.50	{"x": 19, "y": 119, "width": 251, "height": 29}	NaN	NaN	rect_i5	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
10	annot_cue_011	NaN	interval	rectangle	141.0	142.50	1.50	{"x": 595, "y": 45, "width": 64, "height": 21}	NaN	NaN	rect_f5	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN

That is the entire idea: columns are a source artefact; fields are a TimeToAlign! artefact. column_specs is the bridge. A bare type is the gentlest entry — the in-depth CSV/TSV how-to covers the full resolution chain (composite columns, semantic pitch/id fields, and the Step-2 field_specs promotion stage) for richer formats.

Nested JSON Column Access with `Field`

The Thoresen data has a rect_coords_json column containing pixel coordinates as JSON:

{"x": 10, "y": 90, "width": 148, "height": 55}

Use Field("column", "nested_field") to point a coordinate attribute straight at a nested value. TimeToAlign! parses the JSON automatically. ComputedField lets you derive a coordinate from a small formula over those nested values:

from timetoalign.loader import ComputedField, Field  # noqa: E402


class ThoresenGraphicalLoader(TsvLoader):
    """Loader using PIXEL coordinates from a nested JSON column."""

    # Nested fields are addressed directly; JSON is parsed automatically.
    start_column = Field("rect_coords_json", "x")
    end_column = ComputedField(
        "end", formula="rect_coords_json.x + rect_coords_json.width"
    )

    _default_unit = TimeUnit.pixels
    coordinate_type = NumberType.float
    default_event_type = "Rectangle"


graphical = ThoresenGraphicalLoader()
graphical.load(THORESEN / "thoresen_test.tsv")

{
    "unit": str(graphical.unit),
    "coordinate_range": graphical.events.coordinate_range(),
}

{'unit': 'pixels', 'coordinate_range': (10, 760)}

Two Coordinate Systems from One File

The same TSV file backs timelines in different coordinate systems — seconds from the time columns, pixels from the JSON column:

# Physical timeline (seconds)
typed.events.to_dataframe()

	id	name	temporal_type	event_type	start	end	duration	rect_coords_json	layer_order	text_content	graphical_element_id	image_filename	alignment_group_id	text_anchor_xy_json
0	annot_cue_001	NaN	interval	rectangle	0.0	5.00	5.00	{"x": 10, "y": 90, "width": 148, "height": 55}	NaN	NaN	rect_a	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
1	annot_cue_002	NaN	interval	rectangle	1.5	5.50	4.00	{"x": 40, "y": 37, "width": 127, "height": 21}	NaN	NaN	rect_b	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
2	annot_cue_003	NaN	interval	rectangle	3.5	5.50	2.00	{"x": 111, "y": 60, "width": 57, "height": 23}	NaN	NaN	rect_c	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
3	annot_cue_004	NaN	interval	rectangle	34.6	39.80	5.20	{"x": 145, "y": 90, "width": 160, "height": 58}	NaN	NaN	rect_a2	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
4	annot_cue_005	NaN	interval	rectangle	43.5	48.00	4.50	{"x": 385, "y": 46, "width": 139, "height": 20}	NaN	NaN	rect_h2	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
5	annot_cue_006	NaN	interval	rectangle	71.0	75.75	4.75	{"x": 310, "y": 93, "width": 154, "height": 18}	NaN	NaN	rect_d3	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
6	annot_cue_007	NaN	interval	rectangle	76.0	83.50	7.50	{"x": 456, "y": 69, "width": 229, "height": 18}	NaN	NaN	rect_b3	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
7	annot_cue_008	NaN	interval	rectangle	90.5	94.50	4.00	{"x": 14, "y": 115, "width": 127, "height": 31}	NaN	NaN	rect_i4	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
8	annot_cue_009	NaN	interval	rectangle	113.4	116.40	3.00	{"x": 663, "y": 82, "width": 97, "height": 23}	NaN	NaN	rect_a4	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
9	annot_cue_010	NaN	interval	rectangle	121.0	128.50	7.50	{"x": 19, "y": 119, "width": 251, "height": 29}	NaN	NaN	rect_i5	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN
10	annot_cue_011	NaN	interval	rectangle	141.0	142.50	1.50	{"x": 595, "y": 45, "width": 64, "height": 21}	NaN	NaN	rect_f5	thoresen_2010_form-building-patterns_p90-91_pa...	NaN	NaN

