Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
task_categories:
|
3 |
+
- text-classification
|
4 |
+
- text-generation
|
5 |
+
- token-classification
|
6 |
+
- feature-extraction
|
7 |
+
- sentence-similarity
|
8 |
+
language:
|
9 |
+
- en
|
10 |
+
tags:
|
11 |
+
- nlp
|
12 |
+
- parsing
|
13 |
+
- grammar
|
14 |
+
size_categories:
|
15 |
+
- 1M<n<10M
|
16 |
+
---
|
17 |
+
# Sentence Parse Trees
|
18 |
+
|
19 |
+
This dataset contains high-quality English sentences and their corresponding parse trees. It is derived from the [agentlans/high-quality-english-sentences](https://huggingface.co/datasets/agentlans/high-quality-english-sentences) dataset,
|
20 |
+
with parse trees generated using [spaCy](https://spacy.io/) `en_core_web_sm` package. Abnormally long or short parses have been removed.
|
21 |
+
|
22 |
+
## Dataset Description
|
23 |
+
|
24 |
+
- **Size:** 1 502 479 entries
|
25 |
+
- **Source:** [agentlans/high-quality-english-sentences](https://huggingface.co/datasets/agentlans/high-quality-english-sentences)
|
26 |
+
- **Parsing Tool:** spaCy
|
27 |
+
|
28 |
+
## Data Format
|
29 |
+
|
30 |
+
Each entry in the dataset is a JSON object with two fields:
|
31 |
+
|
32 |
+
1. `text`: The original sentence
|
33 |
+
2. `parse`: The parse tree in Lisp-style S-expression format
|
34 |
+
- Each list contains: syntactic dependency relation, the word, and the parse tree of its descendants
|
35 |
+
- Words are enclosed in double quotes and escaped where necessary
|
36 |
+
|
37 |
+
### Example Entry
|
38 |
+
|
39 |
+
```json
|
40 |
+
{
|
41 |
+
"text": "At present, this mechanism has become one of the three major pillars underpinning the development of China-UK relations.",
|
42 |
+
"parse": "(ROOT \"become\" (prep \"At\" (pobj \"present\")) (punct \",\") (nsubj \"mechanism\" (det \"this\")) (aux \"has\") (attr \"one\" (prep \"of\" (pobj \"pillars\" (det \"the\") (nummod \"three\") (amod \"major\") [...]"
|
43 |
+
}
|
44 |
+
```
|
45 |
+
|
46 |
+
## Usage
|
47 |
+
|
48 |
+
This dataset can be used for various natural language processing tasks, including:
|
49 |
+
|
50 |
+
- Syntactic parsing research
|
51 |
+
- Grammar analysis
|
52 |
+
- Language model training
|
53 |
+
- Linguistic studies
|
54 |
+
|
55 |
+
## Licence
|
56 |
+
|
57 |
+
Please refer to the license of the original [agentlans/high-quality-english-sentences](https://huggingface.co/datasets/agentlans/high-quality-english-sentences) dataset.
|