Update dataset card (final)

README.md

@@ -95,10 +95,6 @@ configs:
   data_files:
   - split: train
     path: phon/train-*
-- config_name: phon-long
-  data_files:
-  - split: train
-    path: phon-long/train-*
 - config_name: read
   data_files:
   - split: train
@@ -212,39 +208,13 @@ vowels `æ ə ɜ ē ɘ ɨ ā`, ...).
 
 ### Configurations (corpus parts)
 
-Each corpus part is a **config**; load one with the `name` argument. `phon-long`
-is a derived long-form view of `phon` (same audio, regrouped per recording with
-timestamped transcripts) — it adds no new hours.
+Each corpus part is a **config**; load one with the `name` argument.
 
 - **`dial`** — Dialectal speech — regional varieties of Lithuanian. Transcripts use an extended phonetic/dialectal notation (stress marks, palatalization, dialect-specific vowels). One phrase per file.
 - **`phon`** — Phonetically annotated speech. Full recordings are sliced into phrases using the Praat TextGrid `ortho` (orthographic) tier; the grids' `phono` (phonetic-transcription) tier is aligned to each phrase by time overlap and exposed in the **`phono`** column alongside `text`. The source grids additionally carry `phones`, `syll` and `words` tiers.
-- **`phon-long`** — Long-form (per-recording) view of the **`phon`** part: one row per **full, un-sliced recording** paired with a **timestamped** transcript. `text` and `phono` hold one `[s.sss-s.sss] …` line per phrase, `transcript` is the plain (untimed) transcript, and `transcript_vtt` is a WebVTT rendering. The bracket timestamps are collision-free (labels never contain `[`/`]`/newlines), so they parse unambiguously without a separate structured column. Same audio as `phon`, regrouped for long-form / streaming ASR and forced alignment (it does not add new hours — it re-views the `phon` audio). Timestamp format is derived from the data: phon recordings are sub-minute with millisecond boundaries, so brackets are seconds with 3 decimals.
 - **`read`** — Read speech (prompted reading). One phrase per file with normalized, lower-cased transcripts.
 - **`spon`** — Spontaneous speech (radio, TV, dictaphone, etc.). One phrase per file with normalized, lower-cased transcripts.
 
-#### `phon` vs `phon-long` — same audio, two views
-
-`phon-long` does **not** add new recordings or hours: it repackages the **same
-source audio** as `phon`, kept as whole recordings instead of per-phrase slices.
-That audio is therefore stored **twice** in this repo (~74 GB in each config), by
-design — so you can pick whichever granularity fits without slicing/concatenating
-yourself. **Load one view for your task;** you only need both if you want
-phrase-level clips *and* long-form context.
-
-| | `phon` | `phon-long` |
-|---|--------|-------------|
-| One row is | a single phrase | a full source recording (~30 s, ~15 phrases) |
-| Rows | 1,084,228 | 72,822 |
-| `audio` | phrase slice | whole recording, un-sliced (lossless passthrough) |
-| `text` | phrase transcript | timestamped multiline transcript (`[s.sss-s.sss] …`) |
-| `phono` | phrase phonetics | timestamped multiline phonetics |
-| Timing | `start_ms` / `end_ms` per row | in-line `[s.sss-s.sss]` brackets + `transcript_vtt` per row |
-| Extra fields | `segment_id` | `transcript`, `transcript_vtt`, `num_segments` |
-| Best for | utterance-level ASR / TTS | long-form / streaming ASR, forced alignment |
-
-The two link up via `file_hash`: every `phon` phrase carries the `file_hash` of
-its `phon-long` recording (and a `segment_id` index into that recording).
-
 ### Data Instances
 
 A representative example (the `audio.array` waveform is shown truncated):
@@ -269,18 +239,16 @@ A representative example (the `audio.array` waveform is shown truncated):
 
 ### Data Fields
 
-The configs do **not** all share one schema: `phon-long` carries its own columns,
-and within the other parts several fields are populated only where the source
-encodes them. Field meanings first, then a per-config availability matrix.
+The configs do **not** all share one schema: within each part a field is
+populated only where the source encodes it. Field meanings first, then a
+per-config availability matrix.
 