# Graphical timeline (pixels)
graphical.events.to_dataframe()

	id	name	temporal_type	event_type	start	end	duration	layer_order	text_content	graphical_element_id	image_filename	event_id	alignment_group_id	duration_sec	start_time_sec	text_anchor_xy_json	description
0	e000000	NaN	interval	rectangle	10	158	148	NaN	NaN	rect_a	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_001	NaN	5.00	0.0	NaN	NaN
1	e000001	NaN	interval	rectangle	40	167	127	NaN	NaN	rect_b	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_002	NaN	4.00	1.5	NaN	NaN
2	e000002	NaN	interval	rectangle	111	168	57	NaN	NaN	rect_c	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_003	NaN	2.00	3.5	NaN	NaN
3	e000003	NaN	interval	rectangle	145	305	160	NaN	NaN	rect_a2	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_004	NaN	5.20	34.6	NaN	NaN
4	e000004	NaN	interval	rectangle	385	524	139	NaN	NaN	rect_h2	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_005	NaN	4.50	43.5	NaN	NaN
5	e000005	NaN	interval	rectangle	310	464	154	NaN	NaN	rect_d3	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_006	NaN	4.75	71.0	NaN	NaN
6	e000006	NaN	interval	rectangle	456	685	229	NaN	NaN	rect_b3	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_007	NaN	7.50	76.0	NaN	NaN
7	e000007	NaN	interval	rectangle	14	141	127	NaN	NaN	rect_i4	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_008	NaN	4.00	90.5	NaN	NaN
8	e000008	NaN	interval	rectangle	663	760	97	NaN	NaN	rect_a4	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_009	NaN	3.00	113.4	NaN	NaN
9	e000009	NaN	interval	rectangle	19	270	251	NaN	NaN	rect_i5	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_010	NaN	7.50	121.0	NaN	NaN
10	e000010	NaN	interval	rectangle	595	659	64	NaN	NaN	rect_f5	thoresen_2010_form-building-patterns_p90-91_pa...	annot_cue_011	NaN	1.50	141.0	NaN	NaN

Creating Timelines

Use create_timeline() to turn loaded events into a Timeline object:

# Physical timeline (seconds)
physical_tl = typed.create_timeline(uid="thoresen_physical")
physical_tl

ContinuousPhysicalTimeline[thoresen_physical] (11 events)
                      0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 142.5 seconds

physical_tl.get_timestamp_table()

pyarrow.Table
axis: double
thoresen_physical: double
----
axis: [[0,1.5,3.5,5,5.5,...,116.4,121,128.5,141,142.5]]
thoresen_physical: [[0,1.5,3.5,5,5.5,...,116.4,121,128.5,141,142.5]]

# Graphical timeline (pixels)
graphical_tl = graphical.create_timeline(uid="thoresen_graphical")
graphical_tl

DiscreteGraphicalTimeline[thoresen_graphical] (11 events)
                      0 ∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶∶ 760 pixels

Note: Both timelines represent the same 11 events in different coordinate systems: - Physical: 0 - 142.5 seconds (audio time) - Graphical: 10 - 760 pixels (image coordinates)

TimeToAlign! uses these dual representations to align graphical annotations with audio.

Child Timelines from Column Values with `group_by`

When your data carries events from multiple sources (images, pages, tracks), pass group_by to create_timeline() to split the events into one child timeline per unique value. The Thoresen data has events from five different image files:

from timetoalign.timelines import create_timeline  # noqa: E402

grouped_tl = create_timeline(typed, group_by="image_filename")
grouped_tl

ContinuousPhysicalTimeline[cpt1] (11 events, 5 children)
                      0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 142.5 seconds
  ├─ thoresen_...     0 ~                                5.5 (3 events)
  ├─ thoresen_...     0 ~~~~~~~~~~                       48 (2 events)
  ├─ thoresen_...     0 ~~~~~~~~~~~~~~~~~~               83.5 (2 events)
  ├─ thoresen_...     0 ~~~~~~~~~~~~~~~~~~~~~~~~~~       116.4 (2 events)
  └─ thoresen_...     0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 142.5 (2 events)

# Each child timeline represents events from one image
{
    "parent_id": grouped_tl.id,
    "n_children": grouped_tl.n_children,
    "children": {
        child.id: len(child._events) if child._events else 0
        for _, child in grouped_tl.iter_children()
    },
}

{'parent_id': 'cpt1',
 'n_children': 5,
 'children': {'thoresen_2010_form-building-patterns_p90-91_page1_1.jpeg': 3,
  'thoresen_2010_form-building-patterns_p90-91_page1_2.jpeg': 2,
  'thoresen_2010_form-building-patterns_p90-91_page1_3.jpeg': 2,
  'thoresen_2010_form-building-patterns_p90-91_page1_4.jpeg': 2,
  'thoresen_2010_form-building-patterns_p90-91_page2_1.jpeg': 2}}

Summary

Goal	How
Load an ms3 score TSV	`Ms3Loader()` + `loader.load(path)`
Map your own columns	Subclass `TsvLoader` / `CsvLoader`, set `start_column` etc.
Promote a column to a typed field	Name it in `column_specs` (`{"col": int}`)
Reach a nested JSON value	`Field("column", "nested")` as a coordinate attribute
Derive a coordinate	`ComputedField("end", formula="...")`
Split into child timelines	`create_timeline(loader, group_by="column")`

from timetoalign.loader.tabular import TsvLoader

class MyLoader(TsvLoader):
    start_column = "onset"
    duration_column = "dur"
    column_specs = {"pitch": int, "velocity": int}

Key Takeaway: Tabular loaders map CSV/TSV columns to TimeToAlign! events declaratively. Unnamed columns ride along as property columns; column_specs promotes the ones you want as typed fields; Field reaches nested JSON. For the full column-to-field mechanism — composite columns, semantic field types, and Step-2 promotion — continue to the in-depth CSV/TSV how-to.

How to Load Tabular Data

TL;DR

Setup

Loading Notes from TSV

Converting to pandas

Quick Statistics

Creating Timelines

Custom Loaders for Non-Standard Formats

The Simplest Custom Loader

Promoting Columns to Typed Fields with column_specs

Nested JSON Column Access with Field

Two Coordinate Systems from One File

Creating Timelines

Child Timelines from Column Values with group_by

Summary

Promoting Columns to Typed Fields with `column_specs`

Nested JSON Column Access with `Field`

Child Timelines from Column Values with `group_by`