Datasets:

nctahiti
/

Miroir-IME

Modalities:

Dataset Preview

Split (1)

train

The full dataset viewer is not available (click to read why). Only showing a preview of the rows.

The dataset generation failed

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
sample_id: string
label: string
char_count: int64
stroke_count: int64
strokes: list<item: struct<points: list<item: struct<x: double, y: double, t: int64, p: double>>>>
  child 0, item: struct<points: list<item: struct<x: double, y: double, t: int64, p: double>>>
      child 0, points: list<item: struct<x: double, y: double, t: int64, p: double>>
          child 0, item: struct<x: double, y: double, t: int64, p: double>
              child 0, x: double
              child 1, y: double
              child 2, t: int64
              child 3, p: double
metadata: struct<source_app: string, captured_at: timestamp[s], template_spacing: int64, dpi: int64>
  child 0, source_app: string
  child 1, captured_at: timestamp[s]
  child 2, template_spacing: int64
  child 3, dpi: int64
capture_source: string
original_label: string
model: string
version: int64
corrected: bool
bounds: struct<w: double, h: double>
  child 0, w: double
  child 1, h: double
dataset_id: string
to
{'version': Value('int64'), 'dataset_id': Value('string'), 'capture_source': Value('string'), 'model': Value('string'), 'label': Value('string'), 'original_label': Value('string'), 'corrected': Value('bool'), 'strokes': List({'id': Value('int64'), 'points': List({'x': Value('float64'), 'y': Value('float64'), 't': Value('int64'), 'p': Value('float64'), 'tilt': Value('float64'), 'orient': Value('float64'), 'dist': Value('float64'), 'z': Value('float64')})}), 'bounds': {'w': Value('float64'), 'h': Value('float64')}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1816, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              sample_id: string
              label: string
              char_count: int64
              stroke_count: int64
              strokes: list<item: struct<points: list<item: struct<x: double, y: double, t: int64, p: double>>>>
                child 0, item: struct<points: list<item: struct<x: double, y: double, t: int64, p: double>>>
                    child 0, points: list<item: struct<x: double, y: double, t: int64, p: double>>
                        child 0, item: struct<x: double, y: double, t: int64, p: double>
                            child 0, x: double
                            child 1, y: double
                            child 2, t: int64
                            child 3, p: double
              metadata: struct<source_app: string, captured_at: timestamp[s], template_spacing: int64, dpi: int64>
                child 0, source_app: string
                child 1, captured_at: timestamp[s]
                child 2, template_spacing: int64
                child 3, dpi: int64
              capture_source: string
              original_label: string
              model: string
              version: int64
              corrected: bool
              bounds: struct<w: double, h: double>
                child 0, w: double
                child 1, h: double
              dataset_id: string
              to
              {'version': Value('int64'), 'dataset_id': Value('string'), 'capture_source': Value('string'), 'model': Value('string'), 'label': Value('string'), 'original_label': Value('string'), 'corrected': Value('bool'), 'strokes': List({'id': Value('int64'), 'points': List({'x': Value('float64'), 'y': Value('float64'), 't': Value('int64'), 'p': Value('float64'), 'tilt': Value('float64'), 'orient': Value('float64'), 'dist': Value('float64'), 'z': Value('float64')})}), 'bounds': {'w': Value('float64'), 'h': Value('float64')}}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