 | Field | Type | Description |
 |-------|------|-------------|
 | `audio` | `Audio` | Decoded to a waveform `array` + `sampling_rate` (44100), backed by mono FLAC bytes. |
-| `text` | `string` | Transcript (normalized; `dial` = phonetic/dialectal; `phon-long` = timestamped multiline, one `[s.sss-s.sss] …` line per phrase). |
-| `phono` | `string` | Phonetic transcription, SAMPA-style, from the TextGrid `phono` tier (`phon` = per phrase; `phon-long` = timestamped multiline). |
-| `transcript` | `string` | Plain transcript, phrases joined, no timestamps (`phon-long`). |
-| `transcript_vtt` | `string` | WebVTT (`HH:MM:SS.mmm`) rendering of the timestamped transcript (`phon-long`). |
-| `category` | `string` | Corpus part: `dial` / `phon` / `phon-long` / `read` / `spon`. |
+| `text` | `string` | Transcript (normalized; `dial` = phonetic/dialectal). |
+| `phono` | `string` | Phonetic transcription, SAMPA-style, from the TextGrid `phono` tier (`phon` = per phrase). |
+| `category` | `string` | Corpus part: `dial` / `phon` / `read` / `spon`. |
 | `speaker_id` | `string` | Speaker code, where the path encodes one. |
 | `gender` | `string` | `Male` / `Female`. |
 | `age_group` | `string` | `0-12` / `13-17` / `18-60` / `60+`. |
@@ -292,34 +260,30 @@ encodes them. Field meanings first, then a per-config availability matrix.
 | `file_hash` | `string` | Content-hash id of the source recording. |
 | `segment_id` | `string` | Phrase index within the source recording. |
 | `start_ms` / `end_ms` | `int32` | Phrase boundaries within the source recording. |
-| `num_segments` | `int32` | Number of phrases in the recording (`phon-long`). |
 | `duration_ms` | `int32` | Clip duration in milliseconds. |
 
 #### Field availability by config
 
 **●** populated · **○** present but always empty · **–** not a column in this config.
 
-| Field | `dial` | `phon` | `phon-long` | `read` | `spon` |
-|-------|:------:|:------:|:-----------:|:------:|:------:|
-| `audio` | ● | ● | ● | ● | ● |
-| `text` | ● | ● | ● | ● | ● |
-| `phono` | – | ● | ● | – | ○ |
-| `transcript` | – | – | ● | – | – |
-| `transcript_vtt` | – | – | ● | – | – |
-| `category` | ● | ● | ● | ● | ● |
-| `speaker_id` | ● | ○ | – | ● | ○ |
-| `gender` | ● | ● | ● | ○ | ● |
-| `age_group` | ● | ● | ● | ○ | ● |
-| `lang_type` | ● | ● | ● | ○ | ● |
-| `source` | ● | ● | ● | ○ | ● |
-| `region` | ● | ○ | – | ○ | ○ |
-| `county` | ○ | ○ | – | ● | ○ |
-| `file_path` | ● | ● | ● | ● | ● |
-| `file_hash` | ○ | ● | ● | ○ | ● |
-| `segment_id` | ○ | ● | – | ○ | ○ |
-| `start_ms` / `end_ms` | ○ | ● | – | ○ | ○ |
-| `num_segments` | – | – | ● | – | – |
-| `duration_ms` | ● | ● | ● | ● | ● |
+| Field | `dial` | `phon` | `read` | `spon` |
+|-------|:------:|:------:|:------:|:------:|
+| `audio` | ● | ● | ● | ● |
+| `text` | ● | ● | ● | ● |
+| `phono` | – | ● | – | ○ |
+| `category` | ● | ● | ● | ● |
+| `speaker_id` | ● | ○ | ● | ○ |
+| `gender` | ● | ● | ○ | ● |
+| `age_group` | ● | ● | ○ | ● |
+| `lang_type` | ● | ● | ○ | ● |
+| `source` | ● | ● | ○ | ● |
+| `region` | ● | ○ | ○ | ○ |
+| `county` | ○ | ○ | ● | ○ |
+| `file_path` | ● | ● | ● | ● |
+| `file_hash` | ○ | ● | ○ | ● |
+| `segment_id` | ○ | ● | ○ | ○ |
+| `start_ms` / `end_ms` | ○ | ● | ○ | ○ |
+| `duration_ms` | ● | ● | ● | ● |
 
 ### Data Splits