version int64	dataset_id string	capture_source string	model string	label string	original_label string	corrected bool	strokes list	bounds dict
1	0303a44bd13995d5d8a7d19687bb5766e51be1a51fdff4e213bb6c7c117fe73c	onyx-boox-noteair5c	human	aa	aa	false	[ { "id": 0, "points": [ { "x": 0, "y": 0, "t": 0, "p": 0.0408, "tilt": 0, "orient": 0, "dist": 0, "z": 0 }, { "x": 0, "y": 0.3, "t": 0, "p": 0.0408, "tilt": 0, "orient": 0, ...	{ "w": 168, "h": 47.2 }
1	f328189ed22345dea6468c3cb6a20f7cd3891d22b0f3458ec6abbb691178d901	onyx-boox-noteair5c	human	aa	aa	false	[ { "id": 0, "points": [ { "x": 0, "y": 0, "t": 0, "p": 0.1096, "tilt": 0, "orient": 0, "dist": 0, "z": 0 }, { "x": -0.2, "y": -0.4, "t": 0, "p": 0.1096, "tilt": 0, "orient": 0, ...	{ "w": 164.7, "h": 48.2 }
1	9314ade50ab4f7ccedc0f09b1f06c56921b59760d5e71c4778ed03b85835f594	onyx-boox-noteair5c	human	ab	ab	false	[ { "id": 0, "points": [ { "x": 0, "y": 0, "t": 0, "p": 0.1805, "tilt": 0, "orient": 0, "dist": 0, "z": 0 }, { "x": -0.1, "y": -0.2, "t": 0, "p": 0.1805, "tilt": 0, "orient": 0, ...	{ "w": 157.7, "h": 128.9 }
1	d87a9dac37c3d1f173b6e899e00c8e38bec677d32f02f7a472e40641be468f3a	onyx-boox-noteair5c	human	ab	ab	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.0869,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 149.8, "h": 129.4 }
1	a14129068c94a5fd961429f9df7b8ac53baba5c20eff47fd99d20e06f266bb59	onyx-boox-noteair5c	human	ab	ab	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.0984,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 136.2, "h": 129.9 }
1	bcfc3cc62c3173b555779344cf1d404beaeb872e687a79d58aa7c41671c8e93e	onyx-boox-noteair5c	human	ac	ac	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.0772,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 137.9, "h": 44.9 }
1	3dc8dcf0fb59699c4ac267be480b7847f0ada959a47981a8317be64d730db798	onyx-boox-noteair5c	human	ac	ac	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.1924,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 160, "h": 47.5 }
1	2dab71f4d18c62a7ada620f28ef36df78aa88c56f8f4de6d909c04f48c0bd327	onyx-boox-noteair5c	human	ac	ac	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.0427,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 156.4, "h": 44.6 }
1	743bf2babce119ba9e0a2ee10158b39181e196aa0592cf9afa016cfc0cb6d3a8	onyx-boox-noteair5c	human	ad	ad	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.0381,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 177, "h": 122.8 }
1	bdce3b09ef58fb0cc16fecbae54b068adbff8c749a8eac033463170ae46ce05a	onyx-boox-noteair5c	human	ad	ad	false	[{"id":0,"points":[{"x":0.0,"y":0.0,"t":0,"p":0.0835,"tilt":0.0,"orient":0.0,"dist":0.0,"z":0.0},{"x(...TRUNCATED)	{ "w": 153, "h": 121.2 }

End of preview.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Miroir-IME — Dataset d'écriture manuscrite (ODbL)

Dataset de paires (geste, label) capturées sur Boox Note Air 5C (stylet e-ink) via le Miroir IME. Format parnasse-dataset.v1.

📜 Licence

ODbL (Open Database License) — attribution requise, partage à l'identique.

📊 Contenu

Fichier	Échantillons	Points
`parnasse-dataset-full-*.jsonl`	~400	~420 000

Chaque ligne = un échantillon JSON :

{
  "sample_id": "sha256...",
  "label": "philosophie",
  "char_count": 12,
  "stroke_count": 8,
  "strokes": [{
    "points": [
      {"x": 496.6, "y": 1146.9, "t": 0, "p": 0.98},
      ...
    ]
  }],
  "metadata": {
    "source_app": "com.onyx.android.note",
    "captured_at": "2026-07-04T18:15:00Z",
    "template_spacing": 0.0,
    "dpi": 0.0
  }
}

🖊️ Format

x, y : coordonnées absolues (pixels)
t : timestamp absolu (ms depuis epoch)
p : pression [0.0, 1.0]
sample_id : SHA-256 du contenu (intégrité)
strokes : liste de strokes, chaque stroke = liste de points

🔒 Anonymisation

Les données personnelles (noms, numéros, CB) sont exclues via le flag 🔒 dans le Miroir IME. Les échantillons publiés ne contiennent que des mots isolés — pas de phrases continues, pas de noms propres, pas de données sensibles.

🧠 Usage

import json

with open('parnasse-dataset-full-*.jsonl') as f:
    for line in f:
        sample = json.loads(line)
        label = sample['label']
        strokes = sample['strokes']
        total_points = sum(len(s['points']) for s in strokes)
        print(f'{label}: {len(strokes)} strokes, {total_points} pts')

Pour l'entraînement HTR (Handwritten Text Recognition) :

Les strokes peuvent être convertis en images (rendu bitmap)
Les séquences de points peuvent être utilisées directement pour des modèles sequence-to-sequence
La pression et les timestamps permettent des features avancées

🏗️ Source

Produit par le Miroir IME — un clavier de capture manuscrite pour Android (e-ink). Le dataset est généré via l'outil de recyclage intégré :

Écrire au stylet dans le Miroir IME
Corriger/annoter les transcriptions (📌) ou exclure les données personnelles (🔒)
🌾 Récolter → le dataset est sauvegardé dans Downloads/Parnasse/

🌐 Communauté

Code : github.com/nctahiti/Miroir-IME
Branche : recyclage-dataset-odbl
Discussions : Hugging Face

« Le Miroir est la surface où la pensée se reflète. » — Capitaine

Downloads last month: 62

Total file size:

47.2 MB